Rollo
"of Normandy" Rollo
"of Normandy"Although he is often referred to as the first duke of Normandy, that title is an anachronism. The title "count of Rouen" which is sometimes assigned to him [e.g., in the genealogical table in van Houts (2000), 292] probably accurately reflects his area of authority, but he has not been documented with the title of count. The term princeps Nortmannorum has the double advantage of being justified by the contemporary authority of Flodoard's annals for both Rollo [s.a. 925: MGH SS 3, 375, van Houts (2000), 44] and his son William [s.a. 933, 940: MGH SS 3, 381, 386, van Houts (2000), 45-6], and of also being similar to the anachronistic title by which they were both later called, although this description should not be confused as a hereditary title, as these words were also used, for example, to describe Ragenold, the leader of a different group of Vikings on the river Loire [Flodoard's Annals, s.a. 923: MGH SS 3, 372, van Houts (2000), 42]. See Helmerichs (1997) for a detailed study of the titles by which Rollo and his successors are described in the sources. The subject of many later legends as the result of the writings of Dudo and others, the contemporary sources for Rollo are meager. Probably about 911 [see Douglas (1942), 426-31], king Charles the Simple of France ceded a district around the city of Rouen to Rollo, which eventually evolved into the duchy of Normandy. He is said to have been baptized in 912, assuming the Christian name Robert [Dudo ii, 30 (p. 50)]. He was still living in 928, when he was holding Eudes, son of Heribert of Vermandois, as a captive [Flodoard's Annals, s.a. 928: MGH SS 3, 378, van Houts (2000), 45], and was probably dead by 933, when his son William was mentioned as leading the Normans [Flodoard's Annals, s.a. 933: MGH SS 3, 381, van Houts (2000), 45].
Date of Birth: Unknown.
Place of
Birth: Unknown.
Date of Death: 928×933.
[see above]
Place of Death: Unknown.
Father: Unknown (see Commentary)
Mother: Unknown (see Commentary)
Partner/Spouse:
"Poppa" (See her page for more details)
See Commentary for supposed other wife.
Children:
[Dudo ii, 16 (p. 39) makes Poppa the mother of William, but does
not give the mother of Gerloc/Adele. Guillaume de Jumièges [GND
ii, 6 (v. 1, pp. 64-5)] makes Poppa the mother of both Guillaume
and Gerloc.]
Guillaume (William) I
"Longsword" of Normandy, d.
942.
Gerloc or Adele, m. Guillaume
(William) Tête d'Étoupe, count of Poitou and duke of
Aquitaine.
[Dudo iii, 47 (pp. 69-70, which mentions the marriage but does
not give her name), p. 201 (note 256, and sources cited therein,
for Adele as the Christian name of Gerloc); GND ii, 13 (v. 1, pp.
68-9)]
See Commentary for supposed additional child.
The origin of Rollo is controversial, and requires a detailed discussion. The vast majority of secondary sources would give his parentage as follows, following the Icelandic sources discussed in more detail below:
Supposed father (doubtful): Rognvaldr, jarl of Møre.
Supposed mother (doubtful): Ragnhildr or Hildr.
More recently, David Crouch [Crouch (2002), 297-300] has accepted the account of Richer of Reims, who would make Rollo a son of a certain Catillus:
Supposed father (doubtful): Ketill (Catillus).
Although Rognvaldr and Catillus are the two obvious fathers explicitly named for Rollo in (non-contemporary) primary sources, it is not necessary that one of these two candidates must be accepted, and other possibilities (such as "unknown") must be given serious consideration if that is where the evidence leads. As can be seen from the following brief notices, the various primary sources offer very contradictory information about Rollo's origin.
The earliest author to attribute an explicit origin to Rollo was Richer of Reims, writing between 996 and 998, who called Rollo the son of another Viking invader of France named Catillus (presumably representing the Norse name Ketill) ["duce Rollone filio Catilli" Richer, Historia, i, 28 (vol. 1, p. 62)]. Since Richer is not generally reliable for the relevant time period, and Catillus appears to be a legendary individual with no clear identity in the contemporary sources, this account has generally been discredited, probably correctly [see Douglas (1942), 420-1].
