Board of Trustees of Western Michigan University through its Medieval Institute

Publications

The Ancestry and Kinship of Tancred, Prince Regent of Antioch

Author(s): Francesca Petrizzo
Source: Medieval Prosopography , 2019, Vol. 34 (2019), pp. 41-84
Published by: Board of Trustees of Western Michigan University through its Medieval
Institute Publications

Stable URL: https://www.jstor.org/stable/10.2307/27041706

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/10.2307/27041706?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Board of Trustees of Western Michigan University through its Medieval Institute Publications is
collaborating with JSTOR to digitize, preserve and extend access to Medieval Prosopography

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
 The Ancestry and Kinship of Tancred,
Prince Regent of Antioch
Francesca Petrizzo
Leverhulme Study Abroad Scholar, Università di Roma 2 Tor Vergata

Abstract: Tancred, youngest of the leaders of the First Crusade, is at

once one of the better-known and less studied members of the cru-
sader movement. Closely related to well-known Normans of South-
ern Italy such as Bohemond and Robert Guiscard, Tancred sits at an
important knot in the Hauteville family tree, but inquiry into his
family connections has been fragmented. This article aims at thor-
oughly reviewing the primary and secondary evidence, producing
a new and updated family tree for Tancred, and clarifying many of
the obscure points about his kin and ancestry. By reviewing Latin,
Greek, Arabic, and Armenian sources, and bringing together mate-
rial from both Southern Italy and the crusader East, the article will
investigate Tancred’s probable date of birth; the debated parentage
of his mother Emma, with the consequences for Tancred’s relation-
ship with Bohemond; the biography of Tancred’s father Odo; the
probable place of Tancred’s birth and raising, and its implications
for his identity; the existence and number of Tancred’s siblings,
both brothers and sisters, with the consequences for Tancred’s rela-
tionship to fellow crusaders Roger of Salerno and Richard of the
Principate; finally, Tancred’s marriage to Cecile of France. Thus,
the article will examine Tancred’s family tree as a case study for an
ambitious, far-reaching Hauteville kin group’s connections across
the Mediterranean.

T ancred, prince of Galilee, prince regent of Antioch, is at

once one of the better-known and least studied leaders of the First
Crusade. One of the Hautevilles of Southern Italy, related to Guiscard,
Roger of Sicily, and Bohemond, Tancred was a representative of one of
the more ambitious and better-connected kin groups of the medieval
Mediterranean.1 His career as a crusader was both highly visible and var-

1
It is important to stress that, while a useful scholarly convention, the use of “Haute-
ville” as a family name for this clan does not rest on any extensive usage from their time.
Hauteville is only mentioned as a place of origin by one of the narrative sources of the

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
42   Francesca Petrizzo

ied; belonging first to the Southern Italian contingent with Bohemond,

taking the cross in 1096, Tancred then switched to the service of Godfrey
of Bouillon after the conquest of Antioch in 1098, returning to Antioch
and its rule as regent after Godfrey’s death in 1100. One of the founders
of the crusader polities in Outremer, and a popular figure in contempo-
rary and later accounts of the crusades, Tancred has nonetheless been the
object of very little specialized study.
The most recent biography is the one published by Paul Nicholson in
1940; Tancred has since been the object of inquiry jointly with Bohemond,

time, Geoffrey Malaterra, and then only once (Geoffroi Malaterra, Histoire du Grand
Comte Roger et de son frère Robert Guiscard. Vol. I. Livres I&II, ed. and trans. Marie-Agnes
Lucas-Avenel [Caen: Presses universitaires de Caen, 2016], 1.3). De Altavilla was only used
as a family name by two members of the Loritello-Catanzaro branch of the Hautevilles,
descendants of Geoffrey, son of the patriarch Tancred, further discussed below. These, uncle
and nephew, were both named William of Hauteville. The first William, attested from
1085 (Documenti latini e greci del conte Ruggero I di Calabria e di Sicilia, ed. Julia Becker
[Rome: Viella, 2013], no. 9, 60–61), and who was dead by 1133 (Léon-Robert Ménager,
Les actes latins de Santa Maria di Messina (1103–1250) [Palermo: Istituto siciliano di studi
bizantini e neoellenici, 1963], no. 5, 370–77), is probably to be identified with the William
of Hauteville who witnessed a charter in Josaphat in 1114 (Reinhold Röhrict, Regesta regni
Hierosolymitani [Innsbruck, 1893], n. 73a). This is made extremely likely by the fact that
the issuer of the charter, Guido Capriolus, was one of the Caprioli, a family some of whose
members had been rebelling against the rule of countess Bertha of Loritello at the same
time as William, her brother-in-law (Chronica Trium Tabernarium, ed. Erich Caspar, in
“Die Cronik von Tres Tabernae in Calabrien,” Quellen und Forschungen 10 [1907]: 1–56,
at 39–41). As that rebellion appears to have taken place and to have been put down around
1110–1112, it makes it look feasible that William should spend some time abroad immedi-
ately afterward as the dust settled. The second William of Hauteville was the son of Robert
of Loritello, the first William of Hauteville’s brother. This second William had received the
lordship of Biccari from his own brother, Count Robert II of Loritello (Les chartes de Troia
(1024–1266): édition et étude critique des plus anciens documents conservés à l’Archivio Capi-
tolare, ed. Jean-Marie Martin [Bari: Società di Storia patria per la Puglia, 1972], no. 44,
171–72). Finally, a Mauger of Hauteville is attested once, in Antioch, in 1119 (Walter the
Chancellor, Bella Antiochena: Mit Erläuterungen und einem Anhange, ed. Heinrich Hagen-
meyer [Innsbruck: Wagner University, 1896], 2.3); the presence of William de Hauteville
in the same place a few years before suggests he was also a member of this secondary branch
of the family. As other members of the kin group were known by their titles (and indeed the
brothers of the first William of Hauteville were known as Robert, count of Loritello, and
Rao, count of Catanzaro), we may see that while the name of the town of origin of the fam-
ily was preserved in memory, and remained available as a title to minor members of a minor
branch of the family, other, more appealing titles were used when available. As we shall see
below, Tancred’s own approach to his name and titles was constantly evolving.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   43

or on the occasion of larger studies of the principality of Antioch. 2 As

such, Tancred has been written about in Cahen’s and Asbridge’s analyses
of the birth of the principality of Antioch, and Barber’s work on the cru-
sader states; he was also discussed in Bohemond’s two recent biographies
by Flori and Russo.3 This, however, has meant that little attention has been
paid to the tangled question of Tancred’s family tree since Nicholson’s
work, which was itself heavily reliant for the matter of his subject’s geneal-
ogy on De Saulcy’s painstaking, but now very old article from 1842.4
The consequences of this, in practical terms, are that at present
scholarship about Tancred’s origins and family is both scattered and out-
dated, and that several problematic points in the primary sources have
not been thoroughly examined or questioned for at least seventy years.
This has led to an accumulation of preconceived notions about Tancred’s
genealogy, and to what I would posit to be a systematic misunderstand-
ing of it. The study of Tancred’s family tree has also been complicated by
the often disjointed nature of the study of the Normans in Southern Italy
and Outremer, which has meant a lack of integration of the scholarship
on the topic. The result has been not merely a problematic treatment of
Tancred’s family tree, but also a misjudgment of the larger issues concern-
ing the mechanisms and patterns of behavior of the Hauteville kin group,
and of their investment and contribution to the Christian frontier in both
Sicily and Syria.
It will be the object of this article to review existing information
about Tancred’s genealogy, relying on both a reconsideration of the cru-
sader material, and a study of the Southern Italian primary sources edited
and studied since Nicholson’s work. I shall reassess what this informa-
tion can tell us about his kin group and its patterns of behavior, thus dis-
cussing Tancred not simply as an individual but rather as a member of

2
Robert Lawrence Nicholson, Tancred: A Study of His Career and Work in Their
Relation to the First Crusade and the Establishment of the Latin States in Syria and Palestine
(Chicago: The University of Chicago Libraries, 1940).
3
Claude Cahen, La Syrie du Nord à l’époque des croisades et la principauté franque
d’Antioche (Paris: Geuthner, 1940); Thomas S. Asbridge, The Creation of the Principality of
Antioch, 1098–1130 (Woodbridge: Boydell Press, 2000); Malcolm Barber, The Crusader
States (New Haven: Yale University Press, 2012); Jean Flori, Bohémond d’Antioche: Cheva-
lier d’aventure (Paris: Payot, 2007); Luigi Russo, Boemondo. Figlio di Guiscardo e principe di
Antiochia (Avellino: Sellino Editore, 2009).
4
Fèlicien De Saulcy, “Tancrède. 1er article: Origine de Tancrède,” Bibliothèque de
l’Ecole des chartes 4 (1843): 301–15.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
44   Francesca Petrizzo

a complex mechanism of dynastic expansion. I will be analyzing extant

information from both chronicles and charters, and contextualizing
Tancred’s paternity and ties within the larger Hauteville family network
in the Mezzogiorno and Sicily. It will be an aim of this article to produce
a new and updated family tree for Tancred, seeking to provide subsequent
researchers on the topic with a recent and thorough review of the fun-
damental family ties of the crusader ruler, and a renewed understanding
of Hauteville Mediterranean expansion. The endeavor is made challeng-
ing, but also fascinating, by the unique circumstances of Southern Italian
medieval sources.
The survival of numerous cartularies and archives, most of which
have been edited, for the great monastic institutions and cathedrals of sev-
eral Italian cities offers us an interesting glimpse into the dynamics and
society of the Norman invasion; at the same time, the coverage is patchy,
which means that next to well-documented kin groups and families, we
have obscure figures who only appear once or twice. 5 This has meant
that the prosopography of the Normans in the South is painstaking work,
in which disparate charter evidence needs to be collated with the exist-
ing narrative sources (themselves quite selective). 6 The endeavor can be
rewarding if time-consuming: scholars such as Léon-Robert Ménager and
Wolfgang Jahn have compiled thorough reviews of the material, seeking
to piece together as complete a picture as possible of the Normans in the
Italian South before the year 1100.7 As a member of the Hautevilles, who

5
See, for example, the charter collections for single rulers, such as the Documenti latini
e greci, ed. Becker, and Léon-Robert Ménager, Recueil des actes des ducs normands d’Italie,
1046–1127. 1. Les premier ducs (Bari: Grafica Bigiemme, 1980); the collections of single
monastic houses, such as the Registrum Petri Diaconi, ed. Jean-Marie Martin et al. (Monte-
cassino: Archivio dell’Abbazia, Reg. 3, 2016); and the great cartularies of particular cathe-
drals, such as Le Pergamene del Duomo di Bari (952–1264), ed. G. B. Nitto de Rossi and
Francesco Nitti (Bari, 1897). It is to be flagged up that one of the greatest documentary
reservoirs of Southern Italy, the archive of the Abbey of the Trinità di Cava, near Salerno,
is still by and large unedited for this period, having been published only recently to the
second half of the eleventh century (Codex Diplomaticus Cavensis, vols. 9–10, ed. Giovanni
Vitolo and Simeone Leone [Salerno: Laveglia, 1984–1990], vols. 11–12, ed. Carmine Car-
lone, Leone Morinelli, and Giovanni Vitolo [Salerno: Laveglia, 2015]).
6
The main narrative sources for this period in the South, Amatus of Montecassino’s
Ystoire de li Normant, William of Apulia’s Gesta Wiscardi, and Geoffrey Malaterra’s De
rebus gestis Rogerii et Roberti Guiscardi, will all be discussed below.
7
Léon-Robert Ménager, “Appendice. Inventaire des familles normandes et franques
émigrées en Italie méridionale et en Sicile (XIe–XIIe siècles),” in Roberto il Guiscardo e il

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   45

quickly rose from the obscurity of their ancestor, a Cotentin knight, to the
status of dukes of Apulia and kings of Sicily in two generations, Tancred
belongs to the best-documented kin group; nonetheless, the Hautevilles
were a prolific family, which, besides the well-known branches of the rul-
ers of Apulia and Sicily, originated numerous comital lines.8 As we shall
see throughout this article, to look at one of the Hautevilles is to inves-
tigate a complex network of interconnected alliances, one which often
makes for difficult work in unearthing the appropriate connections.
Let us begin by discussing Tancred’s age at the time of the First
Crusade, a topic which it is necessary to broach before discussing the
identity and genealogy of his mother, Emma; we shall then move on to
a discussion of Emma, as the fundamental link with the larger Hauteville
kin group with which Tancred was associated throughout his life. The
article will then examine Tancred’s father, the obscure Odo the Marquis,
integrating in this analysis recent scholarship on Roger I of Sicily’s rule
and entourage. This will also be the occasion to debate Tancred’s probable
place of birth and raising. I shall then move on to discussing Tancred’s
possible siblings, first his brothers and then his sister; this will allow me
to shed light on the parentage of Roger of Salerno, Tancred’s kinsman and
successor to the regency of Antioch, whose genealogy is itself controver-
sial. Finally, Tancred’s marriage to Cecile of France will be analyzed, with
the question of their lack of issue.

suo tempo. Atti delle prime giornate normanno-sveve, 1973 (Bari: Edizioni Dedalo, 1991),
260–390; Wolfgang Jahn, Untersuchungen zur normannischen Herrschaft in Süditalien
(1040–1100) (Frankfurt am Main: Peter Lang, 1989); for prosopography applied to indi-
viduals or families, see for example G. Portanova, “I Sanseverino dalle origini al 1125,” Ben-
edictina 22 (1976): 105–49; G. Tescione, “Roberto, conte normanno di Alife, Caiazzo, e S.
Agata dei Goti,” Archivio storico di Terra di Lavoro 4 (1975): 1–52.
8
The counts of Conversano, Loritello, Loreto, Catanzaro, and Principato (the counts
of Conversano and Principato will be discussed below) were all descended from the children
of Tancred, as were the princes of Capua, the second most important family in the Norman
South (for them, see Graham A. Loud, Church and Society in the Norman Principality of
Capua, 1058–1197 [Oxford: Clarendon Press, 1985]). As an introduction to this period in
the history of the Norman South, see Graham A. Loud, The Age of Robert Guiscard: South-
ern Italy and the Norman Conquest (New York: Longman, 2000); while for Norman Sicily,
besides the excellent introduction to the Muslim era by Salvatore Tramontana (L’isola di
Allah: Luoghi, uomini e cose di Sicilia nei secoli IX–XI [Torino: Giulio Einaudi, 2014]), one
unfortunately still has to rely either on Donald Matthew’s unfootnoted The Norman King-
dom of Sicily (Cambridge: Cambridge University Press, 1992), or John Julius Norwich’s
nearly novelistic The Kingdom in the Sun, 1130–1194 (London: Longmans, 1970).

