[In the following essay, Battles examines references to subterranean earth dwellings in several Scandinavian and Old English works including the Volsunga Saga.]

The past three decades have witnessed a bewildering variety of interpretations of the Old English Wife's Lament. We are no longer even certain that it is, in fact, the lament of a “wife”; critics have suggested that the narrator is a lordless retainer, a heathen deity, the soul yearning for the body (or the body for the soul), or a revenant. Numerous syntactical and lexical ambiguities—as well as the poem's abstract diction in general—have led to interpretations which differ dramatically even in regard to the basic facts of the poem.¹

At the heart of the debate over the general circumstances described by the poem is the narrator's dwelling, an eoroscraef or eorosele underneath an actreo in a wuda bearu (ll. 27-29).² The dwelling and its surroundings are described at considerable length (ll. 27-36), and this detailed description contrasts starkly with the rest of the poem, whose abstract language and nonlinear order make it difficult to reconstruct the events being narrated. The eoroscraefleordsele provides a tantalizing clue to the situation of the narrator. At least two recent critics have viewed the eoroscraefleordsele as the key to the poem's meaning,³ and certainly any interpretation of the Wife's Lament must come to terms with the curious dwelling in which the narrator lives. It makes a difference whether we are to imagine her living in a hovel, a grave, or a heathen sanctuary.⁴

Current readings of the eoroscraefleordsele essentially fall into three categories. The most widely accepted interpretation, most forcefully advanced by Karl Wentersdorf, holds it to be a cave.⁵ A second approach, first suggested by R. F. Leslie, interprets the dwelling as a grave, specifically a “chambered barrow”;⁶ this has led several critics to suggest that the narrator is dead.⁷ Finally, Earl R. Anderson has recently suggested that eoroseleleoroscraef is a sunken-featured building, a structure quite common in Anglo-Saxon England.⁸ On purely linguistic grounds, all of these theories could be correct. Eoroscraef has a broad semantic range, denoting any hollow place in the earth (e.g., “cave” in Gen. 2597, “grave” in Wand. 84). The term eorosele occurs in only one poem other than the Wife's Lament, namely in Beowulf, but its meaning there is not entirely clear. This does not mean, of course, that all readings of eoroscraefleorosele are equally likely, or that—as the most recent critic to discuss the issue has suggested—the poet had no particular kind of stucture in mind at all.⁹ Indeed, it has been argued that the eoroscraefleorosele is not a literal dwelling at all, but rather a metaphorical expression of the speaker's state of mind. However, the very specificity with which the narrator describes the eoroscraefleorosele—that it is in a grove of trees (in fact underneath an oak) in a desolate countryside overgrown with briars—indicates something rather more specific than a mental abstraction.¹⁰

In fact, archeological evidence and numerous literary parallels suggest that the eoroscraefleorosele is neither a grave, a natural cave, nor a “sunken-featured dwelling,” but rather a southerrain, an artificial underground dwelling or chamber. During the early Middle Ages, souterrains were quite common in Ireland, Scotland, and Brittany, and were also found in Cornwall, Denmark, and northern England. These structures are documented not only by archeological finds, but also by a wealth of Icelandic, Irish, and English historical and literary sources. Reading the eoroscraefleorosele in the Wife's Lament as a souterrain will shed light on several passages in the poem which have thus far eluded a convincing explanation. In both ON and ME literature souterrains appear as secret hideaways generally associated with fugitives, who are often women; thus, the description of a souterrain as the narrator's dwelling would resonate with several prominent themes in the Wife's Lament, notably confinement and secret hostility.

Before turning to a discussion of souterrains, it would be useful to review the two most recently proposed—and most widely-accepted—readings of the eoroscraefleorosele.¹¹ Earl Anderson has equated this structure with a primitive kind of dwelling very common in Anglo-Saxon England, a “sunken-featured building” referred to as Grubenhaus in continental archeology.¹² Citing an earlier work by Joseph Harris, Anderson states that the

eoroscraef is a partially sunken hut, essentially a carpentered structure used as a permanent dwelling. Tacitus alludes to such sunken huts as winter dwellings, and seems to regard them as structures of the same type as storage pits. … Archaeologists have confirmed the presence, throughout Germanic lands, of sunken dwellings of the sort that Tacitus refers to.

(p. 74)

Anderson and Harris were not the first to posit a Grubenhaus: in the Reallexikon der germanischen Altertumskunde, Johannes Hoops noted that the Wife's Lament mentions “eine Art Grubenhutte, eorpsele, … in der Landarme hin und wieder sich aufhielten.”¹³ Both Harris and Anderson cite Much-Jankuhn's notes on Tacitus's Germania, noting that Tacitus describes a sunken featured building in his discussion of Germanic dwellings.¹⁴ Tacitus writes:

solent et subterraneos specus aperire eosque multo insuper fimo onerant, suffugium hiemis et receptaculum frugibus, quia rigorem frigorum eius modi loci molliunt, et si quando hostis advenit, aperta populatur, abdita autem et defossa aut ignorantur aut eo ipso fallunt, quod quaerrenda sunt.¹⁵

Compare this to an archeological description of the type of dwelling that Anderson and Harris posit:

“Sunken-featured building” is a cumbersome and inelegant term adopted for relatively small structures erected over a subrectangular pit dug in the ground, ranging in size from 3 x 2 m up to the largest so far excavated at Upton (Northamptonshire) which measured 9.1 x 5.5 m (Jackson et al. 1969). The superstructure was supported by 1 or 3 posts at each end, represented by post-holes on either the upper or lower edges of the pit. A gable roof is implied by such an arrangement. …¹⁶

The Anglo-Saxon sunken-featured building is similar to an ordinary hut or house, except that its floor is sunk into the ground. Assuming that the narrator of the Wife's Lament has been sent into hiding by her lord, it would make little sense for her hiding-place to be a sunken-featured dwelling, since it would be plainly visible. Moreover, such a structure would hardly be located underneath an oak-tree, which the poem mentions twice (under actreo in pam eoroscraefe, l. 28; under actreo geond pas eoroscrafu, l. 36); both Harris and Anderson remain silent on this point.

