¹ See, e.g. Hammer, Carl I., ‘Family and familia in early-medieval Bavaria’, in Richard, Wall, Jean, Robin and Peter, Laslett, eds., Family forms in historic Europe (Cambridge, 1983) 217–48CrossRefGoogle Scholar, at 243–6 (‘the distribution of household types seems to provide striking confirmation from the early ninth century of early-modern family history: the simple or nuclear family was the overwhelmingly dominant type’); cf. however, Herlihy, David, Medieval households (Cambridge, MA, 1985) 70–2Google Scholar, who finds evidence of pronounced lateral extension on the estates of St Germain-des-Prés in the early ninth century. See, e.g. Degler, Carl N., At odds: women and the family in America from the Revolution to the present (Oxford, 1980) 8–9Google Scholar (‘the modern American family emerged first in the years between the American Revolution and about 1830’). See also the discussion in Goody, Jack, The development of the family and marriage in Europe (Cambridge, 1983) 24–6.CrossRefGoogle Scholar Whatever the strength of the arguments may be, there is, one feels, some ideological tendentiousness not far below the surface of many works that discover ever-earlier declines of collectivities and risings of utility maximising individuals nurtured in simple households. A new romanticism of the simple household appears to have replaced an older romanticism of cooperating collectivities.

² Jochens, Jenny M., ‘En Islande médiévale: A la recherche de la famille nucléaire’, Annales ESC 1 (1985) 95–112Google Scholar at 106.

³ Ibid. 100, 105–7 and see below, n. 82. Assuming a stationary or expanding population, an early age of marriage for both sexes would make it more difficult for new conjugal units to have their own farms and would tend toward the creation of complex households than would a later age of marriage.

⁴ See, generally, Hajnal, J., ‘European marriage patterns in perspective’, in Glass, D. V. and Eversley, D. E. C., eds., Population in history: essays in historical demography (London, 1965).Google Scholar The difficulty of determining marriage age, among other things, even with much better evidence than is available in Iceland, is exemplified in the spirited debate between Zvi Razi and his critics regarding methods of interpretation of manorial court records in English local studies. The possibility of relatively late age for England has been argued by Smith, R. M., ‘Some reflections on the evidence for the origins of the “European marriage pattern’ in England’, in C., Harris, ed., Sociology of the family: new directions for Britain, Sociological Review Monograph, 28 (Keele, 1979) 74–112Google Scholar, and controverted by the evidence in , Razi, Life, marriage, and death in a medieval parish: economy, society, and demography in Halesowen 1270–1400 (Cambridge, 1980).Google Scholar Razi's methods are criticised in Poos, L. R. and Smith, R. M., ‘“Legal windows onto historical populations?” recent research on demography and the Manor Court in medieval England,’ Law and History Review 2 (1984) 128–52.CrossRefGoogle Scholar Response and counter response have followed: see , Razi, ‘The use of manorial-court rolls in demographic analysis: a reconsideration’, Law and History Review 3 (1985) 191–200CrossRefGoogle Scholar; Poos and Smith, ‘“Shades still on the window”: a reply to Zvi Razi,’ ibid. 409–29; and , Razi, ‘The demographic transparency of manorial-court roles,’ Law and History Review 5, (1987) 523–36.CrossRefGoogle Scholar

⁵ See, e.g. ‘the critiques and attempts at resolution’, in Yanagisako, Sylvia J., ‘Family and household: the analysis of domestic groups’, Annual Review of Anthropology 8 (1979) 161–205CrossRefGoogle Scholar; Verdon, Michel, ‘Shaking off the domestic yoke, or the sociological significance of residence’, Comparative Studies in Society and History 22 (1980) 109–32CrossRefGoogle Scholar; Netting, Robert McC., Wilk, Richard R., and Arnould, Eric J., ‘Introduction’, in Netting, Wilk, Arnould, eds., Households: comparative and historical studies of the domestic group (Berkeley, 1984) xiii–xxxviiiGoogle Scholar; Wilk and Netting, ‘Households: changing forms and functions’, ibid. 1–28; Hammel, E. A., ‘On the *** of studying household form and function,’Google Scholaribid. 29–43; Wall, Richard, ‘Introduction’, in Family forms in historic Europe, 1–65Google Scholar; Hajnal, J., ‘Two kinds of pre-industrial household formation systems’Google Scholar, ibid. 65–104; and Laslett, Peter, ‘Family and household as work group and kin group: areas of traditional Europe compared’Google Scholar, ibid. 513–63.

⁶ See, e.g. Hajnal, ‘Two kinds of pre-industrial household formation systems’, 99–100, who requires the assignment of everyone uniquely to one household and also presupposes only one head per household.

⁷ Verdon's formulation in ‘Shaking off the domestic yoke’, and adopted by Laslett in ‘Family and household’, 515, takes coresidence as the defining characteristic of the domestic group. Hajnal, ‘Two kinds of pre-industrial household formation systems’, 99, defines household as the housekeeping or consumption unit and requires eating together or ‘sharing of meals derived from a common stock of food’.

⁸ See Wall, ‘Introduction’, in Family Forms, 35.

⁹ See Peter, Laslett and Richard, Wall, eds., Household and family in past time (Cambridge, 1972) 28–31Google Scholar; and Wall, ‘Introduction’, Family Forms, 19.

¹⁰ ‘Joint’ in Hajnal, ‘Two kinds of pre-industrial household formation systems’, 68–9, 77–8, corresponds for the most part to Laslett's ‘multiple’ in Household and family, 28–31. A ‘multiple’ household requires two or more conjugal units which in most instances will be married couples but might also be a widow and her children. I make no arguments that depend on the distinction being maintained.