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
46   Francesca Petrizzo

By doing this we can obtain a rounded perspective of Tancred’s

family history, and at the same time discuss him prosopographically as a
link in a larger, and tightly interconnected family network spanning the
Mediterranean and the royal families of France and Jerusalem.

Tancred’s Age at the Time of the First Crusade

Thanks to Fulcher of Chartres we have a precise date for Tancred’s death:
12 December 1112, as per Hagenmeyer’s thorough discussion of the
issue.9 This happened, Albert of Aachen tells us, after he was taken by a
violent illness; and afterwards, the prince regent was buried in the portico
of the Church of the Apostles, in Antioch.10 While we therefore have a
fairly precise picture of the way Tancred’s life ended, we cannot say the
same for the way it began. Tancred’s date of birth is a matter for informed
guesswork, and it is fundamental to establishing a number of facts about
his mother and her role in the kin group.
The overwhelming impression gained from the crusader sources is
that at the moment he took the cross with Bohemond (after the siege of
Amalfi, in 1096) Tancred was very young.11 Albert of Aachen describes
him as “tyro illustris” at his first appearance in the Historia Hierosolimitana,
a “distinguished young knight” in the Edgington translation.12 Guibert of
Nogent showed Tancred as a wise but highly vivacious, impatient youth:
“Tancredus ille, qui in bellis dominicis titulum sagacissimae iuventutis
nunc usque meretur et meruit, propriae vivacitatis impatiens ilico erupit,”
he says, having him impetuously interrupt his elders’ lengthy counsels.13
Ralph of Caen, Tancred’s biographer, appears to give us a more pre-
cise indication, albeit one which needs to be contextualized.14 His account

9
Fulcheri Carnotensis Historia Hierosolymitana (1095–1127), ed. Heinrich Hagen-
meyer (Heidelberg: Carl Winters, 1913), 2.47; and see Hagenmeyer’s note in ibid., 562n2.
10
Albert of Aachen’s History of the Journey to Jerusalem, ed. and trans. Susan B. Edg-
ington (Oxford: Clarendon Press, 2007), 12.8, 836.
11
The Deeds of the Franks and Other Pilgrims to Jerusalem, ed. and trans. Rosalind
Hill (Oxford: Clarendon Press, 1967), 1.4, 7.
12
Albert of Aachen’s History, 94–95.
13
Guibert de Nogent, Dei gesta per Francos et cinq autres textes, ed. R. B. C. Huygens
(Turnholt: Brepols, 1996), 194, lines 763–65.
14
This work exists in a widespread translation by Bernard S. Bachrach and David S.
Bachrach: The Gesta Tancredi of Ralph of Caen: A History of the Normans on the First Cru-

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   47

of Tancred’s life begins with a brief description of his family and of his life
before the crusade. “Adolescens iuvenes abilitate armorum, morum gravi-
tate sensu transcendebat,” “when he was an adolescent he was superior to
the young in his skill with arms, and to the old in wisdom,” says Ralph
about Tancred’s education.15 Such an indication of his age is immediately
followed, after a description of Tancred’s doubts about the compatibil-
ity of his military life with his Christian faith, with the narration of how
Tancred came to take the cross. 16 While therefore we are not explicitly
told by Ralph that Tancred went on crusade when he was still an adoles-
cent, and therefore under twenty, we would expect the highly laudatory
account to contain a description of his deeds after adolescence, if these
had occurred. Implicitly, I believe we should take Ralph’s indication to
refer to the age Tancred was when he took the cross.
Southern Italian maturity for males seems to have come at sixteen:
it is at this age that future king Roger II, for example, probably took over
comital rule from his mother Adelaide, in 1112, and at sixteen Bohemond
II took up his titles. 17 Engaging in war as soon as they were of age was
a family tradition for the Hautevilles: Geoffrey Malaterra, a chronicler
for Roger I of Sicily who seems to be the best informed about the fam-
ily, has it that “quam cito aetas permisit,” “as soon as their age allowed,”
they descended upon Italy.18 As an adolescent ready to depart for war, and
thus of age, who was very possibly already engaged in conflict under the
command of his relatives Bohemond and Roger I at the siege of Amalfi in
1096, Tancred was therefore probably born between 1076 and 1080.

sade (Burlington: Ashgate, 2005). While the translators do warn the reader that theirs is a
free translation, “sensum pro sensu rather than verbum pro verbo” (“Introduction,” 16), it is
necessary to flag up that, while quite readable, their work is often imprecise, does not give
the reader a sense of Ralph’s sophisticated, highly literary style, and often glides over the the
complexities of the Latin and the shades of its meaning. Therefore, while the translation
can be useful as a general introduction at undergraduate level, it is not helpful in getting
to grips with the text in a scholarly manner, and a direct approach to the Latin remains
necessary.
15
Radulphi Cadomensis Tancredus, ed. Edoardo D’Angelo (Turnholt: Brepols, 2011),
6, lines 18–19.
16
Tancredus, 6–7, lines 31–53.
17
Hubert Houben, Roger II: A Ruler Between East and West, trans. Graham A. Loud
and Diane Milburn (Cambridge: Cambridge University Press, 2002), 30–31.
18
Malaterra, Histoire du Grand Comte, 1.11.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
48   Francesca Petrizzo

It is worth noting, however, that Tancred appears to have had more

experience of command than we would expect for someone who had just
come of age: as soon as they crossed into the Balkans Bohemond entrusted
him with command of part of the army, which suggests he was already
convinced he could rely upon him, and Tancred immediately proved him-
self able to take care of his men independently.19 I would therefore suggest
that Tancred was at least eighteen at the time of the First Crusade, and
that he was therefore born in the late 1070s. This gives us a timeframe for
his mother, Emma, a woman who was of childbearing age in 1078.

Tancred’s Mother: Emma of Hauteville

Ascertaining the exact parentage of Emma of Hauteville, Tancred’s
mother, is of fundamental importance to correctly place Tancred within
the Hauteville family tree. The matter, however, is complex, and De Saulcy
had understandably given it up as undecidable.20 I believe, however, that
there is one hypothesis which we can regard as more likely given both the
primary sources and the circumstantial evidence.

Emma’s Father
One of the thorniest issues surrounding Emma is whether she was Robert
Guiscard’s daughter, or his sister, and therefore another daughter of
Tancred of Hauteville, the family’s patriarch. This is of course determinant
in ascertaining whether she was, in fact, Bohemond’s sister or his aunt,
with the necessary consequences for his relationship to Tancred. The issue,
as we shall see, is undoubtedly complex.
Most sources for the First Crusade describe Tancred as Bohemond’s
nephew. We find him identified as such in Albert of Aachen, who is
well informed on Tancred’s career and presumably had sources among
those who served with him under Godfrey of Bouillon (of Tancred and
Bohemond’s relationship he says, “iuxta eundem Boemundum avun-
culum suum,” and “Tancradum filium sororis mee,” this explicitly used
by Bohemond in a speech to describe his standard of behavior towards
his own family), and in Matthew of Edessa, whom we would expect

19
The Deeds of the Franks, 2.5, 10–11.
20
De Saulcy, “Tancrède,” 313.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   49

to be well informed as a writer close to Tancred’s zone of influence in

Antioch.21
The information is supported by Guibert de Nogent (“Tancredus … ex
Boemundi, nisi fallor, sorore filium”), Robert the Monk (“Tancredus vide-
licet nepos suus”), Baldric of Bourgueil (“Tancredum … nepotem suum”),
a Genoese chronicler, Cafarus (“et Tanclerium, nepotem Baiamundi ex
sorore”), and significantly Anna Komnene, whose writing is independent
from the crusader sources, and who had a keen interest in the Hautevilles
as some of the fiercest enemies of the Byzantine empire.22 Anna declares
that she does not know whether Tancred was the son of Bohemond’s
brother or sister, which, while suggesting that she has given the matter some
thought and is not idly reporting a rumor, also confirms the impression that
Bohemond and Tancred’s relationship was clear to an extent even to those
who did not know many details about their family tree.23
Fulcher of Chartres, neutrally, reports that Tancred was “eius
[Boamundi] propinquus,” that is closely related to him, thus neither deny-
ing nor outright supporting this declaration. 24 William of Tyre, a later
author, but one who had information on Tancred independent of the cru-
sader sources and was keenly interested in him, also follows the trend that
Tancred was Bohemond’s nephew (“domino Tancredo nepoti suo”).25
A more complicated picture, however, comes from three sources
particularly close to the Southern Italian perspective, which cannot there-
fore be easily discounted.
The Gesta Francorum, written by a man who served under both
Tancred and Bohemond and who came from Southern Italy, is unhelpful

21
Albert of Aachen’s History, 2.22, 94, and 4.15, 270; Ara Edmond Dostourian, Arme-
nia and the Crusades, Tenth to Twelfth Centuries: The Chronicle of Matthew of Edessa (Lan-
ham: University Press of America, 1993), 2.110, 164.
22
Guibert de Nogent, Dei gesta per Francos, bk. 3, lines 64–65; The Historia Ihero-
solimitana of Robert the Monk, ed. D. Kempf and M. G. Bull (Woodbridge: The Boydell
Press, 2013), bk. 11, 17; The Historia Ierosolimitana of Baldric of Bourgueil, ed. Steven Bid-
dlecombe (Woodbridge: Boydell Press, 2014), bk. 1, 12; Luigi Tommaso Belgrano, Annali
Genovesi di Caffaro e de’ suoi continuatori dal MXCIX al MCCXCIII (Rome, 1890), 113;
Annae Comnenae Alexias, ed. Diether R. Reinsch and Athanasios Kambylis, Part 1 (Berlin:
De Gruyter, 2001), 12.7, 378.
23
“οὐ γὰρ οἶδα σαφῶς εἰ πατρόθεν καì ἢ μητρόθεν τὴν πρὸς τὸν Βαἲμοῦντον ὁ Ταγγρὲ
συγγένειαν ἐκέκτητο,” Annae Comnenae Alexias, 12.7, 378.
24
Fulcheri Carnotensis Historia, 2.7, 393.
25
William of Tyre, Chronicon, ed. R. B. C. Huygens (Turnhout: Brepols, 1986), 11.1, 495.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
50   Francesca Petrizzo

in giving us no information at all about their degree of kinship. The author

is definitely interested in Tancred, on whom he kept tabs up until the siege
of Jerusalem, even after Tancred had switched his allegiance to Godfrey,
and the author was serving under Raymond of Toulouse.26 However, the
author identifies Tancred and his brother William simply as Marchisi fili;
nor is his lack of interest in degrees of kinship limited to them.27 He men-
tions no degree of kinship for Richard of the Principate either, despite our
knowledge of the fact that he was also related to Tancred and Bohemond.28
While puzzling, however, this lack of information does not help us sway
the debate one way or the other. I would suggest that the author was more
interested in his commanders’ military prowess and stated identity than
their complex dynastic relations; the Gesta Francorum’s simple, straight-
forward style and lack of sophistication was after all a problem for many of
its contemporaries, and several histories of the First Crusade expressly set
out to rewrite and deepen its narration.29
The most problematic, isolated, and conflicting testimony, how-
ever, comes from the Gesta Tancredi. Placed in a position of confidence
with both Bohemond and Tancred, Ralph is therefore arresting when he
says of Tancred’s parentage something which is worth quoting at length:
“a patre quidem haud ignobilis filius, a maternis autem fratribus nepos
longe sublimior. … matris vero fratres miliciae suae gloriam extra supraque
patriam, id est Normanniam, extulerunt” (“not a lowly son on his father’s
side, he was by far a more illustrious nephew from the maternal brothers.
Indeed, the brothers of his mother made known the glory of their war-
rior prowess outside and beyond their motherland, that is Normandy”).30
Here the language seems fairly irrefutable: Ralph refers to the “maternal
brothers” of whom Tancred is the illustrious nephew, the “brothers of the

