With these promises of good things, it could be tempting for first-time users of to go immediately to the entries. But in reality the first place to go is the , which contains a Its function is to explain the editorial guidelines used by the compilers, the structure of the entries, and the standard symbols and abbreviations. Presumably the reason for publication of the as a separate booklet is the expectation that it will undergo constant modification as publication of the main volumes proceeds and indeed it has already burgeoned from 122 pages on first printing to 190 in association with Vols 1 and 2 (Cathey 2002 413) and to 229 in association with Vols 1, 2, and 3.

Each entry, then, as carefully detailed in the comprises what are termed a lemma, a body, and a tail. The lemma naturally starts with the headword. A helpful feature, considering the comparatively large number of homographs in the language, is the addition of prefixed superscript numbers to distinguish them: thus, for example, 1eisa gløder // embers, from 2eisa ?fare voldsomt frem // ?rage forth. The orthography of headwords has been normalized to represent the language of Norway and Iceland ca. 1200-1250; if there are two deviating forms the more conservative (usually Icelandic) is used. Thus we expect to look for berg, not bærg (cf. Noreen 97 §108). This policy affects the sequence of letters at the end of the alphabet—þ, ǽ, ǿ, ǫ—which may disconcert first-time users when they confront ǽ and ǿ instead of æ and œ respectively. So marked a shift from more familiar normalization practices is defended by the editors on the grounds of its greater fidelity to the spelling of the older manuscripts and soundness practically and pedagogically. Certainly it makes for a clear indication of vowel length and connects with the well-known treatment of the phonology in the ( 218-221), followed by Adolf Noreen in his standard manual (38-39, §§29-31).

In the alphabeticization vowels with and without acute accents are treated as the same letter, so that dúfu-ungi young dove immediately precedes duga make an effort. Also treated as the same letter are d and ð, so that, for instance, eðalborinn ?ædelbåren…// ?noble by birth…, edd bogstavnavn på kapilælen d…// name of the majuscule (small capital) d …,” edda oldemor // great-grandmother and (proprial.) (navn på værk af Snorri Sturluson // name of a work by Snorri Sturluson), edik eddike // vinegar, ediksgerð tilberedning af eddike // preparation of vinegar, and eðilborinn ædelbåren…// noble by birth and rank… occur in succession (the last-named and first-named being cross-referenced—and one might wonder, incidentally, at the difference in glosses). This intermingling of d and ð is potentially confusing for anyone not experienced in reading manuscripts (cf. Adams 334) and lacks precedent in such authorities as Noreen (1923). Given the inherent inconsistencies in any system, one would have to weigh the benefits of the orthography chosen for over a system closer to that of () or other editions that most readers will be using.

As the next component in the lemma, salient aspects of morphology are noted and this documentation is fuller than in earlier dictionaries, albeit couched in a highly compressed format that may seem delphic if one fails to consult the Then, at least in some entries, supportive quotations, references to secondary literature, and editorial comments are inserted. For instance, where manuscript forms could be confused, as in some forms of bregða and bresta, note is made to that effect. Absent from the lemma (and indeed from the entry as a whole) are suggestions as to possible cognates and etyma. Given the abiding uncertainties of etymologizing and the availability of De Vries’s etymological dictionary, restraint is clearly the best policy. At the same time, total purity of methodology is made difficult by citations that themselves contain commentary on or explanations of words. In the interesting case of dǿgn and dǿgr, the citations include one from the where Sigvatr’s use of dǿgn is interpreted as a nonce-variant of dǿgr. Accordingly, the two articles might have been cross-referenced to advantage, as is done with other doublet forms, such as berg and bjarg. Etymology creeps in in a different way when, as noted by James Cathey (1998: 288), weak verbs are classified in the (2004: 26-27) on a diachronic basis, with reference to their categories in proto-Germanic, respectively ō, ja, ē. Yet strong verbs, including the preterite-present type (exemplified by superscript number 2-->2eiga), are not categorized at all, and the classification of nouns is purely synchronic, in terms of gender (masculine, feminine, neuter) and the declensional types weak versus strong ( 2004: 20-23). Perhaps these inconsistencies, though minor in the great scheme of things, reflect some wider uncertainty as to how far to admit the diachronic dimension.

Between lemma and body of the entry, in the case of semantically complex words such as bregða, a brief table of contents for the body is given. This, it has to be said, marks a major advance in format over Fritzner, whose fine-grained distinction of sub-senses is obscured by the close-packed lines of Frakturschrift.