Anderson also cites a gloss eorphus = ypogaeum vel subterraneum, and concludes that this “must refer to the same type of house” (pp. 74-75), i.e., to a sunken-featured dwelling, but the only evidence he adduces is a line from Bald's Leechbook: Romane him worhton eorphus for oaere lyfte wylme (2. 146 l. 16). However, while the identification of the eoroscraefleorosele with the term eordhus is most likely correct, we have no archeological evidence whatever to suggest that the Romans built sunken-featured dwellings; they are Anglo-Saxon structures. One might argue that the Leechbook author could well have mistaken an earlier Anglo-Saxon structure for a Roman one, but would he have attributed the relatively small, timbered, sunken-featured buildings to Romans? These humble structures certainly do not compare to the awe-inspiring eald enta geweorc of OE poetry. In Beowulf, the eorahus—the dragon's den—is certainly not a sunken hut; neither the ME eorphus nor the ON cognate jarohus refer to a “sunken-featured dwelling.” It is therefore unlikely that the narrator of the Wife's Lament is living in this kind of structure.

In his discussion of the eordscraefleordsele, Karl Wentersdorf discusses both the “sunken-featured dwelling” (which he calls “subterranean or semi-subterranean earth-houses,” p. 501) and a kind of structure known as denehole, “vertical shafts 2-4 feet in diameter and 30-150 feet in depth, widening out at the bottom into one or more chambers 15-50 feet in diameter” (p. 500). He states that a denehole would be an unlikely candidate for the eoroscraefleorosele, and that a sunken-featured dwelling would have been too small “to suit the poet's vision of the wife in WL as wandering geond pas eoroscrafu” (p. 501). He thus concludes that the dwelling described by the Wife's Lament cannot be man-made, but must be a natural cavern. As R. F. Leslie has pointed out, however, the adjective eald (l. 29a) suggests an artificial structure,¹⁷ and Wentersdorf's argument that a natural cave might be described as being eald because of its occupation by humans (p. 501) is not entirely convincing. At any rate, Wentersdorf bases his argument for a natural cave primarily upon negative evidence: having excluded deneholes and sunken-featured dwellings from consideration, he argues that the eoroscraefleorosele cannot be an artificial structure.

Yet Wentersdorf did not consider the most likely kind of artificial underground structure, namely a souterrain. Simply stated, souterrains are underground buildings. It is beyond the scope of this paper to offer a full description of the various types of souterrains, or to dwell extensively on their construction; in Scotland alone, souterrains come in a bewildering variety, and to date no one has worked out an acceptable system by which they may be categorized.¹⁸ They range from cellars to granaries to hideaways to genuine dwellings; they may or may not have above-ground structures associated with them; they may be lined with stone or timber and may range in size from the cramped “rathole”¹⁹ to multi-chambered, two-story structures,²⁰ and can reach a length of 190 feet.²¹ As noted, they are found in Ireland and Scotland (250 known examples alone in Ireland's Co. Louth²²), as well as in Brittany, Denmark (where only a few examples have been found, all in the Vendsyssel region of the Jutland peninsula), Cornwall, and Iceland. Nine souterrains have been discovered in the northern-most region of Northumbria, and it is likely that others were destroyed before they were recorded;²³ souterrains were known to later English authors such as Lagamon and the Sir Tristrem poet. Souterrains are well-documented both archeologically and historically, and they are also described by various medieval authors. The evidence from these sources suggests that the eoroscraefleorbsele in the Wife's Lament is indeed most likely a souterrain.

Historical sources document the use of souterrains as hideaways during the time of the Viking incursions in Ireland. The Irish Annals of Ulster record the following incident during the year 863:

Uamh Achaidh Alddai 7 Cnodhbai 7 uam Feirt Boadan os Dubadh 7 uam Mna Angobann ro scruidiset Gaill, quod antea non perfectum est, i. a fecht ro slatsat.iii. righ Gall fercnn Flaind m. Conaing, i. Amhlaim 7 Imhar 7 Auisle; 7 Lorcan m. Cathail leo occa, ri Mide.²⁴

Though there is some debate about dating Irish souterrains, recent archeological investigations have been able to provide evidence that one in Coolcran, Co. Fermanagh, was constructed between 813 and 831,²⁵ and it is not unreasonable to suppose that they would have been used during the Viking raids. The Icelandic Landnamabok offers an illuminating parallel to the description in the Annals of Ulster; there has even been some speculation that the two passages describe the same incursion,²⁶ but considering the paucity of evidence, this must remain speculation:

… Leifr for i hernao i vestrviking; hann herjaoi a Irland, ok fann par jarohus mikit; par gekk hann inn f, og var myrkt, par til er lysti af sveroi bvi, er maor helt i. Leifr drap Pann mann ok tok sveroit, ok mikit fe af honum; sioan var hann kallaor Hjorleifr.²⁷

The Icelandic jarohus can refer either to a souterrain or to a cellar-like chamber underneath a house;²⁸ however, since this is a description of an Irish structure like the one mentioned in the Annals of Ulster, it most likely refers to a souterrain rather than to a cellar-chamber.

Liber VII of Saxo's Gesta Danorum gives a description of the construction and function of such a souterrain. When King Regnald hears of the impending incursion of Gunnar, a ferocious and bloodthirsty Swede, he sends his daughter Drott into hiding:

Sed et rex Normanorum Regnaldus, ultime senectutis, audito tyranni studio, filiam Drotam facto specu claudendam curauit, aptumque ei, prouisis in longum sumptibus, famulicium tribuit … Et ne conspicua cauee pateret elacio, solidiori terre tumorem equabat.²⁹

When Gunnar has killed Regnaldus and hears of Drott's hideaway, he searches for and finds it:

Quo eueniti ut, dum ipse quoque lustrandi inter alios officio fungeretur, subterranei murmuris sonum dubia procul aure colligeret. Tunc sensim progressus cerciorem humane uocis habitum deprehendit. Cumque subiectam pedibus humum ad solidum fodi iussisset, subito patente cauea cuniculatos animaduertit anfractus. Trucidatis itaque famulis, dum detectos antri aditus tueri cupiunt, puella cum repositis inibi spoliis cauo extrahitur.³⁰

The description of this souterrain suggests a spacious, well-provided structure, and if one imagines the narrator of the Wife's Lament in a similar dwelling, it becomes clear why the narrator would have room enough to wander geond (l. 36) its confines; likewise, “winding tunnels” (cuniculati anfractus) may well be the reason why the narrator of the Wife's Lament speaks of her dwelling in both the singular and the plural;³¹ the structure as a whole is an eoroscraef, but it is also composed of several individual eoroscrafu.