26
On the Anonymous’s abandonment of the Southern Italian contingent, Rosalind
Hill, introduction to The Deeds of the Franks, xiii–xv; the Anonymous was still reporting on
Tancred all the way up to the siege of Jerusalem and the fighting at Ascalon (The Deeds of
the Franks, 10.38, 91–92, and 10.39, 95).
27
The Deeds of the Franks, 1.3, 5; 1.4, 7; 3.9, 21. In this last instance the author does
clarify that William was “frater Tancredi.”
28
The Deeds of the Franks, 1.3, 5 and 7. On page 7 Richard also has a brother explicitly
mentioned, Rainulf. See below, pp. 76–77, for a more detailed discussion of Tancred and
Richard’s kinship.
29
Rosalind Hill, introduction to The Deeds of the Franks, ix–xi.
30
Tancredus, 6, lines 2–6.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   51

mother” who came from Normandy. These must be the sons of Tancred
of Hauteville, conquerors of the South; and indeed, Ralph goes on to
describe how Guiscard was the scourge of Greeks and Germans, while
“Reliqui vero fratres numero XI Campaniam, Calabriam, Apuliam con-
tenti sunt debellare. Excipiendus est Rogerius, cui subacta gentilitas Sicula
gloriam peperit inter fratres a Wiscardo secundam.”31 But here is the rub:
Tancred of Hauteville had indeed twelve sons, but only eight (William
Iron-arm, Drogo, Humphrey, Geoffrey, Robert Guiscard, William,
Mauger, and Roger) descended on Italy.32 The rest remained in Normandy.
While setting out to glorify his patron’s family, Ralph appears to be less
informed than we would give him credit for. Either this, or what we are
facing here is a copy containing at least one fundamental mistake.
Is it conceivable that if the surviving text of Ralph’s biography is in
fact wrong about Guiscard’s family, it may in fact be wrong about his two
patrons’ relationship, or at least less clear-cut than he seems at first? We
have to consider three things: in the first place, that the Gesta Tancredi is
an exiguously transmitted text, of which only one, damaged copy survives,
making it impossible to check for copyists’ mistakes; secondly, that the
syntax of the paragraph is ambiguous; and finally, that the Hystoria de via
et recuperatione Antiochiae atque Ierusolimarum (also known as Historia
belli sacri) which is both heavily indebted to the Gesta Tancredi, and tied
to a Southern Italian environment, probably the Abbey of Montecassino,
tells a slightly different and no less complex story.33
Once more, it is necessary to quote the passage at length: “Adhesit
ei [Boamundus] quidam exadelfus suus, Tancredus nomine, nepos eius-
dem patris Robberti ducis, ex sorore sua nomine Emma genitus, a patre
quidam, qui Marchisus dictus est, non ignobilis filius; a matris autem fratri-

31
“The remaining brothers, eleven in number, limited themselves to conquering
Campania, Calabria, and Apulia. But we must single out Roger, who, having obtained Sicil-
ian fame, had glory second only to Guiscard.” Tancredus, 6, lines 11–14.
32
Graham A. Loud, The Age of Robert Guiscard: Southern Italy and the Norman Con-
quest (Harrow: Longman, 2000), 3, 299.
33
Edoardo D’Angelo, introduction to Hystoria de via et recuperatione Antiochiae
atque Ierusolymarum (olim Tudebodus imitatus et continuatus) (Firenze: SISMEL-Edizioni
del Galluzzo, 2009), xiii–xv, xxx–xxxiv. This text also survives in only one complete manu-
script, plus a partial copy.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
52   Francesca Petrizzo

bus nepos longe sublimior. … Reliqui fratres, octo numero, Campaniam,

Apuliam, Calabriam devellavere.”34
First, although the closeness of the two texts is self-evident, here
we see as well the Hystoria correcting the Gesta: while it still seems that
the Hystoria claims there were nine brothers in Italy, the author may have
easily miscounted and simply meant eight, or again, it may have been pos-
sible to be Southern Italian, in a prime Hauteville zone of influence (albeit
clearly writing at least twenty or more years after Tancred had died) and
yet get such details about them wrong.35
Either way, the Hystoria is clearly not following the Gesta in this
detail. More to the point, I believe that we can see here that the Hystoria
is either clarifying, or outright contradicting the Gesta when it comes to
Tancred and Bohemond’s relationship. Utmost care is necessary here in
translating the terms: Du Cange has it that exadelphus can mean “fratris fil-
ius, patruelis”; therefore nephew or, according to the definition of patrue-
lis, which is itself ambiguous, paternal uncle, brother-in-law or cousin.36
While D’Angelo, the editor, seems to take it that this may mean cousin, I
disagree and support Nicholson in translating as nephew.37 And while it
is true that nepos may mean nephew, it may also mean grandson.38 Once
this has been taken into consideration, I would thus translate the passage:
“Tancred, nephew of Bohemond, who was grandson of Guiscard, by his
(Bohemond’s) sister, by the name of Emma.” Although the syntax is no
doubt complex, and the unfortunate placing of the parenthetical phrase
concerning Guiscard might invite us to refer ejusdem to him, the main
meaning of exadelphus as nephew, and the possible alternative meaning of
nepos, mean that I believe we can take the Historia to also support the pic-
ture of Tancred as Bohemond’s nephew, rather than cousin. Once we take
this into account, the matris fratribus, instead of “brothers of the mother,”
could be the brothers on the mother’s side, something which Ralph’s first

34
Hystoria de via, 1.55–57, 9.
35
For the date of composition of the Hystoria de via, D’Angelo, introduction, xv–xvi.
36
Du Cange et al., Glossarium mediae et infimae latinitatis (Niort: L. Favre,
1883–1887), online ed., http://ducange.enc.sorbonne.fr/EXADELPHUS (accessed 19
July 2017); Jan F. Niermeyer and Co van de Kieft, Mediae Latinitatis Lexicon Minus (Lei-
den: Brill, 1976), 1012; Charlton T. Lewis and Charles Short, A Latin Dictionary (Oxford:
Clarendon Press, 1963), 1317.
37
D’Angelo, Hystoria de via, 1.55, 9n; Nicholson, Tancred, 12n6.
38
Niermeyer and van de Kieft, Lexicon, 934; Lewis and Short, Dictionary, 1200.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   53

choice of term maternis fratribus seemingly allows, while the double mean-
ing of nepos could allow us to construe Ralph as depicting Tancred as the
grandson, and great-nephew, of the sons of Tancred of Hauteville. All of
this seems to be undone, however, by what the Hystoria does in the midst
of the narration, having Tancred introduce himself as “Wiscardi nepos ex
sorore.”39 While this would seem to make the previous argument null and
void, it is important to keep it in mind when we consider something else:
the Hystoria is, in fact, a compilation of different texts spliced together.40
What we may be seeing here could be the internal confusion of the text
together with a strangely worded internal coherence, which, taken for all
in all, further invalidates the case for taking the Hystoria at face value.
Nor is the Hystoria de via the only source which features discrepan-
cies in its own text. As De Saulcy correctly flags up, Ralph of Caen refers
to Tancred as both Marchisides and Wiscardida, employing a patronymic
he could hardly apply if Tancred were not Guiscard’s grandson, and which
he indeed uses, in the variant Wiscadigena, to refer to Bohemond, who
was undoubtedly Guiscard’s son. 41 Most telling are two instances. In
lines 250–51, where we find Bohemond referred to as Wiscardigena, and
Tancred and his brothers as Wiscardidas, the closeness of the two attrib-
utes very clearly suggests a common origin, and while there is an easily
explainable distinction for the man who was actually a son of Guiscard,
we could hardly expect Tancred and his brothers to be granted such a title
from being Guiscard’s nephews.42 What is more, in line 88, Ralph thus
announces Bohemond and Tancred’s departure to the East: “confederati
ambo Wiscardide.”43 United by a common purpose and a common name,
the two are here posited as equally descended from Guiscard, which given
Bohemond’s paternity could only possibly be in the vertical line.
While we could suggest an instance of poetic license, Ralph is a
highly literate writer whose classical references are both plentiful and pre-
cise, and we could hardly expect him thus to abuse a classical patronymic,
which he correctly employs within the same text to refer to Guiscard’s son.44

39
Hystoria de via, 17.19, 114.
40
D’Angelo, introduction to Hystoria de via, xxiii–lv.
41
De Saulcy, “Tancrède,” 210–12.
42
Tancredus, 13.
43
Tancredus, 8.
44
D’Angelo, introduction to Tancredus, lxxvii–lxxxii.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
54   Francesca Petrizzo

Muratori, one of Ralph’s editors, did support the view of Tancred as

Bohemond’s cousin, but he was also hard-pressed to explain away the
patronymic, admitting that “is autem apertissime scribit Emmae fratrem
fuisse Robertum Guiscardum ducem, quanquam postea Tancredum
saepe Wiscardidam appellet, qui tamen ab ibso Guiscardo, si ei acqui-
escimus, minime descendebat.”45 In general, Ralph seems keen to present
Bohemond and Tancred’s relationship as a radically close one: Bohemond
is Tancred’s “cognatus sanguis,” quite literally blood-kin.46 Cousins are
of course also kin, but there seems to be a desire in Ralph to show them
sharing the same blood of Guiscard; something which Tancred could have
done only as a grandson, not a nephew.47 This point of view is further
strengthened by three further references to Tancred’s relationship with
Guiscard, even if more oblique ones.
Tancred is referred to as “Wiscardi sobole,” “offshoot of Guiscard,”
which again seems to emphasize direct descendance, with Tancred as a
cutting off Guiscard’s tree; when Arnulf of Chocques accuses Tancred of
duplicity and bad faith in appropriating loot from the Temple in Jerusalem,
he says of him “At indulgendum est Wiscardidae: secutus est enim patrum
suorum uestigia,” (“And this must be said for the Wiscardida: truly he has
followed in his fathers’ footsteps”), which once more ties Tancred directly
to Guiscard’s heritage, and which more firmly seems to put the duke in

45
“Despite the fact that he writes most clearly that Duke Robert Guiscard was
Emma’s brother, nonetheless afterwards he often calls Tancred Wiscardida, who, if we listen
to him, did not at all descend from the same Guiscard,” Ludovico Antonio Muratori, Gesta
Tancredi principis in expeditione Hierosolymitana, Rerum Italicarum Scriptores (Milan,
1724), 5:285–333, at 286.
46
Tancredus, 21, line 559.
47
The Normans in the South clearly did not have a simply patrilineal conception of
kin: as clearly shown not just by Tancred, but by the counts of Conversano, descendants
of a sister of Guiscard and Roger, who remained deeply entangled with their relatives, the
Hautevilles’ sense of dynasty was bilateral and cognatic, not agnatic as among, for instance,
the Franks (see Constance Bouchard’s important book Those of My Blood: Creating Noble
Families in Medieval Francia [Philadelphia: University of Pennsylvania Press, 2001]). This
was enduring through time, as Cuozzo has shown by recognizing the “horizontal concep-
tion of the family” of the Normans in the South over a hundred years after the first inva-
sion, in the 1150s (Errico Cuozzo, La cavalleria nel Regno normanno di Sicilia [Atripalda
Avellino: Mephite, 2002], 198). Nonetheless, Ralph of Caen (significantly not himself a
Southern Norman) through the employment of patronymics seems to seek to establish the
same kind of direct male descendance one may find among the Homeric heroes he liberally
mentions throughout the text.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   55

the roster of his ancestors, as uncles are hardly one’s “fathers”; finally, of
Tancred, his brother William, and Bohemond: “audaciam cognatione
generis germanam,” the “sister boldness from the kinship of their family,”
which, tied to the patronymics Ralph uses to refer to them all discussed
above, once more strongly implies that all men discussed descend from the
same ancestor, Guiscard.48
Context also strengthens the impression that it is Guiscard, and
Guiscard alone, with whom Tancred can claim such close blood-kinship:
should he be equally, for instance, a nephew of Guiscard and Count Roger
I of Sicily, it would become unclear why Tancred should answer particu-
larly for Guiscard’s perceived heinousness, instead of bearing the mantle of
being also a nephew of the conqueror of Muslim Sicily. It is true that we
must be careful in ascribing the characteristics of holy war to the conquest
of the island: William of Apulia only dedicated a few throwaway lines to
Roger’s lifelong war against Muslim Sicily in his Gesta Wiscardi, despite
it being dedicated to Pope Urban II, and Geoffrey Malaterra, Roger’s
own chronicler, candidly admitted that Roger’s first motive in wanting
Sicily was greed.49 Nonetheless, in the moment in which Ralph describes
Tancred being upbraided for his relation to the notorious Guiscard by his
own teacher Arnulf of Chocques, we would expect his hero, if it were pos-
sible, to exhibit a more crusade-appropriate relation as a defense. But it is
of his kinship with Guiscard, and his similarity with him, that Tancred is
accused, and of this kinship and no other that he must answer, as further
contextual proof of his exclusive closeness to the duke of Apulia. The longer
one looks at the Gesta Tancredi, the more one comes to suspect that what
we are looking at here is a coherent description of Tancred as Guiscard’s
grandson, with the first few lines engendering the confusion then possibly
replicated in the Hystoria through lack of revision and obscure syntax.
Summing up our considerations about the Gesta Tancredi and the
Hystoria de via, the only seemingly outright denial that Tancred and

48
Tancredus, 95, line 3226; 113, lines 3832–33; 13, lines 253–54.
49
Malaterra, De rebus gestis Rogerii Calabriae et Siciliae Comitis et Roberti Guiscardi
ducis fratris eius, ed. Ernesto Pontieri (Bologna: Nicola Zanichelli, 1925–1928), 2.28 (this
older edition has been used, awaiting the completion of the new edition of the work by
Lucas-Avenel, of which thus far only the first two books have been made available); Wil-
liam of Apulia, Gesta Roberti Wiscardi = Guillaume de Pouille, La Geste de Robert Guiscard,
ed., trans., and comm. Marguerite Mathieu (Palermo: Istituto Siciliano di Studi Bizantini e
Neoellenici, 1961), bk. 3, lines 194–203.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
56   Francesca Petrizzo

Bohemond were uncle and nephew comes from a source which, however
close to them, survives in one flawed copy; said copy contains at least
one significant error about their family; the syntax of the piece is both
complex and ambiguous; the use of patronymics and other images within
it seems to suggest an alternative interpretation; the other source (itself
sparsely attested) which follows Ralph of Caen appears to correct it and
may well be interpreted in a different way; and well-attested sources, con-
taining detailed information about Tancred, contradict it.
We have, however, two other testimonies we need to deal with.
Orderic Vitalis mentions Tancred’s father, Odo, and calls him “sororius”
of Guiscard, that is, his sister’s husband.50 It is here necessary to say that
Odo and his degree of kinship are mentioned on the occasion of Orderic’s
wholly fanciful tale of Sichelgaita as poisoner. Jamison observes that the
outlandishness of the occasion does not keep Orderic from observing cor-
rectly other characters’ relations to Guiscard, which is fair; but the cir-
cumstance does also suggest that Orderic is not, to put it mildly, wholly
concerned with the truth or reliability of the matter at hand, and that it is
hardly feasible to suggest that a man so cavalier about one matter might be
entirely scrupulous about another.51 What is more, Orderic could use the
word sororius elastically when he chose to: in narrating Bohemond’s rebel-
lion against his half-brother Roger Borsa, Orderic told of Bohemond’s
putative alliance with Jordan, prince of Capua. We know for sure that
Jordan was a nephew of Guiscard, through the latter’s sister Fressenda;
this is reflected by Orderic when he refers to him as Bohemond’s “conso-
brinus,” his cousin, but then, inexplicably, he calls him “sororius.”52 Either
Orderic is using the term to mean a generic relative, or he is less informed
on the Hautevilles’ relations than we might believe him to be.
Finally, we have Waleran of Noblat’s writing, the Scriptum Galeranni
Episcopi de miraculo Boimundi, an early twelfth-century source, celebrat-