¹¹ Bonds based on concubinage, like those founded on marriage, were invoked in gathering support for feud. See, e.g. Njáls saga 98:251–52, Íslenzk fornrit 12 and also Guðmunder saga dýura St. 1: 15:192, where A asks B to support him in a lawsuit because, says the saga, two of B's children were by A's sister. The family sagas are cited by chapter and page number in Íslenzk fornrit (Reykjavik, 1933-), hereafter IF. The chapter divisions in these editions are maintained in most accessible English translations of the sagas. I supply the IF volume number for the first citation of a saga or páttr (short saga). Sagas in the Sturlunga compilation are sigilled by St. before volume, chapter and page in Jðn, Jðhannesson, Magnús, Finnbogason, and Kristján, Eldjám, eds., Sturlunga saga (Reykyavik, 1946).Google Scholar

¹² I am extending the household to include what Laslett and Wall call the houseful; that is, all those residents who are not spouse, child, relative, or servant of the household head (Wall, Family forms, 35; e.g. Berkner, Lutz K., ‘The use and misuse of census data for the historical analysis of family structure’, Journal of Interdisciplinary History 4 (1975) 721–38CrossRefGoogle Scholar; Our information is seldom circumstantial enough to know whether residents who are not children are kin to the head or not. We simply do not know whether servants who could claim some kind of kinship with the householder were more privileged than servants who could not.

¹³ Even the best of sources have their problems. It is well-known that census data might under-represent the significance of joint householding. Berkner, ‘Use and misuse’, 726; also Robert Wheaton, ‘Family and kinship in Western Europe: the problem of the joint family household’, ibid. 601–28, at 606–9.

¹⁴ The laws are preserved in two main manuscripts. They are known by convention as Grágás and date mostly from the twelfth and thirteenth centuries. The surviving manuscripts are not official compilations. They are remarkable for the detail of their provisions and the sheer number of them. For example, the court procedure section alone numbers more than one hundred pages in the standard edition, the rules and procedures governing rights in land another seventy. In contrast to the barbarian codes, the difficulties interpreting Grágás tend not to involve frustrating ellipsis and terseness, but rather the usual problem of how to assess the connection between legal norms and social practice. The mass of detail does create its own problems. Internal inconsistencies suggest that Grágás preserves some obsolete provisions. But there is no reason to doubt that most of the provisions were in effect some time within the twelfth and thirteenth centuries. Citations of Grágás are to the volume and page number of the editions of Vilhjalmur Finsen: Grágás: Islandernes lovbog i fristatens tid, udgivet efter det kongelige bibliolheks haandskrift (Copenhagen, 1852), hereafter Grágás 1a and 1b; Grágás efter det Arnamagnœanske haandskrift 334 fol., Staôarhðlsbðk (Copenhagen, 1879) hereafter Grágás II; and Grágás, Stykker som findes i det Arnamagnæanske haandskrift 351 fol. Skálholtsbák (Copenhagen, 1883), hereafter Grágás III. All three volumes were reprinted in 1974 by Odense Universitetsforlag. Sections 1–116 of Grágás 1a have recently been translated in Dennis, Andrew, Foote, Peter and Perkins, Richard, Laws of early Iceland: Grágás (Winnipeg, 1980).Google Scholar There are also some sixty or so máldagi, i.e. charters, dating from the twelfth and thirteenth century, briefly evidencing gifts to the bishoprics; they are unfortunately of minimal value in this study.

¹⁵ See Grágás II 145–6; Njáls saga 142:386. For brief treatments see Jðhannesson, Jðn, A history of the old Icelandic commonwealth, tr. Bessason, H. (Winnipeg, 1974) 347–9.Google Scholar Our only evidence of the precariousness of the lot of the kotkarl or cotter must be extracted from a simile recording the reluctant parting of an old chieftain from his land: ‘Glum sat in the high seat and did nothing to prepare for his departure, even though he was called. He had the hall decorated with tapestries; he did not want to depart his land like a kotkarl’ Viga-Glúms saga 26:89, 1F 9.

¹⁶ See above, n. 11.

¹⁷ Literary sources, in a way little different from law codes and custumals, raise problems of how accurately they reflect reality. To reject a source because it is literary is a luxury of those historians who have what, by the conventions of the trade, are assumed to be better sources, if for no reason other than that they are duller. The historian of medieval Iceland has no choice. But this is not as bleak as it sounds. The sagas, unlike romance and even most heroic fare, are consciously realistic. The narrator purports only to report the disputes of Icelandic farmers. The action, except for occasional trips to the continent, takes place in Iceland, in familiar countryside, in familiar interiors. There is simply no reason for the redactor to fashion the cultural and social setting of his story ex nihilo in order to tell his tale. Fictionalising dialogue, inventing characters, does not mean fictionalising the setting in which the action takes place. It does not even mean fictionalising the structures of possible action. In other writings I have made use of comparative materials, usually anthropological ones, to provide a sense of the limits of the possible in a stateless feuding society such as Iceland was. The sagas fall well within the range of described patterns in the ethnographic literature. They look credible in most matters and although some of that credibility might be the conscious creation of sophisticated writers, it is just as likely to be owing to accurate description. For more on the problems in saga scholarship regarding the historicity of the sagas see generally Clover, Carol J., ‘Icelandic family sagas’, in , Clover and Lindow, John, eds., Old Norse-Icelandic literature: a critical guide (Ithaca, 1985) 239–315Google Scholar; and also, , Clover, ‘Introduction,’ to volume on ‘Early law and society’ in Scandinavian Studies 58 (1986) 97–100.Google Scholar

¹⁸ The deteriorating climate was aided by the ineptitude of the native response to resource management; see Gunnarsson, Gísli, ‘A study of causal relations in climate and history with an emphasis on the Icelandic experience’, Meddelande frán Ekonomisk-Historiska Institutionen 17 (Lund, 1980)Google Scholar and Thomas H. McGovern et al., ‘Northern islands, human error, and environmental degradation: a view of social and ecological change in the medieval North Atlantic’, unpublished MS presented at the 1985 American Anthropological Association Meeting, Washington, D.C.