The body of the entry begins with glosses. The policy is to offer a translation equivalent rather than the formal or encyclopaedic definition one would expect to see in dictionaries for native speakers. English-language glosses appear immediately after the Danish ones, as if on an equal footing, but the amply clarifies that in some respects the English definition is secondary in relation to its Danish counterpart. Some editorial comment, bibliographical references, and cross-references are presented in Danish only, which makes sense as a means of controlling the sheer bulk of . Another reason for not confining one’s attention to the English glosses is that occasionally they give the impression of being hasty afterthoughts and a few seem not to have been checked by a native speaker (thus, s. eitr, spit by a dragon, translating spyet af drage, where spat would be correct). Clarity can also be an issue. For instance, bryggju-ker, explained in Danish as brokiste, tømret kasse ofte fyldt med sten til understøttelse af skibsbro/brygge, appears in English as log container often filled with stones to support a quay/wharf ; here wooden instead of log might have been clearer. The English in the could occasionally benefit from greater attention too. An instance is the translation of the section on where the imprecise phrases these details for bøjningsangivelsen and background material for anførte bøjningsformer obfuscate the point that is being made ( 2004: 16-17).

The glosses for many words are divided into a series of sub-senses. The various syntactic constructions into which the headword can enter are methodically exemplified within these sub-senses, as we see in the entries on bera, bregða, drepa, and other high-frequency verbs, some of which run to many pages. Each set of glosses is supported by a generous selection of citations. Here manuscript variant spellings are respected, as we see in a citation such as wæll ek lægherstað j sancti Haluarðz kirkiu neer mæistara Oghmunde bærsærk frenða minom, I choose a resting place in Saint Hallvard’s church near Master Oghmund berserk my kinsman. Anyone who finds such orthography obscure should study the information on phonology, graphology, and morphology given by primers such as those by E.V. Gordon, Michael Barnes, and Odd Einar Haugen before trying to use .

Citations are tagged with a siglum that pinpoints the specific edition or manuscript or, where different sections of the manuscript are attributed to different periods, the section of a manuscript from which it has been excerpted. The earliest attestation of each headword is always included in the entry, with an identifying icon. The key to the sigla is supplied in the Indices volume, where we also find datings of the manuscripts. This very compact and precise documentation puts the user in contact with the witnesses in a way not feasible for most scholars up till now. In addition, the citations take account of manuscript variants, and a telling example is the horse-descriptor in . This we find glossed fairly standardly as (om hest) mørk gråbrun med sort stribe langs ryggen // (of a horse) dark grey-brown with a black stripe along the spine, but with the added information that the adjective has a possible variant reading, b[leik]álóttr – something not evident in the edition, in the rival dictionaries, or in E.V. Gordon’s primer. The latter reading is glossed in its place in the dictionary (though without mention of ) as (om hest) lys…med mørk stribe langs ryggen // (of a horse) light-coloured with a dark stripe down the back.

Some words are marked by initial question mark, to show that they are of uncertain status, and here too users of are in a position to benefit from the very extensive excerpting from manuscripts. An instance is ?einkostr betingelse uden alternativ // condition with no alternative, which appears uniquely as a variant in a single manuscript. In such cases the questioned headword is followed by the presumed more correct form. A related category equally deserving of vigilance is the starred word, meaning a ghost word, resulting from a previous editor’s misreading, erroneous reconstruction, or adoption of an attestation in a late manuscript (where the text is now known also to be preserved in older, better manuscripts). Instances are eimun, which is traced back to an erroneous reading of emivn (= emjun) on the part of Jón Sigurðsson (thus correcting Fritzner); einasta, which is documented as appearing uniquely in an eighteenth-century paper manuscript (thus correcting ); and einkleyfr, which appears as a variant in an eighteenth-century edition but could not be confirmed by the editors of .

Some words are registered without full treatment, so that there is no gloss, but references to relevant literature are supplied, in particular other dictionaries that have treated the word. This type of headword includes non-assimilated foreign words, marked alien. Similarly, onomastic material is limited to bynames and names given to artifacts. Thus nicknames such as eikikrókr are listed, as cogn[omen], but not glossed. So too, with query, its more dubious variant eikikroppr. By contrast, blóðøx does get glossed, because it occurs once as a common noun, in addition to its application to Eiríkr blóðøx. The sword-name ekkisax is in effect glossed by the citation. In the case of edda, the meanings great-grandmother and name of a work by Snorri Sturluson are both mentioned, as we have seen, but not the theory that they might be identical.

A more drastic exclusion is that of poetic material. This decision goes back to the dictionary’s inception and was made on the grounds that the poetic vocabulary had recently been treated in Finnur Jónsson’s revised . It is certainly true that poetic vocabulary, especially the innumerable heiti, would have been an awkward and lumpy ingredient in the plan, but at the same time reliance on so late as 1939 seems strange. Finnur Jónsson’s , on which, as noted, the revised is based, had sustained damaging critique during the 1920s and 30s, not merely in the shape of polemics from E.A. Kock in his long-running but also in more reasoned and constructive criticism from other noted skaldic scholars such as Hans Kuhn (1937) and Konstantin Reichardt (1928).