50
Orderic Vitalis, Ecclesiastical History [Historia Ecclesiastica], ed. and trans. Mar-
jorie Chibnall (Oxford: Clarendon Press, 1978), bk. 7, 32; Niermeyer and van de Kieft,
Lexicon, 1273.
51
Evelyn Jamison, “Some Notes on the Anonymi Gesta Francorum, with Special Ref-
erence to the Norman Contingent from South Italy and Sicily in the First Crusade in Stud-
ies,” in French Language and Medieval Literature Presented to Professor Mildred K. Pope
(Manchester: Manchester University Press, 1939), 183–208, 196.
52
Orderic Vitalis, Historia Ecclesiastica, 4.32, 168. Fressenda will be further discussed
below.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   57

ing the intervention of St. Leonard in freeing Bohemond from captivity,

which Luigi Russo has persuasively argued seems to come from a source
close to Bohemond himself, given the amount of detail present in it
regarding obscure Southern Italian churches, unlikely to be known out-
side of it.53 Waleran is therefore arresting when he declares Tancred to be
Bohemond’s “consobrinus,” and indeed it is on his basis, and Ralph’s, that
Russo subscribes to the view of Tancred and Bohemond as cousins.54 Here
again we might argue about the exact meaning of the word, as consobri-
nus can, however rarely, mean nephew (and indeed Anna Komnene her-
self, while making clear that she thought that Tancred was Bohemond’s
nephew, since she wondered whether his mother or his father was
Bohemond’s sibling, had used a word, ἀνεψιαδὴς, which in most cases
means cousin, not nephew); we therefore have to register that the text is
quite suggestive.55
Ultimately, what we have is a number of sources, both related and
unrelated to each other, supporting Bohemond’s role as Tancred’s uncle;
two closely interrelated but problematic and ambiguous sources which
might, or might not, deny this; another source apparently further stating
the matter, but in circumstances hardly likely to inspire trust, and having
misused the term at least once; a final source seemingly strengthening the
case. It is necessary to underline something: while below I will make a case
for the fact that Roger of Salerno was, perhaps wilfully, misrepresented as
Tancred’s nephew rather than cousin to strengthen his claim to Antioch,

53
Waleran of Noblat, Scriptum Galeranni Episcopi de miraculo Boimundi, edited in
Acta Sanctorum, Novembris III, 160E–168F [= BHL 4874]; Luigi Russo, “Il viaggio di
Boemondo d’Altavilla in Francia (1106): Un riesame,” Archivio storico Italiano 163 (2005):
3–42, cf. 18–21.
54
AASS, Nov. III, coll. 160e–168f [= BHL 4874]: 162D; Luigi Russo, “Tancredi e i
Bizantini: Sui Gesta Tancredi in expeditione Hierosolymitana di Rodolfo di Caen,” Medioevo
Greco 2 (2002): 193–230, at 197n29.
55
The University of Chicago’s Logeion reports the meaning as registered by DMLBS,
http://logeion.uchicago.edu/index.html#consobrinus (accessed 23 July 2018), and Du
Cange, Glossarium, defined it with the ambiguous patruelis, discussed above, http://
ducange.enc.sorbonne.fr/consobrinus (accessed 23 July 2018). For ἀνεψιαδὴς see Annae
Comnenae Alexias, 11.3, 329; E. A. Sophocles, Greek Lexicon of the Roman and Byzantine
Period (New York: Charles Scribner’s Sons, 1900), 167; Lorenzo Rocci, Vocabolario Greco-
italiano (Rome: Società editrice Dante Alighieri, 1939), 148. Tancred’s mother, and the
passage which wonders unambiguously whether she is Bohemond’s sister (“ἀδελφὴ τοῢ …
Βαἴμούντου,” Annae Comnenae Alexias, 12.8.2, 378), are discussed below.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
58   Francesca Petrizzo

there would have been absolutely no reason to do the same for Tancred.56
Despite his young age, as repeatedly stressed, Tancred immediately rose
to the forefront of the crusader effort; while it was his uncle who brought
him on crusade, it was through his own strength that Tancred immedi-
ately blazed a trail, conquering Tarsus, rising among Godfrey of Bouillon’s
collaborators, and becoming known to the Byzantines as an enemy to be
feared.57 The fact that their relationship is not mentioned by the Gesta
Francorum makes it even more remarkable that those who reworked the
Gesta, such as Guibert of Nogent or Robert the Monk, included the infor-
mation, which suggests it was well advertised; non-Latin sources as sepa-
rate as Anna Komnene and Matthew of Edessa report it, and Albert of
Aachen, who was wholly independent from his contemporaries, close to
the Lotharingian faction Tancred temporarily served, and thoroughly well
informed on his career, also confirms it. From the sheer weight of evidence
alone, one would be tempted to throw in one’s lot with those who say that
Tancred was, indeed, Bohemond’s nephew; I would suggest, moreover,
that it is the circumstantial evidence about the Hautevilles’ ages that ulti-
mately makes it very unlikely that Emma could have been Guiscard’s sister
rather than daughter.
Tancred of Hauteville, the patriarch, married twice: first Muriella
and then Fressenda. An extraordinarily prolific man, he had twelve sons
and several daughters; their age gaps could be considerable, and the chil-
dren of Tancred held sway in Southern Italy from the 1030s, when William
Iron-arm first came to Italy, to 1101, when Roger I of Sicily died.58 The
only two daughters of the family we know of were married by the 1050s:
Fressenda, married to Richard prince of Capua, and the unknown sister
who bore Geoffrey of Conversano, both had sons of fighting age in the late
1060s.59 The last of the children of Tancred to arrive in the South and the

56
And indeed, it is my contention that because the famed Tancred and Bohemond
were very well known as nephew and uncle in the crusader states, it made it easy to suppose
the same relationship bound Tancred to the more obscure Roger of Salerno.
57
The Deeds of the Franks, 10.35, 84–87; Albert of Aachen’s History, 5.35, 383–84;
7.16–17, 506–10.
58
Loud, Age of Robert Guiscard, 77, 278.
59
For Fressenda’s marriage, Aimé du Mont-Cassin [Amatus of Montecassino], Ystoire
de li Normant, ed. Michèle Guéret-Laferté (Paris: Honoré Champion, 2011), 2.45, 112;
7.1, 292 and William of Apulia, Gesta Roberti Wiscardi, bk. 3, lines 637–40, which clearly
speak of this marriage without mentioning her by name; however, she is several times

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   59

youngest was Roger, who is attested in Southern Italy from circa 1055.60
Assuming that Malaterra was right when he claimed that the Hautevilles
went south as soon as they came of age (and we have no reason to doubt
him on such a practical point) this places Roger’s birth in the late 1030s.
Even admitting that Emma may have been his younger sibling, and she
may therefore have been born in the 1040s, this would put her in her late
thirties by the time Tancred was born; while it is of course possible for a
woman of that age to bear children, Tancred had at least one, and possibly
two, younger brothers.61 The number of children Emma appears to have
borne in the late 1070s–early 1080s suggests to me that she was, more
likely, a much younger woman: Guiscard’s daughter and not his sister, in
accordance with the indication of most of the sources.

Emma’s Mother
Once we accept that Emma was Guiscard’s daughter, we must then estab-
lish the identity of her mother. Guiscard married twice: first a Norman
noblewoman, Alberada of Buonalbergo, ca. 1051; when their marriage
was annulled because of consanguinity in 1058, he married Sichelgaita,
daughter of the Lombard prince of Salerno. 62 Nicholson categorically
denies that Emma could be Alberada’s daughter, given Malaterra’s state-
ment that “ex qua habebat filium” Bohemond, taking it to mean she only
gave Guiscard the one child.63 However, Malaterra’s wording suggests that
he meant that she gave him one son, a legitimate eldest whose claim would
later cause problems for the succession; he was presumably uninterested

mentioned in the charters of her son Jordan (see Graham A. Loud, “A Calendar of the
Diplomas of the Norman Princes of Capua,” Papers of the British School at Rome 44 [1981]:
99–143, notes 21, 26, 27, 30, 31, 36, 38, 23–26). Jordan was old enough to defend the city
of Aquino on his own in 1066 (Aimé, Ystoire, 4.11); Geoffrey was immediately identified as
Guiscard’s nephew at his first appearance in 1068 in Malaterra, De rebus gestis, 2.39, where
he had already accrued a number of castles through his own valor (strenuitate propria) and
would be expected to be in his twenties.
60
Julia Becker, Graf Roger I. von Sizilien. Wegbereiter des Normannischen Königreichs
(Tübingen: Max Niemeyer Verlag, 2008), 38. Becker still allows that a case can be made for
Roger’s arrival as early as 1052, which would put his birth even earlier.
61
See below, pp. 71–76.
62
Loud, Age of Robert Guiscard, 112–13; Hubert Houben, “Adelaide ‘del Vasto’ nella
storia del regno Normanno di Sicilia,” in Hubert Houben, Mezzogiorno normanno svevo.
Monasteri e castelli, ebrei e musulmani (Napoli: Liguori, 1996), 81–113, 85n21.
63
Nicholson, Tancred, 14; Malaterra, Histoire du Grand Comte, 1.30.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
60   Francesca Petrizzo

in any daughters, without any claims to it. De Saulcy and Rey supported
this point by suggesting that Alberada may well have given Guiscard other
children in their seven years together; but what is more, I find the point to
be conclusively supported by William of Apulia.64
Writing a panegyric poem for Guiscard, William appears to be well
informed about his family, and he explicitly states that Sichelgaita gave
Robert “quinque puellas,” five daughters.65 We have information about
five such daughters: Matilda, who married Ramon Berenguer, count of
Barcelona; Mabilia, married to William of Grantmesnil; an unknown
daughter married to Hugo of Este; Sybilla, married to Ebles of Roucy;
and finally Olympias, engaged to Constantine Doukas, son and heir to the
Byzantine emperor.66 This would suggest that Emma, his sixth daughter,
was in fact an issue of his first, Norman, marriage. Bohemond was born
after 1052 and before 1057, so we can locate his sister’s birth anywhere
between those dates or as late as 1059, assuming she might have been born
after the marriage was annulled.67 This would place her in her late teens
or early twenties in 1078, in prime childbearing age. It would also eas-
ily explain the close relationship between Bohemond and Tancred, his
nephew by his full sister, and it would furthermore explain why Emma
obtained a much less prestigious marriage to a more obscure man tied to
her uncle, Roger I, who gave many of his daughters in marriage to local
noblemen useful to his rule.68
Emma embodied a crucial intersection in the Hauteville kin group:
the fruit of Guiscard’s earlier, less ambitious but fundamental marriage

64
De Saulcy, “Tancrède,” 308; E. Rey, “Résumé chronologique de l’histoire des
princes d’Antioche,” Revue de l’Orient Latin 4 (1896): 321–476, 325.
65
“Edidit haec pueros sibi tres et quinque puellas,” William of Apulia, Gesta Roberti
Wiscardi, bk. 2, line 442.
66
Bünemann, in gathering the information concerning Guiscard’s daughters, also
suggests as his daughter the once-attested Gaitelgrima (Richard Bünemann, Robert Guisk-
ard 1015–1085. Ein Normanne erobert Süditalien [Köln: Böhlau Verlag, 1997], 254–58).
However, the only charter attesting her (Cava, Arm. Mag. C.2, n. 26) was identified as a
forgery by Graham Loud in “The Abbey of Cava, Its Property and Its Benefactors in the
Norman Era,” ANS 9 (1987): 143–77, 156. The most recent editors of the Cava charters,
Carlone, Morinelli and Vitolo, hypothesize that the charter was drawn on the model of
an original issued by Gaitelgrima, daughter of Gaimar IV of Salerno (Codex Diplomaticus
Cavensis, 11:105–7).
67
Flori, Bohémond, 27.
68
See below, p. 64.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   61

into a Norman kin group which remained his ally even after the bond was
annulled (and significantly, Alberada was buried with Guiscard and several
of his brothers in the family foundation in Venosa); she was Bohemond’s
closest relation; and she appears to have been unique among Guiscard’s
daughters in serving a marriage policy closer to the interests of his brother
Roger than his own.69
What is more, we have evidence that Emma may have also had an
active role to play in the Hauteville enterprises. In describing the Byzantine
offensive against Southern Italy and Bohemond in 1105, Anna Komnene
tells us that the Byzantine commander, Konstostephanos, attacked
Otranto, the closest port to the Balkans on the Italian peninsula.70 There,
however, the Byzantines were held up by a woman, who defended the for-
tress: “the mother, so it was said, of Tancred, whether she was a sister of
the notorious Bohemond or not I cannot tell, for I do not really know if
Tancred was related to him on his father’s or his mother’s side.”71 Anna
Komnene’s descriptions can have the self-admitted tint of myth, such as
the reported account of Sichelgaita brandishing spears and encouraging