¹⁹ See Jðhannesson, A history of the old Icelandic commonwealth, 303–05 and Gunnarsson, ‘A study of causal relations’, 14–23.

²⁰ See Cleasby, R. and Vigfusson, G., Craigie, Wm. A., ed., An Icelandic-English dictionary, 2nd ed. (Oxford, 1957)Google Scholar and De Vries, Jan, Altnordisches Etymologisches Wönerbuch (Leiden, 1961).Google Scholar

²¹ See Eyrbyggia saga IF 4, 14: 24: ‘I am not willing to divide Helgafell (the farmstead); but I can see we are not suited to have a tvíbýli together so I wish to buy you out.’

²² See, e.g. n. 32 and accompanying text and compare text quoted at n. 123.

²³ An earlier form, búandi, shows the clear link to bú. Bðndi is the present participle of búa, ‘to have a household’.

²⁴ The griðmaðr of the laws usually appears as a húskarl or heimamaðr in the sagas. A heimamaðr need not imply low social rank; like griðmaðr it means someone formally lodged (á vist) in another's household. See, e.g. Thorvard, a chieftain, who is called a homeman (heimamaðr) of Kolbein; Íslendinga saga St. I: 177: 501.

²⁵ The householder was thus obliged to attend the local thing, or send a proxy if he was not able to attend, to support his chieftain (OI goði, pl. goðar). In fact the obligation meant much more than just attending the thing. Both thingmen and goðar looked to each other for support in feud as well as law and one's goði was a nearly indispensable party to disputes with other householders. Thingmen were free to change chieftains upon formal announcement; Grágás 1a 140–41.

²⁶ By ‘the merely juridical household’ I mean to indicate the theoretical household defined by minimal property requirements that give rise to the obligation to declare one's thing attachment as provided in the Grágás provision just cited in the text.

²⁷ This also suggests that access to bóndi status was not especially difficult to achieve and in no way depended on neolocal marriage. See text at n. 79.

²⁸ Eyrbyggja saga 15: 26. This case is also indicative of the possible range of extension in some larger households.

²⁹ Both men in this passage have to be bændr or the passage is without motivation, since in order for there to be a problem about whom to call both would have to be eligible to be called, i.e. bændr.

³⁰ See the description of the farmhouse at Flugumyr where nine men suffocate in the gestahús which apparently was not detached from the main building (Íslendinga saga St. I: 173: 492) and cf. Hoskuld Hvitanessgodi's establishment at Ossaboer where he lodges guests in útibúr (outlying buildings) because his hall was being repaired (Njáls saga 109: 277). Bath-houses were often detached from the main dwelling; see Gestsson, Gisli, ‘Fjórar Baðstofur’, in Kolbeinsson, Guðni, ed., Minjar og Menntir: Afmælisrit helgað Kristjáni Eldjárn, (Reykjavik, 1976) 190–207.Google Scholar

³¹ See, e.g. Eldjárn, Kristján, ‘Two medieval farm sites in Iceland and some remarks on tephrochronology’, in Alan, Small, ed., The fourth viking congress, 1961, (Edinburgh, 1963) 10–19Google Scholar; and Grimsson, Porkell, ‘Miðaldabyggð á Rayðarfelli’, in Minjar og Menntir, 565–76.Google Scholar

³² ‘If two men have a household together and one of them is a landowner and the other a tenant, then the landowner is to be called. If two landowners have a household together or two tenants who can legally be called, then he shall call the one who owns the greatest portion of the household. But if they have equal shares in the household then he shall call whom he wishes from the two of them even though they have no servant. If two men have a household together who are obliged to pay the pingfarakaup [a tax to be paid by those thingmen who meet a certain property requirement in lieu of thing attendance to help defray costs of those who do attend] who do not have a servant it is legal to call one of them. The other shall then pay all the costs that are necessary in proportion to the share he has in the household.’

The same provision, several lines later, contemplates the presence of a married daughter and her husband living with her father: ‘For the household of a man incapable of attending the thing it is lawful to call the following four men if they have their residence there: i) the man's son, ii) his stepson, iii) the near affine who has married his daughter, and iv) his legal fosterson whom the householder has raised’.

³³ See, e.g. Eyrbyggja saga 15: 26; Gisla saga IF 6, chs. 9–10; Guðmundar saga dýra St. 1: 11: 181–82; Njáls saga 14: 45–47.