This passage in the Gesta Danorum also offers a more general parallel to the situation of the narrator in the Wife's Lament, for Drott, like the narrator, was commanded to dwell in a subterranean structure. Of course, the situation is not very clearly defined in the Wife's Lament, so that we cannot be sure that the narrator was sent into her eoroscraefleorosele because of a threatening situation as Drott was, but the statement Sceal ic feor ge neah / mines felaleofan faehou dreogan (ll. 25-26) may well refer to just that.³² It is possible that the narrator of the Wife's Lament has been sent into her eoroscraefleorbsele in order to escape capture by an enemy; in Icelandic sagas, for instance, this happens frequently. In Volsunga saga, when he hears of the approach of an enemy army under Alf, Sigmundr konungr … dro saman her, en Hjordisi var ekit til skogar vio eina ambatt, ok mikit fe for med peim. Hun var par, meoan peir boroust.³³ In a passage from Floamanna saga reminiscent of the description in the Gesta Danorum, vikings marauding in Ireland discover two women and great treasure in a hidden souterrain (jarohus), and, after killing or driving off the men who are also in the souterrain, they capture the women.³⁴ In Gongu-Hrolfs saga, women are hidden i einus jardhus³⁵ from the attacking Scottish raiders under Tryggvi. It is not clear what kind of jarohus is being referred to here, since the saga mentions no further details, but the very brevity with which the author treats this scene suggests that hiding a woman (or several women) in a jarohus is a commonplace, or at the least not something that would seem unfamiliar to the audience.³⁶

Icelandic sagas also offer interesting parallels to another aspect of the Wife's Lament which has been the subject of much unsatisfactory debate. The poem repeatedly mentions the grove of trees and the oak-tree (hearheard, l. 15b, on wuda bearwe, / under actreo, 27b-28a, under actreo, 36a) near which the eoroscraefleorosele is located. Some critics have read this as an indication that the structure is a pagan shrine,³⁷ but there is a more humbly realistic reason why the narrator might emphasize the trees, namely because they are dominant features of the landscape surrounding her dwelling. Of course, that is not to say that the trees in the Wife's Lament have no metaphorical associations whatsoever, but rather that there is no need to assume that the mention of the trees is so unusual as to evoke connotations of a pagan shrine. Souterrains are quite frequently located in forests, since they are generally intended as hideaways and a forest would provide natural camouflage for an undergound dwelling. Here again, Icelandic sagas offer illuminating parallels; in Volsunga saga, Sigmund is hiding from King Siggeir, and meets Signy:

For hun nu ok hittir broour sinn, ok taka Pau Pat rao, at hann gerir parjarohus i skoginum, ok ferr nu pvi fram um hrio, at Signy leynir honum har ok faer honum hat, er hann purfti at hafa. En Siggeirr konungr aetlar, at peir se allir dauoir Volsungar.³⁸

Later on, of course, Signy visits Sigmund in disguise and seduces him, conceiving Fitela from the union. The construction of a jarohus in a forest is paralleled by a similar description in Droplaugarsona saga:

Nu logousk peir yfir ana meo porkel trana ok komu a pann boe, er a Bakka heitir, fyrir austan fljotit ok gengu par i fjos ok toku par pal ok reku ok foru a brott

sioan ok paoan ut til Oddmarsloekjar fyrir vestan Eioaskog. Vio loekinn grofu Peir ser jarohuss ok foerou mold alla ut a loekinn. Vildu Peir eiga fylgsni pat, ef peir pyrfti til at taka.³⁹

This surely is not a sophisticated structure, but it would evidently serve in a time of need. Hrolfs saga kraka mentions a jarohus in a forest used by Halfdan's sons, Hroar and Helgi, to avoid detection by King Frooi.⁴⁰ Perhaps the clearest illustration comes from Floamanna saga; trying to avoid Viking raiders, a group of Irish are hiding in a jarohus on top of which stands a tree. The Northmen

… komu til Irlands um sumarit; var par skogr fyrir, er beir komu at; gengu sioan upp i skoginn, ok i einum stao sa peir fallit lauf af tre. peir kippa upp eikinni ok finna par jarohus undir. Peir sja menn meo vapnum niori i husinu. porgils gerir sinum monnum kost, at sa skal eignaz iii. kostgripi, er fyrstr gengr i husit, en allir jatta pvi nema Gyror. Eptir Pat hljop porgils i husit, ok varo par engi motstaoa. par la klaeoi blatt ok a tveir gullhringar ok svero gott. par varu ok tvaer konur; var onnur ung ok frio, en onnur gomul ok po frio. porgils gekk um husit, ok var vioa berg undir. Hann hafoi i hendi eina rotakylfu ok baroi henni a baoar hendr ok stokk flest undan. Porsteinn for meo honum; ok er peir gengu or jarohusinu, toku Peir konu Pi hina yngri ok fluttu meo ser til skipa ok sva hina eldri.⁴¹

Interestingly, a similar kind of device to hide the entrance to a jarohus is found in Reykdoela saga and Groenlendinga pattr, although there it is bushes (actually, simulated bushes in Reykdoela saga), not a tree, that conceal the entrance.⁴²

Middle English literature offers two quite striking parallels to the eoroscraefleordsele of the Wife's Lament. Sir Tristrem, a late thirteenth-century adaptation of Thomas's Tristan, describes an erpe house to which Tristrem and Ysonde flee in order to escape from King Mark; here too, the structure is located in a forest:

A forest fled pai tille, Tristrem and Ysonde he schene … In on erpe house Pai layn, Per hadde hai ioie y [nou.sub.3]. Etenes bi old dayn Had [wrou.su.b3]t it, wip outen [wou.sub.3]. Ich [ni.sub.3]t, sop to sain, Per til Pai bope [drou.sub.3] Wip miht. Vnder wode [bou.sub.3] Pai knewen day and [n.sub.3]t