69
G. Antonucci, “Note critiche per la storia dei Normanni nel Mezzogiorno d’Italia. I.
Alberada,” Archivio Storico per la Calabria e la Lucania, 4.1–3 (1934): 11–26. The study of
the importance of women as links in the marriage policy of the Hautevilles has been uneven
but persistent: see for example Catherine Heygate, “Marriage Strategies Among the Nor-
mans of Southern Italy in the Eleventh Century,” in Norman Expansion: Connections, Con-
tinuities and Contrasts, ed. Keith J. Stringer and Andrew Jotischky (London: Routledge,
2013), 165–86; Aurélie Thomas, “La carrière matrimoniale des fils de Tancrède de Haute-
ville en Italie méridionale: rivalités fraternelles et stratégies concurrentes,” in Les stratégies
matrimoniales (IXe–XIIIe siècle), ed. Martin Aurell (Turnhout: Brepols, 2013), 89–100;
Valerie Eads, “Sichelgaita of Salerno: Amazon or Trophy Wife?,” Journal of Medieval Mili-
tary History 3 (2005): 72–87; Vera von Falkenhausen, “Zur Regentschaft des Gräfin Ade-
lasia del Vasto in Kalabrien und Sicilien, 1101–12,” in Aetos: Studies in Honour of Cyril
Mango Presented to Him on April 14th, 1998, ed. Ihor Sevcenko and Irmgard Hutter (Stutt-
gart: B. G.Teubner, 1998), 87–115; Pietro Dalena, “‘Guiscardi Coniux Alberada’: Donne e
potere nel clan del Guiscardo,” in Roberto il Guiscardo tra Europa, Oriente e Mezzogiorno.
Atti del convegno internazionele di studio promosso dall’Università degli Studi della Basilicata
in occasione del nono centenario della morte di Roberto il Guiscardo (Potenza-Melfi-Venosa,
19–23 ottobre 1985), ed. Cosimo Damiano Fonseca (Lecce: Galatine, 1990), 157–80.
70
Annae Comnenae Alexias, 12.8, 378–79.
71
“ταύτην τὴν πόλιν γυνή τις ἐφρούρει, μήτηρ, ὡς ὲλέγετο, τοῢ Ταγγρέ, εἴτε ἀδελφὴ τοῢ ὲν
πολλοῒς ἤδη ῥηθέντος Βαἴμούντου εἴτε καὶ μή, συνιδεῒυ οὐκ ἔχω· οὐ γὰρ οἶδα σαφῶς εἰ πατρόθεν
καì ἢ μητρόθεν τὴν πρὸς τὸν Βαἲμοῦντον ὁ Ταγγρὲ συγγένειαν ἐκέκτητο,” Annae Comnenae
Alexias, 12.8.2, 378. Translation quoted from Anna Komnene, The Alexiad, trans. E. R. A.
Sewter, rev. Peter Frankopan (London: Penguin, 2009), 351.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
62   Francesca Petrizzo

the Normans to charge; however, her account of the woman who defended
Otranto is more sober and perfectly believable.72 Anna describes her buy-
ing time by pretending to parley, while sending secret messages to her
son, who marched to Otranto to relieve her. Hauteville women had occa-
sionally played a role in times of war: Roger I’s wife, Judith, had guarded
Troina, and Sichelgaita had been left in charge of the siege of Trani.73 As
we shall see below, it is possible that Emma may have had more sons than
Tancred and his brother William, and it is plausible that Bohemond, who
was already relying on Tancred to hold Antioch while he raised an army
in Europe, may have turned once more to his sister and her sons to help
guard one of the most important ports in the Mezzogiorno.
While we would expect a sister of Bohemond guarding one of his
castles to be identified as such to the Byzantines, Tancred was no mean
figure in their eyes: a few pages before recounting the incident possibly
featuring his mother, Anna Komnene describes his campaigns against
the empire and praises him as a powerful warrior and talented leader, one
of the strongest men of his age.74 As the one Hauteville the Byzantines
never succeeded in beating, a tale of Tancred’s warlike mother also besting
them would fit in with her overall perception of him and his family. While
Anna Komnene is not wholly sure about her identification of Emma, the
plausibility of the episode must grant it at least our attention; and our
final picture of Emma may well be that of a woman who, besides represent-
ing an important link in the wider Hauteville dynastic policies, was also an
active player in them.

Tancred’s Father: Odo the Good Marquis

Unlike his mother, whom we can contextualize with a degree of accuracy
inside the Hauteville kin group, Tancred’s father Odo the Good Marquis
is at once a better-documented and more shadowy figure. Tancred and his
brother William were sometimes identified, if not by their father’s name,
by his title: besides the abovementioned Gesta Tancredi, in the Gesta
Francorum and the Chanson d’Antioche they are known as “sons of the

72
Annae Comnenae Alexias, 4.6.5, 133–34.
73
Malaterra, Histoire du Grand Comte, 2.31; William of Apulia, Gesta Roberti Wis-
cardi, bk. 3, lines 668–73, 202.
74
Annae Comnenae Alexias, 12.2, see 4–5, 362–64.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   63

Marquis.”75 Even Ralph of Caen, who in his first chapter quickly glides
over Odo to praise Emma’s kin, adopts the classically inspired patronymic
Marchisides to refer to his hero and his brother.76 This is hardly surpris-
ing given that, at least in Southern Italy, Tancred seems to have used the
title himself. In his first attested appearance in a Greek-language charter
from Calabria, before the crusade, Tancred is listed among the witnesses
as “Τανκρεϛ μαρκ(η)σ(η)ς,” “Tankres markeses,” with his father’s title; it is
to be underlined, however, that when he later issued charters in Antioch
he first did it simply under his name, and later, after Bohemond returned
to Italy, as “Tancredus princeps.” 77 If throughout the sources Tancred
is indissolubly tied to the Hautevilles, the kin group to which he more
closely belonged, with whose members he went to war, his father’s mem-
ory appears nonetheless to have remained associated with his name.
While Ralph mentions Odo’s wealth, the charters provide proof
of his influence. Odo Bonus Marchisius appears in six charters of Roger
I, from 1095 to 1099.78 It is to be noted that we have scant survivals of
charters from Roger I’s reign owing to different causes, and indeed the
original copy of three of Odo’s charters was destroyed during the Second

75
The Deeds of the Franks, 1.3–4, 5–8; La Chanson d’Antioche, ed. Suzanne Duparc-
Quioc (Paris: Paul Geuthner, 1976), bk. 57, line 1386.
76
Tancredus, for instance 27, line 768; and 129, lines 429–30.
77
Becker, Documenti latini e greci, no. 6, 53–55. The editor misidentified Tancred as
his cousin Tancred of Syracuse; however, Vera von Falkenhausen (in Cristina Rognoni, Les
actes privés grecs de l’Archivo Ducal de Medinaceli (Tolède): Les Monastères de Saint-Pancrace
de Briatico, de Saint-Philippe-de-Bojoannès, et de Saint-Nicolas-des-Drosi (Calabre, XIe–XIIe
siècles), I [Paris: Association Pierre Belon, 2004], 334–37) had correctly recognized him.
Both suggest we may date the document to 1083; but the charter is in very bad condition
and the date of the year illegible, by their own admission (while the month, February, is
fairly clear). Given that Tancred would have been around five at the time, I suggest we
should rather place it as late as February 1092, as the first witness, Jordan, son of Roger of
Sicily (further discussed below) died in September of that year (Malaterra, De rebus gestis,
4.18). While Tancred would still have been underage at the time, around thirteen or four-
teen, Tabuteau showed that it was possible for Norman charters to be witnessed by under-
age boys, especially if related to the issuers; and Tancred, as the great-nephew of Roger
of Sicily, certainly was (Emily Zack Tabuteau, Transfers of Property in Eleventh Century
Norman Law [Chapel Hill and London: The University of North Carolina Press, 1988],
149–50). For his charters in Antioch, Liber Privilegiorum ecclesiae ianuensis, ed. D. Puncuh
(Genoa: n.p., 1962), no. 25, 42–43; Carte dell’Archivio Capitolare di Pisa, vol. 4 (1101–20),
ed. M. T. Carli (Rome: Storia e Letteratura, 1969), nos. 37–38, 80–83.
78
Becker, Documenti latini e greci, nos. 52, 54, 55, 56, 60, 67.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
64   Francesca Petrizzo

World War. 79 This should not therefore lead us to think that Odo was
only associated with Roger I from the date of his first charter with him, or
that he was a minor personage in his retinue. Indeed, on the basis of the
charters which do survive for him, Becker considers Odo one of the most
important noblemen in Roger’s court, and Ralph of Caen had praised his
wealth.80 It would also be possible to suggest that Odo may be the Othonus
whom Malaterra shows valiantly fighting for Roger near Taormina, circa
1079, a placement which would fit in well with Tancred’s birth date, and
explain why he was so well regarded later.81
For a man of Roger’s following to marry a Hauteville daughter was
not unusual: several of his own daughters married men who, while not
prestigious, were useful to Roger’s conquest of Sicily: Flandina married
his own brother-in-law Henry del Vasto; Adelisa married Count Henry
of Monte Sant’Angelo; Emma married Ralph of Montescaglioso; Judith
married Robert of Bassunvilla; a nameless daughter married Hugh of
Girgea. 82 Compared to Constance, who married Conrad, king of Italy,
and the nameless daughter who married Coloman, king of Hungary, these
marriages served a more precise purpose in the wider dynastic expansion:
tying to the Hautevilles men necessary to consolidate their dominions in
Italy.83 If indeed, as I believe, Emma was a daughter of Guiscard from his
first marriage, it would have made sense for her to be employed in this
more practical and less glamorous side of the family’s marriage policy.
If therefore such a wedding makes perfect sense in the wider con-
text of the Hautevilles’ dynastic strategy, uncertainty comes in with the
issue of Odo’s nationality. For a very long time, it has been customary to
think of him as a Northern Italian Frank, one of the many who answered
Roger I’s call to colonize Sicily. Rogadeo in the early twentieth century
believed Odo to be one of the Aleramici, Adelaide del Vasto’s kin group
that, with her marriage to Roger, came to play a fundamental role in Sicily,
but this explanation was not accepted by some in his own time, nor is it

79
Julia Becker, “Charters and Chancery under Roger I and Roger II,” in Norman
Traditions and Transcultural Heritage: Exchange of Cultures in the “Norman” Peripheries of
Medieval Europe, ed. Stefan Burkhardt and Thomas Foerster (Farnham: Routledge, 2013),
79–95; Julia Becker, introduction to Documenti latini e greci, 16–17.
80
Becker, Documenti latini e greci, 205–6; Tancredus, 6, line 17.
81
Malaterra, De rebus gestis, 3.18.
82
Houben, “Adelaide del Vasto,” 109–12.
83
Malaterra, De rebus gestis, 4.23–25.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   65

current now.84 The identification of Odo as a Northern Italian rather than

Norman seems to rest exclusively on his peculiar title, marquis, seldom if
ever used by the Normans but widespread in Northern Italy. This has been
recently contested by Becker, who rejected such an identification, and
in her edition of Roger’s charters she lists him as a “nobile normanno,” a
Norman nobleman.85 However, Ménager had found no grounds to iden-
tify Odo as such in his extensive inventory of the Normans in the South;
and Jamison believed his Northern Italian origin to be plausible.86
Odo’s title was certainly peculiar enough to be constantly associated
with his sons, to the point that while his name was routinely omitted, his
title survived even into the poetic appellatives of the Chanson d’Antioche,
where Tancred is called “fiux Marquis.”87 This, together with the fact that
the title is almost unique among the Normans in the South, and immi-
gration from Northern Italy into Sicily, leads me to believe that Odo was
indeed from Northern Italy, another useful man on the ground for the
Hautevilles, whose cooperation and support was rewarded with the hand
of a Hauteville in marriage, and whose sons became associated with the
Hautevilles’ enterprises.
The discussion of Odo’s ethnic origin also brings us to a complex
point: the description of Tancred as a “Norman.” The bibliography on
the discussion of Norman identity, both in the North and in the South
of Europe, is long and complex; the discussion of whether the Normans
in the South recognized themselves as Norman at all by the turn of the
twelfth century is an ongoing, and probably unsolvable, debate.88 Given

84
E. Rogadeo, “Gli Aleramici nell’Italia meridionale,” in Rassegna pugliese di scienze,
lettere ed arti 21.5–6 (1904): 133–54, cf. 139–41; Carlo Alberto Garufi, “Gli Aleramici
e i Normanni in Sicilia e nelle Puglie. Documenti e ricerche,” in Centenario della nascita
di Michele Amari (Palermo, 1910), 47–83; Henri Bresc, “Gli Aleramici in Sicilia: Alcune
nuove prospettive,” in Bianca Lancia d’Agliano fra il Piemonte e il regno di Sicilia. Atti del
Convegno (Asti-Agliano, 28/29 aprile 1990), ed. Renato Bordone (Alessandria: Edizioni
dell’Orso, 1992), 147–63.
85
Becker, Documenti latini e greci, 353.
86
Ménager, “Inventaire des familles normandes”; Jamison, “Some Notes,” 196–97.
87
Chanson d’Antioche, bk. 57, line 1386.
88
See for example, as a starting point: Graham A. Loud, “Gens Normannorum: Myth
or Reality?,” ANS 4 (1982): 104–16, 205–9, at 115; Cassandra Potts, “Atque unum ex
diversis gentibus populum effecit: Historical Tradition and the Norman Tradition,” ANS
18 (1995): 139–52; Marie-Agnès Lucas-Avenel, “La gens Normannorum en Italie du Sud
d’après les chroniques normands du XIe siècle,” in Identité et Ethnicité: Concepts, débats