According to Dudo of St. Quentin (writing early 11th century), author of the earliest history of the Normans, Rollo had a younger brother named Gurim, presumed to be the familiar name Gorm. Dudo states that Rollo and Gurim were sons of a man who held many lands in "Dacia" (Dudo's word for Denmark, following other authors), and that after the death of the (unnamed) father of Rollo and Gurim, the king of Dacia fought against the sons, killing Gurim and driving Rollo out [Dudo ii, 2-4 (pp. 26-7)]. Dudo later refers to duke Richard I as being related to a "king of Dacia" named Haigrold [Dudo iv, 84-88 (pp. 114-20 passim)], who must have been the Viking raider of France of that name [Flodoard's Annals, s.a. 945: MGH SS 3, 392, van Houts (2000), 51], and not king Harald "Bluetooth" of Denmark. Note that even if Rollo had such a brother named Gurim, he cannot be the famous Gorm "the Old" of Denmark, who survived Rollo by many years.
The earliest dateable sources which attribute a Norwegian origin to Rollo are from the late eleventh and early twelfth centuries, principally Geoffrey Malaterra [writing ca. 1090, see van Houts (2000), 52] and William of Malmesbury [early 12th century, see WM ii, 5 (p. 125)]. A Norwegian origin for Rollo is also stated in the chronicle Chronicon de gestis Normannorum in Francia [MGH SS 1, 532-6, at 536]. Although this source used ninth century Frankish annnals as a basis, the entries regarding Rollo ("Rotlo", mistranscribed as "Rodo" in the MGH edition) were probably introduced at about the time of the manuscript's compilation in the twelfth century. See Helmerichs (2002) for a discussion of this source.
As is well known, the Orkneyinga Saga (late twelfth century) [OrkS 4 (pp. 29-30)], followed by other Icelandic sources (such as the well known Heimskringla and Landnámabók), gives Rollo the name Hrólfr, and make him a son of Rognvaldr, jarl of Møre, and brother of (among others) jarl Torf-Einarr of the Orkneys [OI 1: 187]. Earlier sources, such as Ari's Íslendingabók (early to middle 12th century), mention Rognvald of Møre and his son Hrollaugr who settled in Iceland, but not the supposed connection to the dukes of Normandy [Ari 49, 61]. A poem allegedly written by Einar mentions his brothers, including a Hrólfr, but does not connect Hrólfr to Normandy, and does not name a Gorm among the brothers. (See the page on Rognvaldr for more on this poem.)
Historia Gruffud vab Kenan (ca. 1250), apparently a Welsh translation and/or revision of an earlier Latin life of Gruffudd ap Cynan, gives Haraldr Hárfagri of Norway ("Harald Harfagyr") a brother named Rodulf (i.e., the Latin form of Hrólfr) who is called the founder of Normandy [HGK, 3-4]. However, this is evidently a corrupt version of the Scandinavian version, and the suggestion that Rollo was a brother of Haraldr Hárfagri need not be given any credence.
The most prominent argument of the case for accepting the Scandinavian account that Rollo was the same person as Hrólfr, son of Rognvaldr of Møre, was given by D. C. Douglas [Douglas (1942), 419-23], and this view has been widely followed by others, including standard genealogical reference sources like ES (generally without any reexamination of the underlying evidence). For explicit arguments against the orthodox view identifying Rollo as a son of Rognvaldr of Møre, see Starcke (1946), 222-7, for an early example, and more recently, Helmerichs (2002). For the last several years, in informal discussions on the internet newsgroup soc.genealogy.medieval, I have argued against the acceptance of the Icelandic account of Rollo's origin, and much of the material presented here is a synthesis of those earlier internet postings.