(2454-55, 2478-86).⁴³

Although we do not know for certain what the poet of Sir Tristrem was adopting from Thomas's Tristan here, as the relevant passage in the Tristan does not survive, the Norse version of the Tristan, generally accepted as a very faithful rendering, describes a souterrain:

Ok sva sem peim likaoi nu petta frelsi i morkinni, sva fundu pau leyniligan stao hja vatni nokkuru ok i bergi pvi er heionir menn letu hoggva ok bua i fyrnsku meo miklum hagleik ok fagri smio, ok var petta allt hvalft ok i jorou til at ganga djupt hoggvit, ok var einn leynistigr langt niori undir; joro var mikil a husinu, ok stoo a sa friasti vior d berginu. …⁴⁴

Sylvia C. Harris has identified this structure as a souterrain, as have Rainer Gruenther and Herbert Kolb, although the latter do not use that term.⁴⁵ Moreover, the Scottish vernacular still refers to local souterrains as “yird-hoose,” “eird-Hoose” and other variations of “earth-house,” the term used by the Tristrem-poet.⁴⁶ The Old French Tristan probably described the souterrain in considerable detail, as the Norse version does,⁴⁷ but the Tristrem poet, who treated his source more freely than the Norse translator, felt no need to describe the structure in detail. This suggests that his audience would have had no trouble envisioning the erpe hous as the kind of souterrain described in the ON Tristramssaga.

A second Middle English parallel to the Wife's Lament is found in Lagamon's Brut. Locrin, having cast off his fiancee Guendoleine in favor of the foreign Astrild, is confronted by Guendoleine's angry father. Subsequently, Locrin is advised by his people to cast off Astrild, though he has sworn oaths of fidelity to her. The similarities to the narrator's situation in the Wife's Lament are tantalizing, and it is worth quoting the relevant passage in full:

per wes moni hah word. 7 ba alre seleste eoden to-sumne. hulden muchel husting. ba hehste of pan hirde. pat heo nolden ipolian, for al-peodene gold. pat peos laooe weore, bi-tuxe Corineo 7 Locrine. Ah we willeo raeden, 7 we wileo runan. pat weo nimen Waendoleine, Locrine to are quene. 7 halden alle vre aoes, bi-twux Corineo 7 Locrine. 7 halden hiredes luue, mid soo-fasten [hui.sub.3]e. 7 senden AEstrilde. vt of hissen londe. 7 Locrin pis biluuede, for hit wes his leodene read. He nom Guendoleine, 7 hefde heo to wiue, 7 he hit seide, soo beih hit nere. pat he wolde AEstrilde, senden vt of londe. Ah ne dude he nawiht swo, for swiken he pohte. ah he nom his enne hired-mon, pe he wel trowede on. 7 hehte hine swioe stille steolen vt of hirede. 7 hehte hine faren to pon tune, pe Trinouant wes ihaten. pe wes on vre leoden, Lundene ihaten. 7 par an [hi.sub.3]inge, purh-ut alle pinges. makian an eoro-hus. eadi 7 feier. pe walles of stone, he duren of whales bone. 7 pat inne swioe feire stude, from socne pes folkes. 7 dude ber-inne muchel col, 7 claoes i-nowe. pzelles 7 purpras, 7 guldene ponewaes. muchel win muchel wex, muchel wunsum ping. 7 seoooen foro-rihtes, wende al bi nihte. 7 mid stilliche ginne, brohte AEstrild ber-inne. pus dude Pes riche mon. swa Locrin hine hefde ihaten. for euer-ulc god mon, ah his lauerdes heste to don. Seouen [er.sub.3] wes Astrild, i hissen eoro-huse. pat neuer ne ferde heo wio-uten dore, ne na mon heo ber nuste. buten he king Locrin, 7 his iueren mid him.

(1161-93)⁴⁸

This passage is interesting not only for its description of the souterrain in which Astrild lives, but also because of the similar circumstances in which she and the narrator are put into their hiding places: after vows of undying love, she and her lover are separated through the interference of kinsmen. Lagamon took the outline of this incident from Wace's Roman de Brut, wherein the making of the underground dwelling is described as follows:

Dont a Locrins, par tel devise, Gendoliene DL feme prise, Mais il n'en a mie oublioge Estril qu'il ot LL sognantee; Par un, son bon familier, Fist a Londre faire un celier, Desos terre parfondement, LA fu Estril bien longement: Set ans la tint issi Locrin Celeement el sosterin.

(1419-28)⁴⁹

Just as the Tristrem-poet simply uses erpe house to describe the souterrain presumably described at length in his source, so Lagamon translates the phrase un celier, desos terre Parfondement as eorohus. Although Lagamon then adds six lines not found in Wace, giving an elaborate description of the construction and appointment of the eorb-hus, there is no hint aside from the term eoro-hus to suggest that this structure is an underground dwelling, as is made clear in Wace. Both Lagamon and the Tristrem-poet⁵⁰ seem to have understood eord-hus (or erpe house) to be a sufficient description of a souterrain not to warrant additional clarifying detail.