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
66   Francesca Petrizzo

the historiographical convention, to identify Bohemond and Tancred as

“Southern Italian Normans” is both a necessity, and useful shorthand, and
it is therefore employed here. It seems however necessary to underline that
Tancred was most likely a descendant of Normans only on his mother’s
side, and that we only ever see him refer to himself as a Norman in the
words of Ralph of Caen, himself a native of Normandy, and one who
was writing after Tancred had died.89 In general, no sources for the First
Crusade associate the Southern Italian contingent with the Norman one,
under Duke Robert Curthose; as Murray has pointed out, the ethnic iden-
tity of the colonizers of the Principality of Antioch was not overwhelm-
ingly Norman, and as Asbridge registered, only the Sourdevals, besides the
Hautevilles, settled there from the crusader contingent.90 Odo’s Northern
Italian origin, and the apparently elastic identity of the entire Southern
Italian crusader contingent, introduce a further element of intersectional-
ity to Tancred’s identity, as a complex agent from a complex environment.
A man who clearly came from a noble background, and who enjoyed
success and trust in his environment, Odo appears to have outlived Emma
and to have remarried: we have an 1127 charter by a “Sichelgaita the
Marchioness, wife of the defunct Marquis Odo.”91 The two apparently had
children, who will be further discussed below. Furthermore, Odo’s posi-
tion, and the charters he witnessed, provide us with additional clues to
Tancred’s life—that is, his probable place of birth and upbringing.

historiographiques, examples (IIIe–XIIe siècle), ed. Véronique Gazeau, Pierre Bauduin, Yves
Modéran (Caen: Publications du Crahm, 2008), 233–64; Ewan Johnson, “Normandy
and Norman Identity in Southern Italian Chronicles,” ANS 28 (2004): 85–100; Elisabeth
van Houts, “Qui etaient les normands? Quelques observations sur des liens entre la Nor-
mandie, l’Angleterre et l’Italie au début du XIe siècle,” in 911–2011: Penser les mondes nor-
mands médievaux. Actes du colloque international de Caen et Cerisy (29 septembre–2 octobre
2011), ed. David Bates and Pierre Bauduin (Caen: CRAHAM, 2016), 29–46, 132–33;
and in the same volume, Houbert Houben, “Le royaume normand de Sicile ètait-il vrai-
ment ‘Normand’?,” 325–39.
89
Tancredus, 95, line 3225.
90
Alan V. Murray, “How Norman Was the Norman Principality of Antioch? Prole-
gomena to a Study of the Origins of the Nobility of a Crusader State,” in Family Trees and
the Roots of Politics: The Prosopography of Britain and France from the Tenth to the Twelfth
Century, ed. K. S. B. Keats-Rohan (Woodbridge: Boydell, 1997), 349–59; Asbridge, The
Principality of Antioch, 175–76; Andrew D. Buck, “Dynasty and Diaspora in the Latin
East: The Case of the Sourdevals,” Journal of Medieval History 44.2 (2018): 151–69.
91
“σηκληγαἴτης μαρκησης καì γυνῆς του ἀπεχωένου ὤτου μαρξέσου,” Syllabus grae-
carum membranarum, ed. F. Trinchera (Napoli, 1865), no. 98, 128–29.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   67

Tancred’s Place of Birth

Most of Odo’s surviving charters come from Sicily; his association with
Roger I is undoubted, and his sons seem to have been tied to his service
too.92 Bohemond took up the call to crusade while besieging Amalfi with
Roger’s army in 1096; we are told that most of the knights present at the
siege joined him, and a few lines later Tancred is mentioned first among his
followers, which suggests he was also at Amalfi and joined him then; and
Geoffrey Malaterra laments how many promising young men left Roger’s
army to join Bohemond.93 Bohemond had been intermittently fighting for
Roger since his uncle had settled his dispute with his half-brother Roger
Borsa in 1089; Tancred, given his position as great-nephew of the count
and son of one of his closer associates, is likely to have earned the fighting
experience discussed above in Roger’s wars on both the Continent and
Sicily.94
The last Muslim stronghold, Noto, had fallen in 1091; the conquest
of Sicily had taken about thirty years, and it was nowhere near fully set-
tled or pacified.95 Indeed, Roger I would maintain his capital at Mileto,
in Calabria, a place in which he clearly felt safer, and in one of his 1090s
charters he laments the wholesale destruction wreaked upon the island by
years of war.96 Such a theater appears ideal for a young, promising war-
rior to earn both experience and the trust of his relatives. What is more,
Tancred’s presence in the charter issued by Roger of Sicily and witnessed
by his son Jordan gives us further, suggestive indications about this.97

92
The exception being a 1096 document from Mileto, which Becker believes to be
false (Documenti latini e greci, 220).
93
The Deeds of the Franks, 1.4, 7; Malaterra, De rebus gestis, 4.24.
94
Ralph (“gentis illius pugnam congressu plurimo expertus,” Tancredus, 9, lines
121–23) seems to suggest Tancred was an expert in fighting the Greeks, something which,
as Nicholson points out, would be unlikely for someone who was an infant at the time of
Guiscard’s war against them (Tancred, 17n2). It is to be wondered whether, by illius gentis,
Ralph may not mean a catch-all, un-Western “other”; Russo, Boemondo, 45–50.
95
Loud, Age of Robert Guiscard, 172.
96
Becker, Documenti latini e greci, no. 33. It is to be underlined that this document
in particular has a complex transmission, and Becker expresses doubts on the authenticity
of its arenga.
97
Becker, Documenti latini e greci, no. 6, 53–55.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
68   Francesca Petrizzo

Jordan, Roger of Sicily’s eldest, illegitimate son by a concubine,

was an important player in the conquest of Sicily. 98 Left in charge of
the island by his father at least twice, even though he had once rebelled
against him, at Roger’s untimely death of illness in 1092 Jordan was his
heir apparent.99Lord of Pantalica and Noto, and later buried in Messina,
while active in his father’s court on the mainland as well as on the island
Jordan was clearly closely tied to Sicily.100 As discussed above, Tancred was
probably underage at the time he witnessed Roger’s charter together with
Jordan; his presence, in the absence of his father, suggests that he was close
to the two, and by then no longer bound to his own nuclear family.101 As
Bates has observed of William the Conqueror, while not of age until six-
teen, it was from around fourteen years old that Norman noblemen were
considered fit for military service.102 It is quite possible to envision that
Tancred, clearly a precocious and talented warrior, could have already
been serving with Jordan and Roger on the Sicilian frontier.
Discussing Jordan, moreover, ties into another important element
of Tancred’s character: according to the Hystoria de via, Tancred knew
Arabic well enough to use it to talk to the inhabitants of Antioch.103 Odo’s
lands between Maganoce and Limone were in one of the most enduringly
Muslim territories in the island, and if his son had grown up there he
would have had ample opportunity to both fight the Muslims on the edges
of Hauteville power, and learn their language at home. 104 In addition,
if indeed Tancred served with Jordan, he would have been dealing with

98
He is explicitly described as “filius … ex concubina” (Malaterra, De rebus gestis),
4.18, 97.
99
Malaterra, De rebus gestis, 3.36–37; 4.16, 18.
100
Malaterra, De rebus gestis, 4.18.
101
See n. 77.
102
David Bates, William the Conqueror (New Haven and London: Yale University
Press, 2016), 58–59.
103
“Riccardus autem de Principatu et Tancredus, qui Syriacam linguam sciebant,
consulebant cotidie ammirario, ut domino Boamundo redderet castrum,” Hystoria de via,
10.35, 70.
104
Becker, Documenti latini e greci, 206–7; Salvatore Tramontana, “Popolazione, dis-
tribuzione della terra e classi sociali nella Sicilia di Ruggero il Gran Conte,” in Ruggero il
Gran Conte e l’inizio dello stato normanno. Relazioni e comunicazioni nelle Seconde Giornate
normanno-sveve (Bari, Maggio 1975) (Rome: Il Centro di Ricerca Editore, 1977), 213–70,
240–41.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   69

Arabic-speakers firsthand: Elias, one of Jordan’s closest companions, was

a Muslim convert, and Jordan in general had such good relations with the
Muslims of Sicily, when he was not fighting them, that when he rebelled
against his father Roger feared Jordan might defect to the Muslims to
escape retribution.105
Even had Tancred not been one of Jordan’s collaborators, the
impression that fighting in Sicily was a valuable learning opportunity is
strengthened by the fact that according to the Hystoria de via, Richard of
the Principate also knew Arabic.106 A son of William, Robert Guiscard’s
younger brother and count of the Principato in Campania, Richard had
a brother, Tancred, who fought for Roger I and was by him made lord of
Syracuse in 1092.107 This Tancred had presumably been associated with
Roger before this to be in a position to receive such favors, and this paves
the way for Richard also to have been associated with Roger’s conquest of
the island; Richard’s knowledge of Arabic also brings us close to a Sicilian
theater for the earlier enterprises of these members of the Hautevilles
before the crusade.
Tancred was associated with Sicily in later sources as well: Paris
based his unsupported theory that Tancred was the son of a Muslim father
on the fact that the Chanson d’Antioche calls him, besides “fiux Marquis,”
“fius a l’amirant.”108 Amiratus, admiral, was indeed an exclusively Sicilian
title in the Italian Norman lands, and derived from the Arabic emir; but
Christians had been using it since Roger had entered Sicily, and at one
point a knight was left as admiral in charge of Palermo.109 Whether Odo
ever carried the title amiratus or not, the author of the Chanson or his

105
Malaterra, De rebus gestis, 3.30, 36.
106
See above, n. 103.
107
Becker, Graf Roger I., 78. Ralph suggests that Richard had held Syracuse first, then
offloaded it to his brother Tancred (Tancredus, 74, lines 1373–75); while this would seem
doubly unlikely, given the way the times line up and Roger I’s tight control over the assign-
ment of island lordships, the claim further strengthens the impression of the brothers’
closeness to Sicily.
108
Chanson d’Antioche, bk. 127, line 2866; Paulin Paris, La chanson d’Antioche (Paris,
1848), 2.372; Nicholson thoroughly refuted the theory in Tancred, 4–9.
109
“Reginam remeat Robertus victor ad urbem, Nominis eiusdem quodam remanente
Panormi Milite, qui Siculi datur amiratus haberi!,” William of Apulia, Gesta Roberti Wis-
cardi, bk. 3, lines 340–43; for a review of the institution of Sicilian admiralty see Léon-
Robert Ménager, Amiratus. L’émirat et l’origin de l’amirauté (Paris: S.E.V.P.E.N., 1960).

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
70   Francesca Petrizzo

source clearly associated Tancred and his family with the Sicilian theat-
er.110
As I have said above, Roger’s court was never permanently based in
Sicily; and Emma’s possible presence in Otranto in 1105 suggests that she
and her family may have maintained a stake on the mainland. At the same
time, Tancred, by the time he was in his late teens, had already acquired
considerable experience in war; he had learnt Arabic; he had a flexible
and practical approach to Muslims, being the only crusader to attempt to
protect them during the siege of Jerusalem, and later governing success-
fully an ethnically diverse populace in Antioch; he was the son of a father
closely associated with Roger I; he appears in a charter with Jordan, both
an ally and a fighter of Muslims in Sicily; he had been fighting and went
on crusade with Richard of the Principate, whose own family was clearly
invested in the conquest of Sicily; in later sources, he was remembered
with a title explicitly tied to Sicily.111
We can have no certainties on the matter, and it could be suggested
that Bohemond, who did not fight in Sicily, had in fact taken his nephew
with him long before Amalfi, as Calabria also contained Arabic speakers
from whom Tancred might have learnt the language, and where he might
have resided when he was called upon to witness his great-uncle’s char-
ter.112 At the same time, I believe that Tancred’s profile suggests that he
was either born and raised, or at least spent a considerable part of his life
before the crusade in Sicily; and that it is on the unstable, diverse Sicilian

110
It is to be noted that the Chanson d’Antioche refers to Bohemond as “Buiemons de
Sesile,” “Bohemond of Sicily” (bk. 51, line 1250), a rather aberrant reference for him, so
while suggestive, we cannot take this to be clinching evidence.
111
Asbridge, The Creation of the Principality, 189–97; The Deeds of the Franks, 10.35,
91–92.
112
Aimé, Ystoire, 5.11; Alex Metcalfe, The Muslims of Medieval Italy (Edinburgh:
Edinburgh University Press, 2009), 105. Even if Tancred and Bohemond had never fought
together before the crusade, their embattled closeness thereafter runs like a faultline
through all accounts of their relationship: as pointed out by Hodgson in her sophisticated
analysis of the Gesta Tancredi, the author, who had served both uncle and nephew, found
himself constantly torn by his loyalties between the two, characterizing Tancred as a rising
member of a different way of fighting on crusade, and the one to successfully embody the
transformation from predatory Norman to fully fledged crusader hero (Natasha Hodgson,
“Reinventing Normans as Crusaders? Ralph of Caen’s Gesta Tancredi,” ANS 30 [2007]:
117–32, cf. 124–28). We can see this tension explicated by the Gesta Francorum as well:
while traditionally viewed as a vindication of Bohemond, the source repeatedly praises
Tancred for his insistence in protecting and providing for his men during the crossing of

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   71

frontier that he acquired the skills which he employed once on crusade.

Ralph of Caen suggests that Bohemond recruited Tancred to follow him
on crusade, rather than having already had him in his service; and if indeed
Tancred, however young, had already proven himself in Roger I’s Sicilian
wars, we can see why his uncle would have thought him a valuable asset for
his campaign East, as the Hystoria de via shows when Tancred and Richard
appear to function as interpreters for Bohemond.113
Nor did Tancred go alone on the campaign: he certainly had with
him one, if not two, of his brothers.

Tancred’s Brother(s)
The issue of how many children Emma and Odo had, and how many went
on crusade is probably unsolvable; but it is possible for us to list at least
one certain brother for Tancred and one probable, and make at least a sug-
gestion for a further candidate.