³⁴ Thus Grágás 1a 126 provides that ‘if men live together [equally possible: household together] and one of them takes in an outlaw and the other does not want to, then the one who does not want to is to name witnesses that he does not want to and that it is without his agreement and say this to five of his neighbours after. Then he is not liable for living with an outlaw as long as he gives him no other aid’. A discussion of the difficulties of repudiating actions of a group of which one was arguably a member can be found in Miller, William Ian, ‘Justifying Skarpheðinn: of pretext and politics in the Icelandic bloodfeud’, Scandinavian Studies 55 (1983) 316–44, at 328–32.Google Scholar For a general discussion of group liability and the class of expiators in the bloodfeud see Heusler, Andreas, Das Strafrecht der Isländersagas (Leipzig, 1911) 57–9.Google Scholar

³⁵ This statement is an impression derived from the sagas and undertaken without a formal study. Group names tend to be used by the narrator and characters who are not group members. This, however, does not prevent members of the group from adopting the name to reference themselves in opposition to others; see, e.g. Ljósvetninga saga IF 10, 23: 69: ‘We Ljosvetnings know that your hostility to us has long been unsparing.’

³⁶ Ljósvetninga saga 22: 64.

³⁷ Hajnal's definition would require one head per household; see above n. 6.

³⁸ Egils saga 56: 151; Laxdæla saga 29: 49.

³⁹ Njáls saga 36: 95.

⁴⁰ See, e.g. Grágás, which provides rules for thing attachment where a bóndi marries a woman who has a bú (1a 139); for men eligible to represent a household headed by a woman (1a 161); governing the killing of foreigners who are lodged in her household (1a 173; II 340) and the disposition of her property (1b 229). The sagas confirm the existence of such households, usually of widows, but unmarried heiresses also headed households; see, e.g. Njáls saga 18:52 (Unn); Guðmundar saga dýra St. I: 5: 168 (Gudrun Thordardottir). One more matter of brief note: Sturla Sighvatsson's household at Saudafell includes, among others, his wife and his wife's mother. Both are designated with the title húsfreyja (Íslendinga saga St. I: 71: 327). Although the mother-in-law might simply bear the title as an honorific and a reflection of past status, this is not altogether clear. Could it be that his mother-in-law had sufficient wealth that she qualified as a head of a separate juridical household? For the possible connection between households headed by widows and witchcraft accusations see Miller, William Ian, ‘Dreams, prophecy and sorcery: blaming the secret offender in Medieval Iceland’, Scandinavian Studies 58 (1986) 101–23, at 114–15.Google Scholar

⁴¹ Íslendinga saga St. I: 69: 324.

⁴² See also Eyrbyggja saga 51: 143–44.

⁴³ ‘If a man had not found a place for those people whom he is responsible for on the previous Moving Day he is to be fined for each of them and the action belongs to anyone who wishes to prosecute.’

⁴⁴ See Jðhannesson, A history of the old Icelandic commonwealth, 355–58.

⁴⁵ See, e.g. Porsteins páttr stangarhqggs IF 11, p. 70 (household of Thorstein); Ljðsvetninga saga 13: 16, 18:51 (household of Thorkel Hake); Hrafns saga Sveinbjarnarsonar St. I: 17: 221 (household of Amundi),

⁴⁶ Saga evidence suggests that many of these were loose marriages; see, e.g. Ljðsvetninga saga 13: 16–17 (marriage); Hrafns saga Sveinbjarnarsonar St. I: 14: 218 (concubinage); Sturlu saga St. I: 4: 65 (concubinage); Njáls saga 39: 103 (concubinage); Reykdæsla saga 30: 24 (marriage). One of Hajnal's essential features of north-west European household formation is the circulation of young people between households as servants before marriage; ‘Two kinds of pre-industrial household formation systems’, 92–9. Service of this sort, he observes, did not exist in joint household systems such as India. Some of the people in service in medieval Iceland were of the north-west European kind, but many did not appear to be. Servants were often poorer kin of the household head, and many seem to have formed a fairly permanent underclass. Although Iceland did not have a joint household system in the manner of India neither was it yet a society whose householding arrangements were typified by the north-west European model.

⁴⁷ Porgils saga ok Hafliða St. I: 1: 12–13, 5: 17.

⁴⁸ See above, n. 43. The laws contemplate separate households for a husband and wife who must find household attachment: ‘If a man has a wife, he shall have found her a position and informed her of it before the fifth day of the week when seven weeks of summer have passed at the latest. If a place has not been obtained for her, so far as she knows, then she has the right to find a position for herself where she wishes’ (Grágás 1a 129). Likewise: ‘If a servant marries and each has lodging in a different place, each shall remain where they are in service if they are bound by their work’ (1a 135). Marriage without coresidence is rare in Europe, if not always so in ethnographic literature; see Wheaton, Robert, ‘Observations on the development of kinship history, 1942–1985’, Journal of Family History 12 (1987) 290.CrossRefGoogle Scholar For the separation of a father and son see Íslendinga saga St. I: 141: 440.

⁴⁹ On the hrepp see Mauer, Konrad, Das Staatsrecht des islāndischen Freistaates (Leipzig, 1909) 499–525Google Scholar; vol. 4 of Vorlesungen über Altnordische Rechtsgeschichte (1907–1910); also Jðhannesson, A history of the old Icelandic commonwealth, 86–9.

⁵⁰ Life-cycle service appears to be what Isolf intended for his daughter in Ljðsvetninga saga 22: 64 in sending her to Eyjolf, although some of these arrangements must have been scarcely distinguishable from concubinage; e.g. consider the succession of housekeepers that Thorgils skardi sought and retained (Porgils saga skarða St. II: 27: 151–52); see also Hávarðar saga Ísfirðings IF 6, 1: 291.