Of course, the Wife's Lament does not use the term eorohus to refer to the wife's dwelling, but rather eordscrwf and eorosele. However, the compound eordsele is clearly parallel in construction to eorohus. Both sele and hus denote “dwelling”; sele generally refers to a spacious dwelling, a hall, whereas hus usually denotes either a smaller dwelling or even a single room. Thus, the term eorosele may refer to a structure larger than but otherwise similar to an eorohus; the description of the narrator pacing geond pas eoroscrafu (l. 36) certainly suggests a sizable dwelling. In Beowulf,⁵¹ all three terms—eoroscraef (eoroscrafa, l. 3046), eorosele (ll. 2515, 2410), and eorohus (eorohuse, l. 2232)—are used by the poet to describe the dragon's lair, and certainly there the structure in question must be sufficiently large to accommodate Beowulf and Wiglaf fighting a dragon, as well as the hoard of gold. It is beyond the scope of this article to offer a comprehensive rereading of Beowulf's fight against the dragon, but the poet's description of the lair suggests that it may well be a souterrain. The most widely accepted interpretation is that the dragon's den is a tomb, but Beowulf never explicitly associates it with burial; indeed, the Lay of the Last Survivor mentions various kinds of treasure that are hidden in the eordhus, but not the bodies of those who owned it. If anyone is buried there at all, it is the last survivor himself, who unblioe hwe(arf) / daeges ond nihtes, oo oaet deaoes wylm / hran aet heortan (ll. 2268-70);⁵² still, we are not told that he died in the eorbhus. Lines 2717-19 may well describe a tomb, but they could also refer to a souterrain: [se aeoeling] seah on enta geweorc, / hu oa stanbogan stapulum faeste / ece eoroeced innan healde. In his note, Klaeber reads “The stone chamber is … contained in the ever enduring (or, primieval) earth-house,” and suggests that stanbogan refers to a kind of vaulting, while stapulas may denote the upright stones used therein. As a parallel to the dragon's den as a dwelling rather than a tomb, Joseph Harris has cited the description of Fdfnir's den in Fafnismal, which refers to a dwelling dug into the ground, though it is of iron, not stone; it has huroir af iarni oc gaetti; af iarni voro oc allir timbrstoccar i husino, enn grafit i ioro nior.⁵³ Interestingly, souterrains are frequently associated with hidden treasure, and the dragon's den is associated primarily with a hord;⁵⁴ of course, tombs also frequently contain treasure, but if the hord mentioned by Beowulf is intended as grave-goods, then it could only honor the last survivor himself. Since there are no striking material differences between tombs and souterrains, the essential distinction can be made only on the basis of function: if someone is buried in this structure, it is a tomb; if not, it is most likely a souterrain. In Beowulf, the textual evidence is inconclusive.

Though each of the texts discussed here differs in some aspects from the Wife's Lament, when taken as a whole they illuminate the circumstances described by the poem. In all but the Groenlendinga pattr and Reykdoela saga, the souterrain is used to conceal human beings from their enemies: in Volsunga saga, Sigmund hides from King Siggeir; in Hrolfs saga kraka, Hroar and Helgi hide from King Frothi; in the Gesta Danorum, King Regnald conceals his daughter from Gunnar; in Floamannasaga and Landnamabok, the Irish conceal themselves from viking raiders; in Sir Tristrem, Tristan and Isolt escape to the forest from Mark's court; and in the Brut, Locrin sequesters Astrild near London to keep her as his paramour. In Volsunga saga, Hrolfs saga kraka, Floamanna saga, and Sir Tristrem, the souterrain is explicitly said to be located in a forest. In several texts—the Brut, Sir Tristrem, Floamanna saga, the Gesta Danorum, and Gongu-Hrolfs saga—the souterrains are associated specifically (though not always exclusively) with women; the passage from Brut is especially interesting, since it depicts a woman who is separated from her lover through the scheming of others, after which her lover orders a souterrain constructed for her, in which she then lives.

These texts can help to illuminate the situation described in the Wife's Lament. The narrator, having followed her husband into a foreign country (ll. 6-10), must now endure separation from him through the machinations of his kinsmen (ll. 11-14). Her husband commands her to take up residence in a souterrain located in a forest (ll. 15-17, 27-32),⁵⁵ and there she bewails her fate (ll. 17-26, 32-41). The poem then closes with her pondering her husband's fate (ll. 42-53). just as the souterrain circumscribes the narrator's physical movement and becomes the dominant feature of her material existence, so too it embodies her two most prominent concerns: loneliness at the separation from her lover and concern about the hostility of her husband's enemies (be they his kinsmen or not). The motifs of loneliness and fear of hostility are associated with the eoroscraefleorasele. The juxtaposition between the kinsmen's dyrne gepoht (l. 12) and her lord's command that she should dwell in the herheard (l. 15) suggests a causal relationship, though this is not made explicit; in lines 25-28, this juxtaposition is repeated:

Sceal ic feor ge neah mines felaleofan faehou dreogan. Heht mec mon wunian on wuda bearwe, under actreo in ham eorbscrzefe.

Likewise, through the poignant contrast in lines 29-36, the dwelling becomes emblematic of the narrator's longing for her lord:

Eald is Pes eorosele, eal ic eom oflongad, sindon dena dimme, duna uphea, bitre burgtunas, brerum beweaxne, wic wynna leas. Ful oft mec her wrahe begeat fromsip frean. Frynd sind on eorpan, leofe lifgende, leger weardiao, ponne ic on uhtan ana gonge under actreo geond has eoroscrafu.

The narrator contrasts her own loneliness and isolation in the joyless dwelling to the lovers who are lying together in bed. Her grief at the departure of her lord, and her longing for him, infuse the eoroscrafleorosele and its surroundings with melancholy; when she walks through her dwelling at dawn, she imagines other lovers in bed, in each others' embrace. But she is lonely not only because her husband is away, but also because she has no friends at all here; she stresses that she walks ana through the dwelling, and in ll. 16-17 she says that she ahte … leofra lyt on pissum londstede, / holdra freonda. Considering the contexts in which souterrains occur in the Scandinavian, Irish and English texts cited above, associating the narrator's dwelling with the motifs of longing and fcar of hostility makes eminent sense.

Of course, the various parallels cited here should not obscure the essential difficulties that remain in interpreting the Wife's Lament. Saxo's Gesta Danorum, Sir Tristrem, Lagamon's Brut, Floamanna saga, the Annals of Ulster, Volsunga saga, Hrolfs saga kraka, Gongu-Hrolfs saga, Landndmabok, Groenlendinga pattr, and Reykdoela saga cannot fill in the narrative gaps which lend the Wife's Lament its enigmatic aspect. They can, however, suggest why the narrator should live on wuda bearwe, / under actreo in pam eoroscreefe (ll. 27-28), and the various contexts in which souterrains are found in these texts may frame a background against which the poem could be interpreted. However modest an increase in our understanding of the poem this represents, it brings us one step closer to understanding the situation in which the narrator finds herself.

Notes

1. For an overview of the scholarship on the Wife's Lament, and of the particular problems in interpreting the pocm, see Anne Klinck, The Old English Elegies (McGill-Queens U. Press, 1992), pp. 49-54, 177-88.