Most of the argument of Douglas consists of accepting the tale of the sagas and rejecting evidence from the Norman sources which contradict the saga version, while explaining away the problems (on which more below). The evidence which Douglas puts forward as "a powerful, if not a conclusive, argument in favor of the identity of Rollo with Ganger-Rolf" concerns a passage in Landnámabók that refers to a daughter of Gongu-Hrólfr:
"... Annarr son Óttars vas Helge; hann herjaðe á Skottland, ok feck þar at herfange Niðbiorgo, dóttor Beolans konungs ok Caðlínar, dóttor Gongo-Hrólfs" (Another son of Óttarr was Helge. He harried in Scotland, and won there as his booty Niðbjorg, daughter of king Beolan and Caðlín, daughter of Gongu-Hrólfr.) [OI 1: 66-7]
This passage, which Douglas attributed to "Ari the Learned" (who may or may not have been the author), is then compared with a passage from the nearly contemporary Plaintsong of Rollo's son William "Longsword" which was written soon after William's death (see also Robert Helmerichs's Planctus website):
"Hic in orbe transmarino natus patre
in errore paganorum permanente
matre quoque consignata alma fide
sacra fuit lotus unda"
(Born overseas from a father who stuck to the pagan error and
from a mother who was devoted to the sweet religion, he was
blessed with the holy chrism.)
[Douglas (1942), 422 (Latin); van Houts (2000), 41 (English
translation)]
After explaining that the two stories are consistent with one another, Douglas then states that "[t]he suggestion of the Landnámabók is thus confirmed by an epic poem composed in Gaul in the tenth century." While it is true that if one ignores all other evidence, then the two accounts as they stand are consistent with each other and with the claim that Rollo and Gongu-Hrólfr were the same man, it is surely a gross overstatement to claim that the Plaintsong "confirms" the other account, for there is not a single statement in the passage from Landnámabók that is confirmed by the Plaintsong. This is a clear case of circular reasoning, for without first assuming that Rollo and Gongu-Hrólfr were the same man, there is no evidence that the two passages have any relation whatsoever. Douglas's case is further undermined by the fact that another source [Laxdœla Saga chapter 32, see OI 1: 246] makes Niðbjorg's mother Caðlín a daughter of Gongu-Hrólfr, son of Oxna-Þórir, directly contradicting the thesis that Caðlín was supposedly a granddaughter of Rognvaldr of Møre. Yet, Douglas apparently regarded this as the strongest part of his argument.
There are three main strands of evidence (somewhat related to each other) against the identification of Rollo with Hrólfr son of Rognvaldr:
1. The discrepancies between the Norman and Icelandic
sources.
Among other contradictions, the Norman sources give Rollo a
brother named Gurim, while the Icelandic sources give Hrólfr
several brothers, none of them named Gormr (the presumed
Old-Norse form for Gurim). Although both of the sources have
their problems, earlier native sources would seem to have a
higher priority than later foreign sources. While many elements
of the Dudo's account are clearly legendary, there appears to be
no clear motive on the part of Dudo (writing less than a century
after Rollo's death) to invent a younger brother for Rollo who is
then immediately killed off.
2. The general unreliability of Norse sources for the
early tenth century.
For the period under consideration, i.e., the early ninth
century, the sagas have a poor record for reliability, even for
Scandinavian history. For example, consider the following words
of Peter Sawyer (written with regard to a different matter, but
true in general), a well known expert on early Viking history:
"... These sagas cannot, however, be accepted as reliable
sources for the tenth century. The only trustworthy evidence for
the tenth century in those sagas are the contemporary verses
around which the saga writers wove their tales." [Sawyer
(1995), 42] None of these verses confirm the identity of Rollo
and Hrólfr. The suspicion is made even larger by the fact that
the Icelandic sources show no knowledge of Norman history other
than facts which would have been well known throughout Europe at
that time (such as William the Conqueror's descent from Rollo).
