Read in this way, the story is quite reasonable and intelligible, and the rest of the Gesta’s account might stand almost intact, except the capture of the duke, which of course is dragged in from the earlier war. The confusion between the Williams of Aquitaine is easily accounted for, and so is that between the Geoffreys of Anjou, especially as all the Geoffreys after Martel occasionally took to themselves his cognomen.
Note D.
THE DESCENDANTS OF HERBERT WAKE-DOG.
Not the least puzzling matter connected with the Cenomannian wars is the genealogy of the sovereign house of Maine. The succession of the counts themselves—Hugh I. (or David), Herbert I. (Wake-dog), Hugh II., Herbert II.—is plain enough, as also that each was the son of his predecessor. But the filiation of the women of the family—Margaret, Gersendis, Paula and Biota—is far from being equally clear.
1. As to Margaret, there is no real doubt. Orderic does once (Duchesne, Hist. Norm. Scriptt., p. 683) call her a daughter of Herbert [II.]; but his own statements in two other places (ib. pp. 487 and 532), as well as Will. Poitiers (ib. p. 190), shew that this is a mere slip. Margaret was clearly a daughter of Hugh II. and sister of Herbert II.
2. As to Biota. Orderic (as above, p. 487) calls her “Hugonis Cenomannensium comitis filiam”; in Will. Poitiers (ib. p. 189) she is “soror Hugonis”; and Mr. Freeman (Norm. Conq., 3d ed., vol. iii. p. 200, and note T, p. 676) adopts the latter version. Biota, then, was a daughter of Herbert Wake-dog and sister of Hugh II. But were Gersendis and Paula her sisters or her nieces?
3. The fullest and most distinct statement of the Cenomannian pedigree is that of Orderic in Duchesne, Hist. Norm. Scriptt., p. 532: “Hugo filius Herberti ... Bertam ... in conjugium accepit; quæ filium nomine Herbertum et tres filias ei peperit. Una earum data est Azsoni Marchiso Liguriæ. Alia nomine Margarita Rodberto filio Guillelmi Ducis Neustriæ desponsata est ... Tertia vero Joanni domino castri quod Flecchia dicitur nupsit.”
With regard to this last marriage, it is to be observed that in the speech which Orderic puts into the mouth of Elias of La Flèche, addressing Hugh of Este (ib. p. 684), he says nothing about his mother at all, but makes him trace his descent from Herbert Wake-dog through his grandmother, whom he calls Herbert’s daughter: “Filia Herberti comitis Lancelino de Balgenceio nupsit, eique ... Joannem meum genitorem peperit.” The name of John’s wife, Paula, comes from another passage of Orderic (ib. p. 768); but he there says nothing about her parentage, merely calling her son Elias “Hugonis Cenomannorum consulis consobrinus.” The houses of Le Mans and La Flèche cannot have intermarried twice in two succeeding generations; one of Orderic’s statements must be wrong; but which, I cannot decide.
The last point is the parentage of Gersendis, the wife of Azzo of Este; and as the whole tone of Elias’s speech (as above) implies that he and her son were related to the counts of Le Mans in the same degree, the solution of this question might almost be held to decide the previous one also. This seems to be Mr. Freeman’s opinion, and he regards Orderic’s statement quoted above as conclusive that Gersendis and Paula were both daughters of Hugh II., and sisters therefore of Margaret and Herbert II., in spite of the biographer of the bishops of Le Mans (Mabillon, Vet. Anal., p. 308), who expressly says that Gersendis was a daughter of Herbert Wake-dog, and the continuator of Will. Jumièges, who says:—“Cenomannenses ... consilium ineunt cum Heliâ filio Joannis de Flecâ ... ut filiam cujusdam comitis Langobardiæ, neptem videlicet Hereberti quondam Cenomannensis comitis ex primogenitâ filiâ, in matrimonium ducat.” Will. Jumièges, l. viii. c. 5 (Duchesne, Hist. Norm. Scriptt., p. 294). This re-appears in R. Diceto (Stubbs, vol. i. pp. 183, 184; Marchegay, Comtes, p. 334) in the following form:—“Helias, filius Johannis de Flecâ, Sibillam, filiam cujusdam comitis Longobardiæ, neptem scilicet Hereberti quondam Cenomannorum comitis, duxit uxorem, et cum eâ comitatum Cenomanniæ suscepit.” But this is certainly wrong; for the first wife of Elias was Matilda of Château-du-Loir, and the second was Agnes of Perche.
What Elias could have had to gain by the marriage thus proposed for him it is impossible to guess, as he himself certainly was quite as nearly related to the counts of Maine as this oddly-described bride could have been. Mr. Freeman (Norm. Conq., 3d ed., vol. iii., note T, p. 676), takes the description as favouring Orderic’s theory, and remarks: “The words could only have been written by one who looked on Gersendis as a sister of Herbert.” “Neptem Hereberti,” then, he interprets, “niece of Herbert [II].” But is it not a much simpler interpretation of the whole phrase—“neptem Hereberti ex primogenitâ filiâ”—to read it “granddaughter of Herbert [I.] through his eldest daughter”? In that case, we should have another witness on the side of the bishops’ biographer.
There is another curious bit of evidence which at first glance seems also to tell in his favour. I do not think that it really proves anything about the matter; but it is worth examining for other reasons. M. d’Arbois de Jubainville (Comtes de Champagne, vol. i. p. 392, note 5), declares it proved on documentary evidence that Stephen-Henry of Blois, the father of our King Stephen, was the son of Theobald III. by his first marriage with Gersendis of Maine. About the marriage itself there is no doubt, nor about the divorce which followed it; and the latter had taken place in 1049 at latest, for Theobald was excommunicated for that very cause by the Council of Reims. Most historians seem however to have supposed that Gersendis was then a mere child, and that the mother of Stephen, as well as of Theobald’s other children, was his second wife, Adela of Valois. M. de Jubainville, in support of his opinion, refers especially to two charters. One is in Gallia Christiana, vol. viii., instr. col. 548. It has no date, and says nothing about Stephen’s mother or his stepmother; I therefore cannot see its bearing on the question. The other is in Bernier, Histoire de Blois, preuves, pp. xiii–xiv. In it Stephen-Henry, in the year 1089, grants certain lands to Pontlevoy “pro animæ meæ et uxoris et Theobaldi patris mei et matris meæ Gandree ... remedio”; and has the grant confirmed “nomine ... Alæ uxoris meæ, Alæ uxoris Thebaudi comitis,” etc. This certainly seems to shew that Adela was not his mother, though it does not necessarily follow that “Gandree” represents Gersendis. If it does, Stephen-Henry must have been born in 1049 at latest, and therefore Gersendis cannot possibly have been a daughter of Hugh II., who was not married till 1040 at the very earliest.