2. This and all subsequent references are from the Anglo-Saxon Poetic Records edition, ed. by George Philip Krapp and Elliott Van Kirk Dobbie (Columbia U. Press, 1936).

3. William C. Johnson, Jr. “The Wife's Lament as Death-Song,” in Martin Green, ed., The Old English Elegis: New Essays in Researck and Criticism (London: Fairleigh Dickingson U. Press 1983), p. 69; Emily Jensen, “The Wife's Lament's Eoroscraef: Figural or Literal Sign?” Neuphilologische Mitteilungen 91 (1990): 450.

4. While some critics have argued that the narrator is a man—see Rudolph C. Bambas, “Another View or the Old English Wife's Lament,” JEGP 62 (1963): 303-7, and Martin Stevens, “The Narrator of the Wife's Lament,” Neuphilologische Mitteilungen 69 (1968): 72-90—this view has not found widespread acceptance in the field.

5. “The Situation of the Narrator in the Old English Wife's Lament,” Speculum 56 (1981): 492-516.

6. Three Old English Elegies, (Manchester U. Press, 1961), p. 56.

7. Elinor Lench, “The Wife's Lament: A Poem of the Livin Dead,” Comitatus 1 (1970): 3-23. Raymonnd P. Tripp, Jr., “The Narrator as Revenant: A Reconsideration of Three Old English Elegies,” Papers on Language and Literature (1972): 339-61.

8. Earl R. Anderson, “The Uncarpentered World of Old English Poetry,” Anglo-Saxon England 20 (1991): 6-5-80. Joseph Harris, “A Note on eoroscraef/ eorosele and Current Interpretations of the Wife's Lament,” English Studies 58 (1977): 204-8.

9. See Jensen, passim.

10. Jensen states: “The interpretation … depends entirely on whether one reads eoroscroef as a literal sign, a word or thing that calls up its linguistic or historical associations with words and things in other texts, poetic and non-poetic; or one reads eoroscraef as a metaphoric sign, a poetic image that calls up its own immediate metaphoric contexts and, perhaps, other similar metaphoric contexts” (p. 450). One wonders why the eordscraef must be either metaphorical or literal (rather than both).

11. A curiously tenacious interpretation has been mentioned above, namely that the dwelling is a grave. For a convincing refutation of this theory, see Wentersdorf and J. Harris.

12. See Torsten Capelle, Die Archaologie der Angelsachsen: Eigenstandigkeit und kontinentale Bindung vom 5. bis 9. Jahrhundert (Darmstadt: Wissenschaftliche Buchgesellschaft, 1990); and Philip Rahtz, “Buildings and Rural Settlements,” in David M. Wilson, ed., The Archaeology of Anglo-Saxon England (London: Methuen & Co., Ltd., 1976).

13. Reallexikon der Germanischen Altertumskunde (Strassburg: Karl J. Trubner, 1913-15), 2:457.

14. Rudolf Much, ed., Die Germania des Tacitus, 3rd ed. rev. H. Jankuhn, Altertumskunde 5 (Heidelberg: Carl Winter, 1937), 3:179-81.

15. Much-Jankuhn, Germania, 172. (“They also have the habit of hollowing out underground caves, which they cover with masses of manure and use both as refuges from the winter and as storehouses for prroduce. Such shelters temper the keenness of the frosts; and if an invader comes, he ravages the open country, while these hidden excavations are either not to exist, or else escape detection simply because they cannot be found without a search.” Trans. H. Mattingly, The Agricola and Germania [London: Penguin Books, 19701, p. 115.)

16. Arnold p. 23.

17. See the note to lines 28-36.

18. For the most thorough description of souterrains in Scotland, see F. L. W. Thomas, “On the Primitive Dwellings of the Outer Hebrides,” Proceedings of the Society of Antiquaries of Scotland 7 (1867-68): 153-94; David MacRitchie, “Description of an Earth-House at Pitcur, Forfashire,” Proceedings of the Society of Antiquaries of Scotland 34 (1900): 202-14; and MacRitchie, “Earth-Houses and Their Occupants,” Proceedings of the Society of Antiquaries of Scotland 51 (1916-17): 178-987. For Irish examples, see MacAlister; Barry; and V. Gordon Childe, Prehistoric Communities of the British Isles (London and Edinburgh: W. & R. Chambers, 1940). For Danish souterrains, see P. V. Glob, Denmark: An Archaeological History from the Stone Age to the Vikings, trans. Joan Bulman (1967; Cornell U. Press, 1971); Gudmund Hatt, “Dwelling-Houses in Jutland in the Iron Age,” Antiquity 11 (1937): 162-73; and Axel M. Holmberg, Nordbon under Hednatiden (Stockholm: T. Theod. Bergelin, 1852). For a grief treatment of structures in Cornwall, see Charles Woolf, An Introduction to the Archaeology of Cornwall (Truro: D. Bradford Barton Ltd., 1970).

19. R. A. S. MacAlister, The Archaeology of Ireland, 2nd rev. ed. (London: Methuen, 1949), p. 273.

20. Ibid., pp. 272-73.

21. F. T. Wainwright, “Souterrains in Scotland,” Antiquity 27 (1953): 221.

22. T. B. Barry, The Archaeology of Medieval Ireland (London and New York: Methuen, 1987), p. 25.

23. Humphrey Welfare, “The Southern Souterrains,” in Roger Miket and Colin Burgess, eds., Between and Beyond the Walls (Edinburgh: John Donald Publishers, 1984), p. 305. I am grateful to Catherine Karkov for supplying this reference.

24. Sean Mac Airt, and Gearoid Mac Niocaill, The Annals of Ulster (To A.D. 1131), Part I (Dublin Institute for Advanced Studies, 1983), p. 318. (“The uam of Achad Aldai, and of Cnodba, and of Boadan's Mound above Dubad, and of Oengoba's wife, were searched by the foreigners—something which had never been done before. This was the occasion when three kings of the foreigners, i.e., Amlaib and Imar and Auisle, plundered the land of Flann son of Conaing; and Lorcan son of Cathal, king of Mide, was with them in this.” Trans. adapted from Mac Airt and Mac Niocaill; I have reinserted the original uam [= “grave,” “cave,” “souterrain”], which had been translated as “caves.”)