3. Rollo and Hrólfr appear to be different names.
The natural Latinization of the name Hrólfr would be Radulfus or
Rodulfus. Yet, the Frankish and Norman sources usually refer to
the founder of Normandy as Rollo. Since these Frankish sources
also include numerous individuals named Rodulfus, and
consistently distinguish the two names, it appears that the names
were regarded as different. Douglas explained this by suggesting
a hypothetical hypochoristic form "Hrolle" of the name
"Hroðwulf" as the basis for the name Rollo, and
provides a single charter in which Rollo is referred to as
"Rolphus" as evidence that the names were the same,
acknowledging, however, that the charter itself was "not
above suspicion." If the names were really regarded as the
same, it would be expected that more convincing evidence to this
effect could be offered. In fact, the Icelandic sources, in
identifying Rollo with Hrólfr, son of Rognvaldr, provide Hrólfr
with a brother named Hrollaugr, which seems like a more plausible
Norse form for the name Rollo than Hrólfr. This Hrollaugr, said
to be an early settler of Iceland, cannot be identified with
Rollo of Normandy, so that the Icelandic sources are providing
the founder of Normandy with a brother whose name is evidenctly
Rollo, further illustrating the problems with the Icelandic
account. Robert Helmerichs wrote: "Is it possible that in
seeking an origin for Rollo/Hrollaugr, the bards found the
Hrollaugr of the Heimskringla, but upon realizing that
it was historically impossible for him to be our Rollo, they
invented a brother for him, based upon the Göngu-Hrólfssaga and
named Hrólfr?" [Helmerichs (2002), note 20] In postings to
the newsgroups soc.genealogy.medieval and soc.history.medieval, I
have suggested the following (obviously conjectural) scenario. As
already mentioned above, the annalist Flodoard mentioned a Viking
named Ragenold, called princeps Nortmannorum like Rollo.
Although no direct connection between Ragneold and Rollo was
suggested by Flodoard, they did occur in nearby passages. To an
Icelandic scholar who happened to get a hold of a copy of
Flodoard's annals (or something else based on them), the names
Ragenold and Rollo occuring so close together might very well
have reminded him about the Rognvaldr and his son Hrollaugr (the
early Icelandic settler) in the Icelandic sources, and this could
have been the beginning of a process of fabrication (perhaps
"brilliant deduction" from the point of view of the
saga writer himself) in which Rollo was identified as a son of
Rognvald, and had his name changed to Hrólfr. Previously, Niels
Lukman had gone even further, and had suggested that Rognvaldr of
Møre may have himself been fabricated,
with the Ragenold of Flodoard's annals as the historical
prototype [Lukman (1976), 42-3]. (While I find Lukman's
suggestion unconvincing, I mention this article because it was
what first led me to consider the above scenario.) The
identification of Rollo with the legendary Göngu-Hrólfr of the
Icelandic sagas provides further reason for suspicion. Two
alleged fathers of this legendary figure, Rognvaldr and
Oxna-Þórir, have already been mentioned above. However, a third
father is provided by the Icelandic saga Göngu-Hrólfs Saga,
which makes Göngu-Hrólfr (who becomes king of Russia in the
end, and has no Norman connection in the saga) one of the sons of
a certain Sturlaugr, king of Hringeríki in Norway, and relates
the well known story that he was so heavy that no horse could
carry him all day [Göngu-Hrólfs Saga, ch. 4 (p. 36)]. It has
been noted [e.g., Christiansen, in his edition of Dudo, pp.
197-8, n. 223] that Dudo states that Rollo was unable to ride in
his old age [Dudo ii, 34 (p. 54)], and it may be that this
passage of Dudo facilitated the identification of Rollo with the
legend of Göngu-Hrólfr (and the change of name from Rollo to
Hrólfr) by the Icelandic saga-writers.
With such contradictory evidence, it is difficult to come to any definitive conclusion, but the above reasons indicate why the view that Rollo was a son of Rognvaldr is open to grave doubt. The saying that "nature abhors a vaccuum" is particularly apt in cases like this, and many genealogists are extremely reluctant to leave a parentage as "unknown" when such an attractive candidate is available to "fill in the blank." Nevertheless, given the the serious problems with the Icelandic sources mentioned above, it is sensible to regard Rollo's parentage as unknown.
Supposed second wife (probably mythical):
Gisla, said to be daughter of Charles the Simple, king of France [Dudo, 46-7, 53]. She is unknown in the Frankish sources. The fact that Charles the Simple's kinsman Charles the Fat had a daughter also named Gisla who married a Viking (Godefridus) in the ninth century has led to the natural suspicion that this Gisla is an invention based on the earlier woman of the name. If she existed at all, there is no reason to believe that she was a mother of any of Rollo's children.