⁵¹ Guðmundar saga dýra St. 1: 10: 179; see also Sturlu saga St. 1: 20: 89; and Porsteins saga hvita If 11, 3: 6. It is of some interest that Solvi is provided with a patronym. The sagas do not care to give us much detail about servants; they are frequently unnamed or if named they are often without patronymic, although in this matter Sturlunga saga is more likely to provide a patronym than the family sagas are. The presence of a patronym is likely to indicate a bóndi's son and thus, perhaps, a life-cycle servant. See, e.g. what appear to be life-cycle services in Sturlu saga St. 1: 12: 78 (Hall?); 15: 72 (Thorolf); Vqðu-Brands páttr If 10, 2: 128–31 (Brand). Younger brothers and kinsman of the powerful Sturlungs could be homeman to their seniors; but there were short-term arrangements and although they might be called homemen they were unlikely to have been servants in any meaningful sense; Íslendinga saga St. 1: 81: 344 (Kolbein and Orœkja). Nevertheless Sturla Sighvatsson can at least contemplate reducing his kinsmen to service; ibid. 125: 407–8.

⁵² Gunnars páttr piðrandabana IF 11, 1:196–7; Sturlu saga St. 1: 2: 64; and 15:81.

⁵³ E.g. Hrafnkels saga If 11, 8: 126 (Eyvind's skósveinn, i.e. shoeboy, servant).

⁵⁴ On the obligation to maintain poor kinsmen see generally Grágás, Ómaga-bálkr', ib 3–28, II 103–51, esp. ib 3–4, 11, 26–27.

⁵⁵ Compare, however, the incredibly low nuptiality rates for Iceland in the census of 1703, suggesting a virtual prohibition of servant marriage as claimed in Gunnarsson, Gísli, ‘Fertility and nuptiality in Iceland's demographic history’, Meddelande från Ekonomisk-Historiska Institutionen, Lunds University, 12 (1980) 7–15Google Scholar and questioned in Hajnal, ‘European marriage patterns in perspective’, 137–38.

A provision like this could not hope to achieve its goal unless there were also effective ways of discouraging illegitimacy. Strictures against fornication and seduction provided for a declining scale of punishment depending on the status of the woman involved. This would have the effect of insulating men of the householding class in their depredations on servant would not do much to discourage illegitimacy. In any event liability for support attached to the father of the child or to his kin or the hrepp if he was unable to assume the obligation. For the manner in which the laws dealt with improverished men who made it a habit of producing bastards, see below, n. 134.

⁵⁶ Laxdœla saga 28: 76; 74: 215; Gisla saga 2: 7 (describing, however, events in Norway). Contrast with this the 1703 census which has 77.8% of children 0–14 living at home; Gunnarsson, ‘Fertility and nuptiality’, 9.

⁵⁷ Droplaugarsona saga IF 11, 5: 150; Hœsa-póris saga If 3, 2: 7; Laxdœla saga 16: 37.

⁵⁸ In addition to the cases in the preceding note see Sturlu saga St. 1: 25: 98.

⁵⁹ Laxœla saga 27: 75; Njáls saga 94: 237; Sturlu saga St. 1: 34: 113; Víga-Glúms saga 12: 40–41.

⁶⁰ Njáls saga 27: 74.

⁶¹ Prestssaga Guðmundar góða St. 1:4: 123.

⁶² Such may have been a contributing motive in the fosterings mentioned in Ljósvetninga saga 22: 63 and Viga-Glúms saga 17: 57.

⁶³ The sagas give us no examples of suing to recover the fóstrlaun and the accounts of kinsmen taking in the children of kin who have met untimely deaths or whose fathers have gone abroad do not indicate whether they were repaid; one case suggests that there was no intention that it should be repaid other than by support; Njáls saga 93:236. See also Laxœla saga 50: 158; porsteins saga hvita 7:17. There is, however, an ironic use of fóstrlaun which might indicate otherwise. Thus after Bolli has killed Kjartan, Thorgerd, who is Kjartan's mother and Bolli's foster mother, thinks Bolli has ‘made a sorry repayment for his fostering’ (sár fóstrlaunin); Laxdœla saga 51: 159.

⁶⁴ E.g. Egils saga IF 2, 40: 101; Njáls saga 9: 29: 39: 103.

⁶⁵ See Hœnsa-póris saga 11: 32; Reykdœla saga 28: 238. These are the only saga examples unless we also include the late Víglundar saga IF 14, 7: 75.

⁶⁶ Víga-Glúms saga 10: 36.

⁶⁷ Ljósvetninga saga 13: 16–17 (Gudrun); Njáls saga 39: 103 (Gudfinna).

⁶⁸ Njáls saga 128: 325.

⁶⁹ Eyrbyggja saga 54: 150; see also ibid. 11: 18 (60 freemen).

⁷⁰ Sqrla páttr IF 10, 1: 109. The number of servants seems excessive, but given the number of servants the number of cows is not excessive. In this regard note that the minimal property requirement obliging a farmer to pay the pingfararkaup was a cow free of debt or the value of a cow for each of his skuldahjún; see the text above at nn. 20–21 and also Jóhannesson, A history of the old Icelandic commonwealth, 289–90.

⁷¹ Páls saga byskups 5: 260 in vol. 1 of Byskupa sōgur, Guōni, Jónsson, ed. (Reykjavik, 1953).Google Scholar

⁷² Jochens, ‘En Islande Médiévale’, 107, would attribute this complex household to the dramatic purposes of an author who needed to have these people residing together so that they could be burned together as per the demands of the plot. Even if, admittedly, the extent and complexity of Njal's household are somewhat aberrational, the fact that it is extended or complex is not. Moreover, Íslendinga saga St. I: 170–73: 481–92 indicates that a group of people who did not share a household could be conveniently burned together as invitees to a feast.

⁷³ Eyrbyggja saga 50: 139.