William
William, filius Marchisi, is the brother whom most sources attribute to
Tancred, and of whose existence and details we can be the surest. Probably
younger given his lack of command and seeming inexperience, William
did not cross into the Balkans with his relatives and the Southern Italian
contingent, but rather joined Hugh of Vermandois and crossed with him
from Bari.114 William was known to writers on the Southern Italian con-
tingent, and his relationship with Tancred was well advertised.115 His join-
ing Hugh suggests that he was a free agent, able to offer his services to any
lord going on crusade.
William’s greatest moment in the spotlight, however, occurred at his
death. A fighter among the crusaders at the siege of Nicaea, William found
himself surrounded by Turks; Tancred tried to rush to his rescue, but he

the Balkans in spite of Bohemond’s orders not to antagonize the Byzantines, something
which could be read as a veiled reproach of the leader of the Southern contingent. The
Deeds of the Franks, 5.7ff., 10–14.
113
Tancredus, 7–8; see above, n. 103.
114
The Deeds of the Franks, 1.3, 5–6.
115
See p. 72.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
72   Francesca Petrizzo

arrived too late to save him.116 The episode, at once tragic and believable, is
portrayed affectingly in the sources: Albert of Aachen has it that Tancred
could only plant his banner by his brother’s body after cutting his way out
of the fray, and praises William’s beauty and daring; Ralph of Caen quotes
the Aeneid wondering how his hero felt seeing his brother fall; and the
Chanson d’Antioche has the young fighter cry out: “Tancred, where is your
standard?” before dying.117 A Hauteville looking for fortune, with the pro-
tection of a great lord of France, William appears to have been young and
rash, in the sources; he was presumably himself only just of age, and so able
to go to war with someone not of his family.
He is the only full brother of Tancred of whose existence we can be
sure, but not the only one of which we know.

Robert
Tancred’s brother Robert is only mentioned once, and then in passing, by
Ralph of Caen, at the moment of his catalogue of the Southern Italians
who went East: “Adde Wiscardidas Tancredum et fratres Willelmum
Robertumque,” “Tancred and his brothers William and Robert.”118 I have
said before that Ralph’s language and construction can be obscure, and
that the one extant version of the text cannot be checked. At the same
time, here the word order and the enclitic -que connecting Robert to
William seem unmistakable, and they appear to suggest that Tancred did
indeed have another brother who appeared on crusade. As a close member
of Tancred’s retinue Ralph would have the means to be informed about
someone too obscure to be noticed by other chroniclers.
If Tancred was in his late teens, and William just about of age,
it’s possible that Robert may have been underage, someone who was the
charge of his relatives, too young both for command and service of oth-
ers. We may be looking at someone who may have returned home before
the campaign got well under way, possibly indeed that son whom we
would find, ten years later, marching to Otranto to relieve Emma, if Anna
Komnene is to be believed.119 Given the many possible explanations for his

116
The Deeds of the Franks, 3.9, 21.
117
Albert of Aachen’s History, 2.39, 130; Tancredus, 28–29, lines 825–28; Chanson
d’Antioche, bk. 97, line 2107.
118
Tancredus, 13, lines 250–51.
119
Annae Comnenae Alexias, 12.8, 379.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   73

absence in other crusader chronicles, we must, at the very least, consider

Robert a possibility.

Rogerius Marchisi
As we approach the topic of the third possible full brother of Tancred of
whom we have traces, I must stress that my suggestion here is very tenta-
tive, as it rests on a single mention in a single charter, and simply on a pat-
ronymic. Rogerius Marchisi appears in 1094 as a witness to one of Roger
I’s documents.120 He appears nowhere else, and Becker could not identify
him other than as another member of Roger’s retinue, with lands in Agira
near Enna.121 My suggestion that we may be looking at another brother of
Tancred rests on four facts:
1. Rogerius Marchisi’s patronymic, which is the same as that by which
Tancred and William were known;
2. the uniqueness of the title, which is borne by no one else in Southern
Italy as either a patronymic or a personal title other than Odo and his
sons;
3. his contemporaneity in place, time and area of influence as Odo,
albeit apparently with an independent fief;
4. the likelihood that Odo, a prominent member in a comital court on
the frontier of a vast and barely conquered territory, might be expected
to be able to hold on to at least one son of his to inherit his territories.
However, Rogerius Marchisi might simply have been the son of another
Northern Italian marquis; or a Northern Italian himself. While I must
therefore reiterate that we are in extremely uncertain waters here, the pres-
ence of one other person carrying the patronymic “Marchisi” at the same
time as Tancred and his brother is striking, and the possibility that this
was also a relative of his must be flagged up.

Sichelgaita’s Children
Having dealt with the children of Emma and Odo, we need to look at the
latter’s second marriage with Sichelgaita. The passage in the charter which
attests Sichelgaita, issued by her in 1127, is worth quoting at length: “τοῦ

120
Becker, Documenti latini e greci, no. 40.
121
Becker, Documenti latini e greci, 165; Becker, Graf Roger I., 85.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
74   Francesca Petrizzo

ἐν μακαρια τὴ μνήμη κόμητος ρωκεριου, καì αυθεντου ἠμων, καì τοῦ αὐτου ὺιου
κόμιτος ρωκεριου, καì αυθεντου ἠμων· καì τῶν τεκνων αὐτου, καì ὔπερ ψυχεικῆς
σωτηριας τῶν πρòαπελθοτων ἠμῶν γοναῖων, καì τοὺ ἐμοῦ ἀυθεντου κυροῦ ὤτου
καì ἐμῆς, καì τῶν ημετερων τέκνων ” (“and [I] make this donation for the
happy memory of count Roger, and my lord, and his son count Roger, also
my lord; and for his children, and for the spiritual health of my departed
parents, and for my lord Odo, and for our children”).122 Clearly, the char-
ter attests that Sichelgaita and Odo had children; what is more, I believe it
suggests the said children were born after Tancred died. The charter refers
to Roger II as ὺιος of Roger I, that is a son; but it refers to his children,
and the children of Sichelgaita and Odo, as τέκνα, a gender-neutral term
usually applied to younger children.123 In 1127 Roger II’s sons were under
ten years old, and Sichelgaita’s status as a widow who used the title of mar-
chioness suggests that she was either the holder of the title with young
daughters, or a regent for at least one son; either way, in order for such
children to count as τέκνα they would have to be under sixteen in 1127,
and thus to have been born in or more probably after 1112.124
It is possible that these children are attested in their adult lives. In
the Catalogus Baronum, a list of the fief-holders who owed military service
to the kings of Sicily dating from the 1150s, we find Hugo and Johannes
Marchisii, apparently brothers given their names, and their children, who
held land in Molise.125 The possibility of the Marchisii being related to
Tancred was first raised in the seventeenth century, with Fernando della
Marra’s Discorsi delle famiglie estinte.126 Evelyn Jamison’s delicate comment
on his work is that the family needs to be studied “more fully”; this is
understandable, given that della Marra’s reconstruction of their ancestry is
fantastical, while still being deserving of attention.127 Della Marra posited

122
Trinchera, Syllabus, no. 98.
123
Henry George Liddell and Robert Scott, A Greek-English Lexicon (Oxford:
Clarendon Press, 1940), online ed., http://perseus.uchicago.edu/cgi-bin/philologic/
getobject.pl?c.76:3:125.LSJ; http://perseus.uchicago.edu/cgi-bin/philologic/getobject.
pl?c.73:5:28.LSJ (accessed 20 July 2017).
124
Houben, Roger II, 35–36.
125
Errico Cuozzo, Catalogus Baronum: Commentario (Rome: Tipografia del Senato,
1966), 207.
126
Ferrante della Marra, Discorsi delle famiglie estinte, forastiere o non comprese ne’ Seggi
di Napoli, imparentate colla Casa della Marra (Naples, 1641).
127
Evelyn Jamison, “Notes on S. Maria della Strada at Matrice, Its History and Sculp-
ture,” Papers of the British School at Rome 14 (1938): 32–97, 39–40.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   75

that the Marchisii were direct descendants of Tancred, who he was con-
vinced had returned from the crusades, and of the princess Cecile; while
this is of course wholly unfounded, his assertion that “the times fitting, it
must be true” raises an important issue.128
These adult landholders in the 1150s could doubtless have been
the children of Odo who had been still underage in 1127. Molise was far
from Calabria, of course; but Roger II had undertaken a radical restruc-
turing of the landholding classes after his violent coming to power in the
1130s, and as the marchioness Sichelgaita had clearly been faithful to
him, it is easy to imagine her sons fighting for him and being rewarded for
it.129 As remarked above, the patronymic “marchisius” is exceedingly rare
in Southern Italy; the very fact that della Marra could immediately point
to Tancred, son of the Marquis, as a match tells us that very few bore it.
While we cannot be wholly sure that Hugh and Johannes were Tancred’s
brothers, the matching of time, place, and likelihood makes it at least a
proposition to entertain carefully.
We can ascribe to Tancred one brother, William, with a degree of
certainty; another, Robert, with plausibility; I suggest we should at least
entertain the possibility that Rogerius Marchisi was another brother of his.
While I feel that, given his presentation in the sources, William was almost
certainly younger than Tancred, and the same must be true of Robert (if
he indeed existed), we cannot conclusively know whether Tancred was
Odo’s eldest. The Hautevilles’ patterns of naming and descendance are
erratic to the point of being non-existent. If Odo was indeed Northern
Italian, he is striking for having given his two (possibly three) known sons
very Norman names: Tancred, William, Robert. All three names occur
several times within the Hauteville family tree. The Hautevilles in general
seemed indifferent to giving their eldest their own name. Roger I, indeed,
only named his very last son Roger.130
Succession on the Hauteville conquest edge could be haphazard:
people could and did die young. Roger I had his bastard son as heir appar-

128
Della Marra, Discorsi, 224–30.
129
For a discussion of this restructuring following a series of baronial revolts against
him, see Hervin Fernández-Aceves, “The Re-Arrangement of the Nobility Under the
Hauteville Monarchy: The Creation of the South Italian Counties,” Ex Historia 8 (2016):
58–90.
130
For a bird’s-eye view of the apparent lack of Hauteville naming patterns, see genea-
logical tables in Loud, Age of Robert Guiscard, 299–300, and Houben, Roger II, table II.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
76   Francesca Petrizzo

ent until he predeceased him; Robert Guiscard had disinherited his eld-
est. 131 We might expect an eldest son to have had incentive enough to
remain in Sicily, a land barely under control; then again Tancred, consist-
ently portrayed as ambitious and acquisitive in the sources, might easily be
pictured choosing to entertain the idea of finding better lands elsewhere.
If Bohemond indeed set out to offer him command of part of his army,
this might have sounded far more attractive than whatever junior role he
might have enjoyed heretofore in his great-uncle Roger’s force.
In the absence of sources, we can say with a degree of certainty that
Tancred is the eldest among the sons of Odo and Emma which we know
of; and that certainly none of them were alive by 1127, when Odo’s sec-
ond wife and widow Sichelgaita held both the title of marchioness and the
care of at least two young children.

Tancred’s Sister
Those writing about Tancred have consistently attributed a sister to him:
Nicholson, Cahen, Asbridge, and Flori all do so.132 This comes from tan-
gential evidence, and from a knotty point in the sources: the parentage of
Roger of Salerno.
Of Roger of Salerno, Tancred’s successor to the regency of Antioch
at his death, we know for sure that he was the son of Richard of the
Principate; and that he and Tancred were therefore related, being cousins
in the third degree.133 However, certain sources suggest that the two were
uncle and nephew, and that Richard had married a sister of Tancred, and
fathered Roger by her. Albert of Aachen, Ibn al-Athir, and the Anonymous
Syriac Chronicle all report this; another Arabic source, Ibn al-Qalanisi,
also claims that Roger was Tancred’s nephew, albeit by his brother.134 More

131
Malaterra, De rebus gestis, 4.18, 97.
132
Nicholson, Tancred, 225; Cahen, La Syrie du Nord, 545; Asbridge, The Creation of
the Principality, 165; Flori, Bohémond, 291.
133
Fulcheri Carnotensis Historia, 2.49, 570.
134
Albert of Aachen’s History, 12.9, 837; The First and Second Crusades from an Anony-
mous Syriac Chronicle, trans. A. S. Tritton and H. A. R. Gibb, Journal of the Royal Asiatic
Society of Great Britain and Ireland 1 (1933): 69–101, 85; The Chronicle of Ibn al-Athir for
the Crusading Period from al-Kamil fi’l-ta’rik, trans. D. S. Richards (Aldershot: Ashgate,
2006–2008), 287; Ibn al-Qalanisi, The Damascus Chronicle of the Crusades, trans. H. A. R.
Gibb (London: Luzac, 1932), 132.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   77

circumspectly, Fulcher of Chartres refers to Roger as “Tancredi cognatus,”

Tancred’s kinsman, and Orderic Vitalis, writing without the fanciful exag-
geration that we saw in his reference to Sichelgaita above, refers to him as
“prefatorum principum consobrinus,” that is, a cousin of the aforemen-
tioned princes, Bohemond and Tancred, which Roger would have been
anyway.135 However, crucial confirmation of his status as Tancred’s nephew
seems to come from Roger himself: in a charter issued to the monastery
of Josaphat, in Syria, he refers to Tancred as his “avunculus,” which usually
has the meaning of “maternal uncle.”136
The evidence, however, is not as clear-cut as it seems, and once more
collateral circumstances help us contextualize it. First of all, the charter:
avunculus can mean uncle, but also cousin. 137 What is more, Tancred’s
sister would have been Richard of the Principate’s second cousin, and in
order to marry her he would have needed a significant dispensation. It is
on these grounds that Edgington takes issue with his stated parentage, and
so does Jamison.138 Whilst the Hautevilles, by the 1080s de iure et de facto
rulers of Southern Italy, could easily have obtained it, the greatest objec-
tion to the possibility of such a marriage is its sheer lack of reason.
Roger of Salerno succeeded Tancred as ruler of Antioch in 1112.
Already a trusted warrior, he must have been at least in his twenties, as
nobody remarks on his youth as they had done with Tancred. He must
therefore have been born in Southern Italy, at least a few years before the
crusade. In Southern Italy Roger was born into the family of the counts
of Principato, by the 1080s one of the most faithful offshoots of the
Hautevilles, who spent their time endowing churches with their ducal
overlords rather than causing trouble.139 Richard of the Principate’s loyalty