25. Barry, p. 26.

26. F. L. W. Thomas, p. 188.

27. Valdimar Asmundarson, ed., Islendingabok og Landnamabok, Islendinga sogur 1 (Reykjavik: Sigurour Kristjanson, 1909), p. 31. (“Leif went viking in the west to plunder. He harried about ireland and found there a great jardhus. He went into it, and it was dark there until light came from a weapon which a man held. Leif killed the man and took the sword and a great deal of money as well. Thereafter he was called Sword-Leif. He also harried widely about Ireland and got great booty there.”) This translation is by the author, as in all cases where no translation is cited.

28. See Cleasby-Vigfusson, Icelandic-English Dictionary (Oxford: Clarendon Press, 1874), s.v. jardhus.

29. Gesta Danorum, ed. Alfred Holder, (Strassburg: Karl J. Trubner, 1886), p. 240. (“But the King of the Northmen, now in extreme age, when he heard how the tyrant busied himself, had a cave made and shut up in it his daughter Drott, giving her due attendance, and providing her maintenance for a long time. … And, to prevent the cave from being noticed by its height, he levelled the hump down to the level ground.” Trans. Oliver Elton, The First Nine Books of the Danish History of Saxo Grammaticus [London: David Nutt, 1894!, p. 288.)

30. P. 241. (“Hence, while he was himself conducting the search with others, his doubtful car caught the distant sound of a subterranean hum. Then he went on slowly, and recognized a human voice with certainty. He ordered the ground underfoot to be dug down to the solid rock; and when the cave was suddenly laid open, he saw the winding tunnels. The servants were slain as they tried to guard the now uncovered entrance to the tunnel, and the girl was dragged out of the hole, together with the booty therein concealed.” Trans. Elton, p. 289.)

31. pam eordscraefe (1. 28, acc. sing.) and pes eordsele (1. 29, nom. sing.), but pas eordscrafu (1. 36, acc. pl.).

32. As Leslie has noted in his edition, faehou is a technical term used to describe a state of feud and hence does not refer to hostility between the narrator and her lord (note to 1. 26).

33. Guoni Jonsson, ed., Volsunga saga, Fornaldar sogur Nordurlanda 1 (Reykjavik: Islendingasagnautgafan, 1981), p. 136. (“King Sigmund … gathered an army. But Hjordis, together with a bond-woman and great wealth, was driven into the forest. There Hjordis stayed during the fighting.”) Trans. Jesse Byock, The Saga of the Volsungs (U. of California Press, 1990), p. 52.

34. See note 41 below.

35. Gudni Jonsson, ed., Gongu-Hrolfs saga, Fornaldar sogur Nordurlanda 3 (Reykjavik: Islendingasagnautgafan, 1981), p. 267.

36. As mentioned above, the term jardhus could denote both a souterrain proper or an underground chamber (cf. note 28). However, the precise nature of the jardhus is less important than the general principle of women being sent into hiding a subterranean hideaway.

37. See Wentersdorf, and R. P. Orton, “The Wife's Lament and Skirnismal: Some Parallels,” in Or Dolum til Dala: Gudbrandur Vigfusson Centenary Essays, ed. by Rory McTurk and Andrew Wawn, Leeds Texts and Monographs 11 [New Series! (Leeds, 1989), pp. 206-37. Both rely heavily on supposedly pagan connotations of heargeard (MS her heard), although this is usually translated as “dwelling in a grove” (habitaculum in nemoribus)—thus C. W. M. Grein, Der Sprachschatz der angelsachsischen Dichter (Kassel, 1864), 2:60; J. R. Clark Hall, A Concise Anglo-Saxon Dictionary, 4th ed. with a supplement by Herbert D. Merritt (U. of Toronto Press, 1960). In his edition, Leslie emends to her eard niman, an idiomatic phrase which denotes “to take up one's abode” in Ps. 131, 15.3 (Par. Ps.) and Chr. 63. Wentersdorf envisions the narrator dwelling in a pagan sanctuary, and Orton compares the Wife's Lament to Skirnismal, and suggests that the narrative of the Wife's Lament may be “derived from a myth about an abandoned fertility goddess or Terra Mater who weeps (like Freyja) for a departed husband or lover, wanders in search of him, and is finally condemned by his withdrawal (like Gerdr) to a lonely, sterile, and death-like existence in a primitive, elemental world” (231).

38. P. 120. (“She went and met with her brother and they decided that he should make an underground dwelling in the woods. This went on for a while with Signy hiding him there and bringing him what he needed. King Siggeir, however, believed that all the Volsungs were dead.”) Trans. Byock, p. 42.

39. Jon Johanesson, ed., Droplaugarsona saga, Islensk fornrit 11 (Reykjavik: Hio islenska fornritafelag, 1950), p. 168. (“Now they crossed the river with Thorkel Trani and came to the place which was called at Bakki, east of the river, and they went into the cow-house and took a spade and a shovel and go away thereafter, out to Oddmarsloek west of the Eitha-wood. Beside the brook they dug out a jardhus for themselves and threw all the dirt into the brook. They wanted to have that hideout, in case they should have need of it.”) Throughout this essay, the Old Icelandic “o-hook” has been replaced by [angstrom] for reasons of typography.

40. Guoni Jonsson, ed., Hrolfs saga kraka, Fornaldar sogur Nordurlanda 1 (Reykjavik: Islendingasagnautgafan, 1981), p. 4.

41. Finnur Jonsson, ed., Floamannasaga, Samfund til Udgivelse af gammel Nordisk Litteratur 56 Copenhagen: 1932), p. 24. (“… came to Ireland during the summer; there was a forest before them, where they arrived; they went into the forest, and in one place they saw leaves fallen from a tree. They turned over the tree and found a jardhus underneath. They saw men with weapons below in the house. Thorgils offered his men this bargain, that he who went first into the house should own all the treasure, and all agreed to this except Gyrth. Thereafter Thorgils ran into the house, and there was no resistance. There lay a blue cloth and two gold rings on it and a good sword. There were also two women; one was young and beautiful, the other old but also beautiful. Thorgils went about the house, and there were rocks in many places underneath. He had in his hand a root-club and he wielded it with both hands and knocked most of it away Thorsteinn went with him; and when they went from the jardhus, they took the younger woman and carried her to the ship, and also the older one.”)