⁷⁴ Ljósvetninga saga 13: 16–17.

⁷⁵ Ibid. 13: 16, 18: 51.

⁷⁶ Njáls saga 120: 305.

⁷⁷ Hrafns saga Sveinbjarnarsonar St. I: 17: 221.

⁷⁸ Hænsa-Póris saga 16: 42: VQōu-Brands páttr 1: 125.

⁷⁹ ‘Neolocalism’, in the jargon of anthropology, refers to the practice of a newly married couple setting up on their own, and living by themselves, not with either set of parents; Laslett, ‘Family and household’, 531. A rule of neolocality implies, therefore, simple households.

⁸⁰ Hungrvaka 1: 2 in Guōni, Jónsson, ed., Byskupa sögur 1 (Reykjavik, 1953)Google Scholar; Cleasby-Vigfusson's gloss, s.v. hjún, I: ‘a married couple should get a house’ is less ambivalent but hard to justify grammatically.

⁸¹ Gísla saga 10: 34: ‘Saman er brœōra eign bezt at líta ok at sjá.’ See below text at nn. 123–29.

⁸² Jochens, ‘En Islande Médiévale’, 100, describes three marriage models typical of Sturlunga saga. In one, men are twenty and women eighteen or younger. In a second a young man marries an older woman, usually a widow with children. The third is the inverse of the second. The second and third type would do little to raise the average age of first marriages, assuming that the younger men who take older women then take younger women at a later stage of their life cycles. Jochens gives no numbers regarding the prevalence of the types, nor is there any way, given the nature of the evidence, that she could. We have no means of determining the number of celibates, nor the prevalence of illegitimacy in any quantifiable way. Jochens' models are taken from Sturlunga saga, a source perceptibly skewed to the wealthy and powerful families, and only for the most wealthy and powerful of these are the accounts sufficiently circumstantial to allow for even a moderately accurate determination of age. Nevertheless, it should also be noted that the laws envisage the possibility, at least, of very early marriages for women. Consider Grágás 1a 224: ‘If a women is married when she is sixteen or younger…,’ ‘If she is widowed when she is younger than sixteen…’ and such marriages are confirmed by the experiences of principal female characters in Laxdæla saga 34: 93 and Njáls saga.

⁸³ In a stationary population roughly 20% of all married couples will produce no children surviving their parents and another 20% will produce only daughters who survive their parents. See Wrigley, E. A., ‘Fertility strategy for the individual and the group’, in Charles, Tilly, ed., Historical studies of changing fertility (Princeton, 1978) 135–54.Google Scholar

⁸⁴ See Miller, William Ian, ‘Gift, sale, payment, raid: case studies in the negotiation and classification of exchange in medieval Iceland’, Speculum 61 (1986) 18–50CrossRefGoogle Scholar, at 49–50. A sale of land would most likely involve a loss of status for the seller, since straintened circumstances, fertility failure, or insufficient strength to stand up to the would-be purchaser would be the usual reasons the land would be available in the first place; see, e.g., Laxdæla saga 75: 218–21; Sturlu saga St. I: 15–16: 81–84 (cases of Erlend prest and Ozur's inheritance); 23:96.

⁸⁵ Our information about medieval Icelandic demography is very limited; see Hastrup, Kirsten, Culture and history in medieval Iceland (Oxford, 1985) 165–77.Google Scholar

⁸⁶ Ibid. 170.

⁸⁷ Hastrup, 172–77, takes this view.

⁸⁸ See, e.g. Jóhannesson, A history of the old Icelandic commonwealth, 29–34, 345–50; Byock, Jesse L., Feud in the Icelandic saga (Berkeley, 1982) 143–50Google Scholar; McGovern, Thomas H.et al., ‘Northern islands, human error, and environmental degradation: a view of social and ecological change in the medieval North Atlantic’Google Scholar; Hastrup, Culture and history in medieval Iceland, 189–96.

⁸⁹ Thorarinsson, S., ‘Tephrochronology and Medieval Iceland’, in Rainer, Berger ed., Scientific methods in medieval archaeology (Berkeley, 1970) 295–328Google Scholar, at 320–5.

⁹⁰ Hastrup, Culture and history in medieval Iceland, 172, deduces the partition of lands from the inheritance rules, but saga evidence suggests that this need not be the case; see below, n. 99.

⁹¹ The examples of neolocal marriage in Jochens, ‘En Islande Médiévale’, 98–101, involve only the wealthiest chieftain families.

⁹² Guōmundar saga dýra St. I: 1: 162. There is no indication of the marital status of the sons, but it should be noted that Eyjolf apparently intended them to have one establishment together.

⁹³ See the references above, n. 88.

⁹⁴ Sturlu saga St. I: 7: 69; Guōmundar saga dýra St. I: 1: 161; but cf. Egils saga 79: 275.

⁹⁵ Egils saga 56: 151. The son could have equal say in the running of the household even though he was unmarried; see Laxdæla saga20: 49.

⁹⁶ Laxdæla saga 35:96–7: 43: 130; Njáls saga 61: 154; 90:225; Sturlu saga 25:91. See also the Grágás provision quoted in n. 32.

⁹⁷ Laxdæla saga 20:49; Íslendinga saga St. 1: 39: 284; 86–7: 358–60; see also Jochens, ‘En Islande Médiévale’, 97, 101.

⁹⁸ See, e.g. Gísla saga 10: 34; Íslendinga saga St. I: 4: 232; 33: 262; 81: 344; 83: 346–7; Sturlu saga St. I: 9: 72. Although it is of little probative value for the medieval period the presence of the householder's siblings was a marked feature of households enumerated in the 1703 census, marked, that is, relative to roughly contemporaneous figures for England, Geneva, and Norway; see Manntaliō 1703, table VIII, and Wall, ‘Introduction’, in Family forms 53.