135
Fulcheri Carnotensis Historia, 2.47, 563; Orderic Vitalis, Historia Ecclesiastica,
11.25, 104; Du Cange, Glossarium, http://ducange.enc.sorbonne.fr/consobrinus (accessed
23 July 2017). Admittedly, cognatus may also bear the meaning of nephew, among others
including “brother-in-law,” which here would not make sense (Niermeyer and van de Kieft,
Lexicon, 257). I believe especially in light of Fulcher’s mistrust of Roger, discussed below,
that he probably meant it as a generic “relative.”
136
Chartes de Terre Sante provenant de l’abbaye de N.D. de Josaphat, ed. H. F. Delaborde
(Paris, 1880), no. 4, 27.
137
Niermeyer and van de Kieft, Lexicon, 75.
138
Jamison, “Some Notes,” 198; Edgington, Albert of Aachen’s History, 837n19.
139
Graham A. Loud, The Latin Church in Norman Italy (Cambridge: Cambridge
University Press, 2011), 93–98; for a Principato family tree see Graham A. Loud, “The
Abbey of Cava, Its Property and Benefactors in the Norman Era,” in Anglo-Norman Studies

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
78   Francesca Petrizzo

to Bohemond or Tancred was never in doubt; there would have been no

reason to bring him closer with such an alliance, and excepting the suspi-
ciously timely discovery of Guiscard’s kinship with Alberada, none of the
Hautevilles seem to have married close kin.
What is more, Odo and Emma’s children clearly carried some
cachet: if their son William had been easily accepted into the retinue of
a prince of France, their daughter could be expected to make a good mar-
riage, and a wealthy Sicilian landowner like Odo could have endowed a
daughter handsomely and made her an attractive prospect. An ecclesiasti-
cally dubious marriage to an already very close and faithful member of the
family would have neither reason nor precedent. Short of a love match, I
cannot see a reason why such a marriage would have taken place.
What I believe we are looking at here is a case of dubious and mis-
taken kinship, one which the Hautevilles themselves may have lost no
time in clarifying. As I have shown above, Bohemond’s link with Tancred
as his uncle was well known, at least among crusader sources; on its basis,
Tancred’s claim to Antioch was never doubted by anyone. Asbridge points
out that the chronicles are keen to emphasize that Roger did have some
claim to the regency of Antioch, while Fulcher portrays him outright
as having usurped Bohemond II’s rule.140 As the highest-ranking, most
efficient Hauteville left in the East, Roger seems to have been associ-
ated by Tancred with his power and designated as heir in the pragmatic,
contextual mechanism of horizontal kinship and inheritance which the
Hautevilles had been practicing for fifty years; to an outsider, however, the
claim might easily seem dubious.141

IX. Proceedings of the Battle Conference 1986, ed. R. A. Brown (Woodbridge: Boydell and
Brewer, 1987), 143–77, at 158.
140
Asbridge, The Creation of the Principality, 139; Fulcheri Carnotensis Historia, 3.3,
622–24.
141
The first three successions of the Hautevilles in Southern Italy were from brother
to brother: William Iron-arm to Drogo, Drogo to Humphrey, and Humphrey’s infant son
Abelard was usurped by Guiscard, who was the first to successfully ensure vertical succession
thanks to the support of Roger of Sicily for his chosen heir Roger Borsa. Just like Bohemond
had taken Tancred as his second, so had Roger done with his nephew Serlo. The phenom-
enon was by no means confined to the Hautevilles, and Cuozzo remarked on the Normans
in Italy’s “concezione orizzontale della famiglia,” a horizontal conception of the family with
succession between peers given the scarcity of available heirs, after examining the twelfth-
century Catalogus Baronum, a list of the fief-holders in the kingdom of Sicily (Cuozzo, La
cavalleria nel Regno normanno, 198). For a more thorough discussion of these issues, see

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   79

With the model of Bohemond and Tancred’s relationship before

them, crusader and Arabic sources alike could easily assume the same kind
of relationship between Tancred and Roger, or outright declare them sim-
ply kinsmen; Roger, anxious to bolster his own claim, would probably
have an incentive in choosing wilfully obscure or ambiguous language in
describing their relationship. Once we have considered all this, I believe
we can discount the existence of Tancred’s sister, and the possibility that
he and Roger of Salerno were both cousins, and uncle and nephew. If
Emma and Odo had daughters besides sons, we seem to have no reliable
surviving trace of them.

Tancred’s Wife: Cecile of France

I have shown throughout this article how discussion of Tancred’s family
tree often hinges on discussion of the larger Hauteville dynastic policies,
and how in many ways his particular genealogy is a very good case study
for the intersection of several trends in the larger Hauteville kin dynam-
ics. If indeed Tancred’s parentage and origin seem to be a link between
Guiscard’s first Norman support network and the establishment of Roger
I’s Sicilian county, and in his life and with his brothers he embodied the
family bid for the crusader East, after Bohemond he represents the second
male Hauteville to achieve a royal marriage.
The Hautevilles had been punching above their weight in marriage
terms for a long time: Guiscard married a princess of Salerno, and one
of his daughters was set to become empress of Byzantium. Daughters of
Roger I became queens of Hungary and Italy; and Roger Borsa had married
a king’s widow, Adela of Flanders, previously queen of Denmark.142 But
Bohemond was the first to marry the daughter of a king; and when he did
so, he arranged a similar marriage for Tancred, albeit a controversial one.
In 1105 Bohemond departed Antioch, stripping it of its resources,
and headed for Europe to raise a crusade against the Byzantines, leaving
Tancred in charge as regent.143 In France he received a warm, enthusiastic

Francesca Petrizzo, “Band of Brothers: Kin Dynamics of the Hautevilles and Other Normans
in Southern Italy and Syria, c. 1030–c. 1140” (Ph.D. dissertation, University of Leeds, 2018).
142
Romuald of Salerno, Chronicon, ed. C. A. Garufi, Rerum Italicarum Scriptores, 127
(Città di Castello, 1914), 200.
143
Tancredus, 128–29. The exact nature of Tancred’s rule after Bohemond’s departure is a
compelx matter which needs further exploration; the term is here used pragmatically to indi-

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
80   Francesca Petrizzo

welcome, with the popularity of his plans and wealth of his prospects leading
him to a marriage with the king’s daughter in 1106.144 Constance, daughter
of Phillip I and his first wife Bertha of Holland, was then in her twenties,
divorced from Hugh, count of Troyes and Champagne.145 At the same time,
Bohemond arranged for Constance’s half-sister Cecile to marry Tancred, and
she was sent East at some point after this.146 Two things are remarkable about
Cecile: her parentage, and the age she was when she was married to Tancred.
Cecile was the daughter of Phillip’s second marriage, that to his
mistress, Bertrade of Montfort.147 Having failed to obtain an annulment,
when Phillip nonetheless insisted on putting his first wife aside, and
then refused to repudiate Bertrade despite Urban II’s attempt to medi-
ate, he was excommunicated.148 The scandalous story was well known, and
William of Tyre also reported it in more delicate terms.149 Daughter of a
marriage not sanctioned by the church, Cecile was, indeed, strictly speak-
ing illegitimate, but the offspring of the king of France and his beloved
mistress, now queen, would have been no wife to scoff at, and might have
conferred a considerable cachet on her children by Tancred. But when she
married Tancred Cecile was probably not of childbearing age at all.
Cecile must have been born after 1092; Fliche, still an authority on
Phillip, states that we do not know exactly when his children by Bertrade

cate Bohemond left Tancred in charge, but remained alive and known as prince of Antioch.
144
Andrew W. Lewis, Royal Succession in Capetian France: Studies on Familial Order
and the State (Cambridge, MA: Harvard University Press, 1981), 52; Russo, Boemondo,
161–64.
145
Michel Bur, La formation du comté de Champagne, v.950–v.1150 (Nancy: Univer-
sité de Nancy II, 1977), 274, 486–87.
146
Constance and Bohemond having married in March, Nicholson suggests that
Cecile may have arrived in Antioch in the autumn of the same year (Nicholson, Tancred,
163); more conservatively, Rey suggests it is more likely the marriage proper took place at
some point in 1107 (Rey, “Histoire des princes d’Antioche,” 337). As we will see below, I do
not think the date of the marriage in itself is especially significant.
147
Historia regum Francorum monasterii S. Dionysii, ed. Georg Waitz, MGH SS,
9:395–406, at 405.
148
Christof Rolker, Canon Law and the Letters of Ivo of Chartres (Cambridge: Cam-
bridge University Press, 2010), 234–38; Robert Somerville, Pope Urban II’s Council of Pia-
cenza (Oxford: Oxford University Press, 2011), 118. The issue is also discussed by George
Duby, The Knight, the Lady and the Priest: The Making of Modern Marriage in Medieval
France, trans. Barbara Bray (Chicago: University of Chicago Press, 1994), cf. 3–19.
149
William of Tyre, Chronicon, 11.1, 495.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   81

were born.150 Nolan believes that in 1093–1094 the couple still had no
sons, which of course tells us nothing about Cecile. 151 Putting her birth
date at 1093 onwards, the princess may have still been a child when she was
sent East. As consummation required puberty, her marriage to Tancred
was probably the kind of solemn, official betrothal a delayed consumma-
tion would have turned into marriage, and the bride to be was sent East
to ensure the promise could not easily be broken.152 The crusader sources
refer to her simply as Tancred’s wife, without mention of either her age or
any doubts as to the status of the marriage.153
As Tancred died in 1112, Cecile may well have been beyond puberty
by then, but it is hardly surprising the two had no issue, as the possibility
remains that the marriage may not have been consummated at all. Cecile,
however, never went back to France: in 1115 we find her married to Pons,
lord of Tripoli.154 William of Tyre has it that Tancred arranged for this
marriage on his deathbed, endowing the bride; Albert of Aachen says that
this was done “on the king’s [Baldwin I’s] advice,” and the delay between
Tancred’s death and Cecile’s remarriage may cast doubts on his having
arranged it.155 Cahen suggests that the idea of Tancred providing for his
widow is but a gallant fiction on William’s part.156
A more nuanced approach, however, is advisable. Cecile represented
a formidable dynastic link with one of the great monarchies of Europe, and
she was therefore an asset. Tancred died of illness, which suggests at least
the chance that he was forewarned and had time to arrange for both the
principality and his wife. As discussed above, Tancred had a proven track
record of being both enterprising and active. Having to relinquish both
his principality and his prestigious wife, it would have been in character
with the rest of his rule to provide for both, as he did by clearly designating

150
Augustin Fliche, Le règne de Philippe Ier, roi de France (1060–1108) (Paris: Société
Française d’Imprimerie et de Librairie, 1912), 90.
151
Kathleen Nolan, Queens in Stone and Silver: The Creation of a Visual Imagery of
Queenship in Capetian France (New York: Palgrave Macmillan, 2017), 20.
152
David L. d’Avray, Medieval Marriage: Symbolism and Society (Oxford: Oxford Uni-
versity Press, 2005), 182. I am grateful to Professor d’Avray for his guidance on this issue.
153
William of Tyre, Chronicon, “uxore sua,” 11.18, 522; Albert of Aachen’s History,
12.19, 854.
154
Albert of Aachen’s History, 12.19, 854.
155
William of Tyre, Chronicon, 11.18, 522–23; Albert of Aachen’s History, 12.19, 854.
156
Cahen, La Syrie du Nord, 256n20.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
82   Francesca Petrizzo

Genealogy of Tancred

Roger of Salerno as his heir.157 At the same time, once Tancred had passed
King Baldwin I became the undisputed dominating force in the crusader
polities; it is possible that he only gave his consent for the marriage to
take place after a few years, or that communications were exchanged with
France concerning the now widowed princess of which we have no trace.
Cecile’s marriage to Pons was both long and fruitful: they had three
children, and we hear of her in 1139, having undertaken an influential
role during her husband’s reign.158 While therefore it appears that Cecile’s
life in Outremer did eventually develop positively for her, her union with
Tancred remains a premature, and unfulfilled dynastic ambition: the first,
sterile shoot of the Hauteville ambition to marry into royal blood in the
Latin East.

Conclusion: Tancred, A New Family Tree

This article has sought to review the existing information about Tancred
of Antioch’s family tree, while at the same time contextualizing him
within the larger kin dynamics and dynastic trends of the Hautevilles.
Context has proven crucial in solving several conflicting points in the

157
Albert of Aachen’s History, 12.12, 840.
158
Cartulaire de l’église de Saint-Sépulchre de Jerusalem, ed. E. de Rozière (Paris, 1849),
“Saint-Sépulchre de Jerusalem”, 92, 183.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
The Ancestry and Kinship of Tancred, Prince Regent of Antioch   83

sources, framing Tancred as a fundamental intersection in the expansion

of the Hauteville clan. A grandson of Guiscard and his first Norman wife,
a son of one of the Northern Italian colonists of Sicily, Tancred probably
grew up on the frontier of Hauteville expansion in Southern Italy, where
he learnt Arabic, and acquired both military skills and experience dealing
with Muslim communities. Having been recruited by his uncle Bohemond,
Tancred went on crusade with one, possibly two of his younger brothers,
and with his cousin Richard of the Principate. As part of this extensive kin
group, Tancred established himself and eventually became prince regent
of Antioch. When Bohemond capitalized on his glory as crusader to
achieve a royal marriage for himself, Tancred also married into the French
royal family; at his death, he left power in the hands of his cousin Roger
of Salerno.
While his untimely death meant Tancred’s career was cut short, his
life and its connections are a prime case study for both the evolution of
Hauteville kin patterns of expansion and the influence of Southern Italian
Normans in the crusader states.

This content downloaded from

87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms
 This content downloaded from
87.13.216.192 on Fri, 24 Jan 2025 22:30:38 UTC
All use subject to https://about.jstor.org/terms