Supposed additional child (doubtful):
Caðlin (Kathleen), said by
Norse sources to have married a certain king Beolan,
who is otherwise unidentified. As discussed above, the evidence
for her is less than satisfactory.
Ari = Halldór Hermannsson, ed. & trans., The Book of Icelanders (Íslendingabók) by Ari Thorgilsson (Islandica, vol. 20, Ithaca, 1930).
Crouch (2002) = David Crouch, The Normans: the history of a dynasty (London, 2002).
Douglas (1942) = D. C. Douglas, "Rollo of Normandy", English Historical Review 57 (1942), 417-36.
Dudo = Eric Christiansen, ed. & trans., Dudo of St. Quentin, History of the Normans (The Boydell Press, Woodbridge, Suffolk, 1998). Citation is by book and chapter of Dudo's work, with the page number in parentheses.
ES = Detlev Schwennicke, Europäische Stammtafeln (neue Folge), (Marburg, 1980-present).
Flodoard's Annals = See MGH SS 3, 363-408 (Latin), and van Houts (2000), 42-51 (English translation of excerpts relating to the Normans).
Helmerichs (1997) = Robert Helmerichs, "Princeps, Comes, Dux Normannorum: Early Rollonid Designators and their Significance", Haskins Society Journal 9 (1997), 57-77.
Helmerichs (2002) = "Rollo as Historical Figure", http://www.mm.com/user/rob/Rollo/HistoricalRollo.html.
GND = Guillaume de Jumièges, Gesta Normannorum Ducum, as edited in Elisabeth van Houts, ed. & trans., The Gesta Normannorum Ducum of William of Jumièges, Orderic Vitalis and Robert of Torigni, 2 vols., (Oxford, 1992). Citation is by book and chapter of Guillaume's work, with the volume and page number of the edition by van Houts in parentheses.
Göngu-Hrólfs Saga = Hermann Pálsson & Paul Edwards, trans., Göngu-Hrolfs Saga (Toronto, 1980).
HGK = D. Simon Evans, ed., Historia Gruffudd vab Kenan (Caerdydd, 1977).
Lukman (1976) = Niels Lukman, "Ragnarr loðbrók, Sigifrid, and the Saints of Flanders", Mediaeval Scandinavia 9 (1976), 7-50.
OI = Gudbrand Vigfusson and F. York Powell, ed. & trans., Origines Islandicae, 2 vols. (Oxford, 1905).
OrkS = Herman Pálsson & Paul Edwards, ed. & trans., Orkneyinga Saga (London, 1978). Citation is by chapter, with the page number in parentheses.
PL = P. Migne, Patrologiae Cursus Completus, series Latina, 221 vols. (Paris, 1844-1859).
Richer, Historia = G. H. Pertz, ed., & J. Guadet, trans. (French), Richer, Histoire de son temps, 2 vols. (Paris, 1845).
Sawyer (1995) = Peter Sawyer, "The Last Scandinavian Kings of York", Northern History 31 (1995), 39-44.
Settipani = Christian Settipani, La préhistoire des Capétiens 481-987 (Première partie - Mérovingiens, Carolingiens et Robertiens) (Villeneuve d'Ascq, 1993).
Starcke (1946) = Viggo Starcke, Denmark in World History (Philadelphia, 1962, a translation of the Danish edition of 1946 by Frank Noel Stagg, Ingeborg Nixon, and Mrs. Elmer Harp).
van Houts (2000) = Elisabeth van Houts, ed. & trans., The Normans in Europe (Manchester & New York, 2000) [gives English translations of many of the primary sources relevant to early Norman history]
WM = J. A. Giles, ed. & trans., William of Malmesbury's Chronicle of the Kings of England (London, 1889). Citation is by book and chapter of William's work, with the page number in parentheses.
Compiled by Stewart Baldwin. Some of the material on this page (the arguments against Rollo being a son of Rognvaldr in particular) is based on postings by me that were made on numerous occasions to the newsgroup soc.genealogy.medieval in the late 1990's and early 2000's.
First uploaded 23 July 2001.
Significant revision of the earlier version first uploaded 8 February 2004.