⁹⁹ Grágás 1a 218; Gísla saga 4: 16; Ljósvetninga saga 22: 62; Sturlu saga St. I: 3: 65 (although brothers mismanage their affairs and must sell land shortly thereafter). Brothers could also formally divide the inheritance. Still, as far as I know, there is only one partition in kind of a family farm at the death of a parent when the decedent's sons were the heirs; Valla-Ljóts saga 3: 241. Among the wealthier families who might control more than one farm, property divisions among brothers at the death of a parent tended to keep farms intact; Hrafnkels saga 10: 133; Njáls saga 78: 192. A practice noted in the family sagas has one brother take his share in livestock, the other his śhare in land; Laxdæla saga 26: 73; Gísla saga 10: 35. Partitions of lands were more likely when the class of heirs included people not related to each other; Víga-Glúms saga 5: 15; see also Eyrbyggja saga 14: 24–25.

¹⁰⁰ The marital status of the brothers is not always determinable; but see, e.g. Gísla saga 4:16; Guōmundar saga dýra St. 1: 11: 181; Heiōarviga saga IF 3, 41: 326; Ílendinga saga St. 1: 46: 294; Laxdæla saga 32: 86; 46: 144; Ljósvetninga saga 20: 54; 22: 62; Njáls saga 47: 120; Sturlu saga St. 1: 3: 65; Porgils saga skarōda 14: 123.

¹⁰¹ Íslendinga saga St. 1: 86: 358 (Br & SiHu), 33: 262 (WiBrs); Laxdœla saga 46: 139 (Fa & So); Egils saga 56: 151 (Fa & So); Laxdœla saga 32: 86 (Br & SiHu); Njáls saga 61: 154 (DaHu); Sturlu saga St. 1: 7: 69 (Fa & So, WiBr, WiBrWi), 25: 97 (DaHu).

¹⁰² Affinal relations, especially those with sisters' husbands and daughters' husbands, tended to be kept in good order. The laws recognise the closeness of these kin (Grágás 1a 47, 62, 201) and the sagas confirm them as particularly active in matters of vengeance; see, e.g. the roles of Kari in Njáls saga, Thorgrim alikarl in Guðmundar saga dýra; Otrygg in Ljósvetninga saga 24: 77.

¹⁰³ Sturlu saga St. 1: 15: 82, 19: 87.

¹⁰⁴ Íslendinga saga St. 1: 106: 383 and 32: 260.

¹⁰⁵ Sturlu saga St. I: 28: 103; Porsteins ðáttr stangarhqggs, 77–8.

¹⁰⁶ See Grágás 1a 246–9, II 85–7 (setting aside transfers of ancestor); 1b 76–7, 410–14 (setting aside transfers of guardian); see also 1b 17–18. Successful reclamations were usually made in blood rather than at law; Eyrbyggja saga 32–7: 87–102; Hávarðar saga Ísfirðings 14: 337–41.

¹⁰⁷ E.g. the estate of Thord Goddi in Laxdœla saga 16: 37; 22: 62 and Reykjaholt in Íslendinga saga St. 1: 16: 241.

¹⁰⁸ E.g. Eyrbyggja saga 31: 84.

¹⁰⁹ Sturlu saga St. 1: 28: 103.

¹¹⁰ See, generally, ‘Omaga-bálkr’, Grágás 1b 3–28, II 103–51.

¹¹¹ Wheaton, ‘Family and kinship in Western Europe’, 611.

¹¹² The notion of individually owned property must be qualified to some extent to take account of the claims of heirs and dependants. I do not wish to enter here into the mostly arid debates about the rise of individualism. I mean only to indicate that it is the difference in the way in which title to a particular piece of property is conceived that determines whether all of it or only some fraction will be subject to the owner's direction or be considered an asset of his estate when he dies.

¹¹³ See, e.g. Bandamanna saga IF 7, 1: 294; Egils saga 40: 102; 58: 173; Eyrbyggja saga 30: 81ff; Laxdœla saga 24: 66 where it is noted that Olaf's father did not envy his son's popularity, implying the likelihood of the contrary.

¹¹⁴ Eyrbyggja saga 30: 82–3 (jointly owned field); Njáls saga 36: 92 (jointly owned woods).

¹¹⁵ See, e.g. Eyrbyggja saga 30: 81ff; Laxdœla saga (Bolli v. Kjartan); Ljósvetninga saga 2–4: 6–15; and generally the dealings of the Sturlusons in Íslendinga saga. Inheritance disputes, almost by definition, frequently pit household members against each other.

¹¹⁶ It may be of some interest that the sagas record no patricides, filicides, or fratricides, although there are some very close calls. See, e.g. Ljósvetninga saga 20: 57 where a father must be restrained from torching a house in which he knows his son is lodged, after the son has refused an offer of free passage; also ibid. 2: 9; similarly Porgils saga skarða St. II: 32: 160 although this time it is the son who is willing to burn his father. We might grant one case of filicide if we are to count sins of omission. Thorkel, a priest, prefers to confess his son and hand him over to his executioners rather than pay them the compensation they request; ibid. 30: 158.

¹¹⁷ E.g. Hrafns saga Sveinbjarnarsonar 8: 285—86 in Gudbrand Vigfusson, ed., Sturlunga saga, vol. II (Oxford, 1878); Ljósvetninga saga 21: 61–62.