42. Finnur Jonsson, ed., Reykdola saga, Islenskar Fornsogur 2 (Copenhagen: Islenzka-Bokmentafelagi, 1881), p. 18. Einar Sveinsson and Matthias Pordarson, eds., Groenlendinga pattr, Islensk fornrit 4 (Reykjavik: Hio islenska fornritafelag, 1935), pp. 289-90.

43. George P. McNeill, ed., Sir Tristrem, Scottish Text Society 8 (Edinburgh and London: William Blackwood & Sons, 1886). (“They fled to a forest, Tristram and the beautiful Ysonde … They lay in an earth-house, where they had enough. Giants in old days had fashioned it, without doubt. Each night, to tell the truth, they went there with might. Under the boughs of the woods they experienced both day and night.”) I am grateful to Tim Jones for supplying this reference.

44. Eugen Kolbing, ed., Die nordische und die englische Version der Tristan-Sage Vol. 1 (Heilbronn: Gebr. Henninger, 1878), p. 79. (“And as they enjoyed the freedom in the forest, they found a secret place near a certain river, in a rocky hillside [cf. S. Harris, p. 311], which heathen men had caused to be hewn out and prepared with great skill and beautiful artifice, and it was all vaulted and the entrance hewn deep into the earth, and a secret passage led far into it; there was much earth on top of the house, and on top of the hillside grew a very beautiful tree. …”)

45. Sylvia C. Harris, “The Cave of the Lovers in the ‘Tristrams saga’ and Related Tristan Romances.” Romania 98 (1977): 306-30, 460-500. Rainer Gruenther, “Das wunnecliche tal,” Euphorion 55 (1961): 341-404. Herbert Kolb, “Das minne hus,” Euphorion 56 (1962): 229-47.

46. MacRitchie, “Earth-Houses,” p. 178; The Scottish National Dictionary, ed. William Grant and David Murison (Edinburgh: Scottish National Dictionary Association, 1952), s.v. “Erd-.”

47. See S. Harris, pp. 306-15.

48. G. L. Brook and R. F. Leslie, eds., Layamon: Brut, EETS 250, Vol. 1 (Oxford U. Press, 1963); Caligula ms. readings cited here. (“There were many proud words. / Then the most senior men strode to an assembly: / They held a special husting, the highest of those heroes. / Said they would not tolerate, just for alien treasure, / That there should be this quarrel between Corineus and Locrin: / ‘But we will give counsel, this advice we confer: / That we give to Locrin Gwendoline to be our Queen, / And keep all the oaths we made to Corineus and Locrin, / And preserve the people's love with most profound intent, / And send Astrild away, out of this country.’/ Locrin had to approve of this, as it was his people's advice: / So he took Gwendoline and had her as his wife, / And he told them this (though it was not true) / That now Astrild would be sent right out of the country. / But this he did not do at all, planning to deceive them, / But took one of his retinue, whose loyal faith he knew, / And commanded him most secretly to steal away from the court, / And commanded him to go to the town then called Troynovant / (Which in our language was given its present name of London) / And there, in great haste above everything else, / Fashion an earth-house, attractive and fine: / The walls made of stone, the doors of whale-bone, / And make it in a pleasant place, away from people's prying. / And put inside plenty of coal and sufficitent clothing: / Coverlets and purple cloths and plenty of golden coins, / Plenty of wine, plenty of wax and plenty of welcome things. / And then afterwards straight, moving always at night, / And by secret devices, convey Astrild inside it. / All this did the noble man, as Locrin had instructed him / (After all, every good man ought to do his lord's command!) / Seven years was Astrild in that same earthhouse, / Without ever going out of doors or anyone knowing her in there, / Except for King Locrin and his companion with him.” Trans. Rosamund Allen, Lawman: Brut [New York: St. Martin's Press, 1992], pp. 31-32.)

49. Wace, Le Roman de Brut, ed. Edouard Frere, Vol. 1, (Rouen, 1836). (“Locrin by that device took her as his wife, but he did not at all forget Estril, whom he had as his concubine. By a trusted servant he had had made in London a hiding place, deep underneath the earth, and there Estril was for a long time. For seven years he had her there, secretly in the souterrain.”)

50. Some critics have suggested that Thomas, the Tristanpoet, drew on Wace's description of the underground hideaway for his depiction of Tristan and Isolt's underground chamber. However, Sylvia Harris (pp. 475-98) has demonstrated that Thomas probably had other sources for his description of the southerrain, even if Wace's use of the motif may have influenced him to include it in his own work.

51. Beowulf and the Fi t at Finnsburg, ed. Fr. Klaeber, 3rd ed. (Lexington, Mass.: D.C Heath and Company, 1950).

52. This is an interesting parallel to the terms in which the narrator of the Wife's Lament describes her existence: … ic on uhtan ana gonge / … geond pas eordscrafu. / Paer ic sittan mot sumorlangne daeg, / paer ic wepan mag mine wraecsipas, / earfopa fela (11. 35-39).

53. P. 208, postscript to fn. 8. Fafnismal cited from Gustav Neckel, ed., Edda: die Lieder des Codex Regius nebst verwandten Denkmdlern, 5th ed. rev. by Hans Kuhn (Heidelberg: Carl Winter, 1983), 44 prose, p. 188 (“The gateposts and gates were iron, as were all the timbers in the house, which was dug down into the earth.”)

54. Reykdoela saga (p. 18) and Groenlendinga Pattr (p. 289) mention provisions Landnamabok (p. 31) describes precious sword and a hoard of money, Floamanna saga mentions golden rings and a sword (p. 24), and Saxo speaks of treasure and swords (pp. 240-41).

55. Because of the reference to burgtunas (1. 31) and a wic (1. 32), some critics have thought that the dwelling is located in an abandoned settlement. However, burgtunas is probably a variation of the preceding duna uphea (1. 30), since the hills near the dwelling resemble walls around a city. Likewise, wic need not denote a settlement, gut can rather suggest a solitary dwelling as well.