¹¹⁸ The dispute between the households of Gunnar and Njal in Njáls saga is perhaps the best example; see also Porsteings páttr stangarhqggs. It is in this context that the goading women and old men of the sagas should be understood for which see my ‘Choosing the avenger: some aspects of the bloodfeud in medieval Iceland and England’, Law and History Review 1 (1983), 159–204.Google Scholar

¹¹⁹ An occasional notice in the sagas suggests that disputes between homeman and householder need not remain repressed. A certain Grim is given an end consonant with his name in this brief notice: ‘he was killed by his housecarls’; Porgils saga ok Hafliða St. 1: 11: 30.

¹²⁰ See, e.g. Eyrbyggja saga 14: 24–6; Laxdœla saga 19: 45–7; but cf. Hrafns saga Sveinbjarnarson St. I: 13: 214; Sturlu saga St. I: 26: 100. The property of minors (below sixteen for males, twenty for females) was administered by a guardian called the fjárvarðveizlumaðr. In the absence of a father the role would be filled by a brother who had reached majority; see Grágás 1a 225–6.

¹²¹ A similar case occurs in Eyrbyggja saga 14: 24–6 where Snorri claims his paternal inheritance from his stepfather who, though willing to pay out the inheritance is not willing to share a household: ‘I am not willing to partition Helgafell’, he said, ‘but I see that we are not suited to have a joint household (tvibýli) together, so I will buy you out.’

¹²² The passage could be more tendentiously but properly rendered: ‘it was agreed later that Kodran should have a household at (bjó i) Modruvellir’. See Cleasby-Vigfusson, s.v. búa A.1.4.

¹²³ Gisla saga 5: 19.

¹²⁴ Geirmund and Gudrid are siblings, very possibly kin to Gisli, who are in service with Gisli and Thorkel. The value of Gisla saga's evidence for householding should perhaps be discounted somewhat because the householding type figures more integrally in the plot than the general run of saga householding evidence and hence is more likely to suffer fictionalising distortions. That said, there is nothing in the description of householding arrangements in the saga that is inconsistent with the accounts derived from safer contexts.

¹²⁵ Gisli, for example, had killed Thorkel's best friend in Norway before they emigrated. The two brothers were frequently at odds over controlling sexual access to their sister Thordis. For the dependants see text below at note 126.

¹²⁶ See Gisla saga, xiv.

¹²⁷ Some notice should be made here of demand divorce. According to the sagas, in pre-Christian times either husband or wife could renounce the marriage simply by doing so in the presence of witnesses. After the introduction of Christianity, separation ostensibly required the bishop's approval except in the following cases: if the couple were unable to support their dependants, if one dealt the other an injury or wound, or if one spouse had wealth and the other did not and the latter's dependent kin were a drain on the property of the other (Grágás ib 39–40). Sturlunga saga indicates that separation was not uncommon; see, e.g. Íslendinga saga St. 1: 3: 231; 53: 304; 75: 335; 82: 346.

¹²⁸ See above, text and notes at nn. 56–63.

¹²⁹ See Miller, ‘Gift, sale, payment, raid’.

¹³⁰ E.g. Vqðu-Brands páttr 2: 130–31.

¹³¹ Instances of uxorilocal marriage were not uncommon; see, e.g. Laxdœla saga 35: 96–7; 43: 130; 69: 203; 70: 207–08; Guðmundar saga dýra St. I: 5: 169; 12: 186. Couples could also relocate to provide care for ageing parents; see, e.g. Laxdœla saga 10: 20.

¹³² Guomundar saga dýra St. 1: 1. 162; Viga-Glúms saga 11: 40.

¹³³ See the second marriage of Gudrun in Laxdœla saga and Hallgerd's second and third marriages in Njáls saga. Wealthy widows had a hard time staying unattached and often entered into, or were constrained to enter into, joint householding arrangements or loose marriages with men intent on their property: e.g. Sturlu saga St. 1: 2: 64; Íslendinga saga St. I: 52: 302.

¹³⁴ Gunnarsson, ‘Fertility and nuptiality’, 15–9, doubts excessive illegitimacy. As has become the frequent refrain in this paper, we just do not know. If we count the children of concubines as illegitimate then the sagas and genealogies suggest the rate was quite high among chieftains. The people most concerned with keeping the rates low were those upon whom the burden of support ultimately rested, that is kinsmen or members of the hrepp. Thus Grágás 1b 26: ‘No one is obliged to accept more than two illegitimate third cousins fathered by one man unless the father of the children is castrated’, and further (1b 28) where it is provided that the kin of vagabonds were not liable for their children unless the parents attached themselves to a household.

¹³⁵ The provisions are brutally unsympathetic. Besides being liable for full outlawry (Grágás 1a 139—40) vagrants and those who showed them charity were subject to a number of legal disabilities. E.g. fornication with a beggarwoman was unactionable (1b 48); it was lawful to castrate a vagabond and it was unactionable if he were injured or killed in the process (1b 103); one could take in beggars only to whip them (ib 179); nor was one to feed or shelter them at the thing on pain of lesser outlawry; their booths could be knocked down and if they happened to have any property with them it could be taken from them without liability (1b 14).

¹³⁶ Fóstbrœbra saga IF 6, 14: 195.

¹³⁷ See above, n. 85.

¹³⁸ Statistical Bureau of Iceland, Manntalið 1703 (Reykjavik, 1960); Larusson, Björn, The Old Icelandic Land Registers (Lund, 1967).Google Scholar

¹³⁹ Jochens, ‘En Islande médiévale’.