As a compromise, registers words found exclusively in poetry with a marking as poet. and with references to the specialist dictionaries and scholarly literature but without glosses or citations. Thus, for example, agentives of the types deyðir, dolgstríðir, eldboði, and eldskerðandi, very few of which occur in prose, are listed in with a bare reference to . This is much better than no documentation at all but remains unsatisfactory. Some ghost-words, for instance those that exist solely in the shape of emendations made by Finnur Jónsson and his predecessors, could have been eliminated to great advantage. The entry s. brandálfr is a case in point: at the very least it should be signalled as a conjectural or dubious form.

The tail of the entry contains information on compounds, other dictionaries or glossaries that have treated the headword, and secondary works consulted in writing the entry. Thanks to the availability of the register of compounds, they receive far fuller documentation (especially where the lemma constitutes the second or later element) than in earlier dictionaries. The secondary works include such standards as but also the volumes of the as they emerge, and the editorial team has been commendably thorough in keeping up to date with recent publications.

In the remainder of this article, it is upon the glosses and the citations that I shall be concentrating. The sub-senses and glosses often represent a complete rethinking of the evidence, as compared with earlier dictionaries. Moreover, the provision of citations is so full that the user has the materials at hand whereby to evaluate and critique the gloss where necessary. Looking at an extremely high-frequency grammar word, ef if, we see that the thorough demarcation of sub-senses is especially helpful in showing how uses of ef shade off into those of at and er, with inclusion of citations that illustrate variations between these words in the manuscripts.

When we ponder lexical words, the fine discriminations and rich documentation in will often help us to a more precise understanding than either or Fritzner. An instance in , ch. 4, is einhleypingr: “Sámr mælti: ‘Hvárt ertu goðorðsmaðr?’ Hann kvað þat fjarri fara. ‘Ertu þá bóndi?’ sagði Sámr. Hann kvazk eigi þat vera. Sámr mælti: ‘Hvat manna ertu þá?’ Hann svarar: ‘Ek em einn einhleypingr. Kom ek út í fyrra vetr. Hefi ek verit útan sjau vetr ok farit út í Miklagarð, en em handgenginn Garðskonunginum. En nú em ek á vist með bróður mínum, þeim er Þorgeirr heitir.’” 111 As translated by Hermann Pálsson, the exchange goes as follows: “‘Are you a chieftain?’ He said far from it. ‘Are you a farmer then?’ said Sam. He said he was not. ‘What kind of man are you then?’ said Sam. ‘I’m a wanderer,’ he replied. ‘I returned from abroad the year before last, after I’d been away from Iceland for six years and travelled south to Constantinople where I was in service with the Emperor of Byzantium. But now I’m staying with my brother Thorgeir.’” 49 Where Hermann offers I’m a wanderer, Terry Gunnell opts for I’m unattached 447 and Gwyn Jones for I am a footloose sort of man 102 . The latter has stronger literary appeal than its competitors, since it can be read as linking onwards in associative fashion to the famous episode of the boil on Þorgeirr’s foot. Nevertheless, all these translators appear to base themselves on the incomplete definitions to be found in the older dictionaries. typically takes his cue from the etymology: one who goes alone, hence a vagabond or person without hearth or home, cp. Scot. landlouper; Geir T. Zoega follows him with single person without hearth or home 340 and Gordon with unmarried man of no fixed abode, landloper; Fritzner has løs og ledig Person, der ikke er gift eller bosat unattached and free person, who is not married or settled in his own house (translations from Fritzner are mine throughout). Absent from these is the idea of an itinerant livelihood that we find clearly identified in ’s gloss, ugift person uden egen husholdning, løsarbejder // unmarried person without own household, jobbing worker. Also helpful are the accompanying citations in , which document the various connotations and associations of einhleypingr, and the adjacent entries on einhleypismaðr and einhleypr. The overall result is to lead us to a better appreciation of the episode. Þorkell is an einhleypingr, therefore free of responsibilities and reluctant to assume the goðorð, yet he is also a man of substance, in virtue of his itinerant livelihood, which lends weight to the words in which he impresses on Þorgeirr his brother where the family responsibilities lie.

A comparably emblematic word, which appeals to us intuitively as epitomizing a character or an episode in a saga, is eptirbátr in its context in , ch. 9 : Gunnlaugr mælti: ‘Hvar kómu feðr okkrir þess,’ segir hann, ‘at faðir minn væri eptirbátr fǫður þíns, hvar nema alls hvergi? Skal ok svá með okkr vera’ 3 80 . Gwyn Jones translates: ‘And what occasion was that,’ demanded Gunnlaug, ‘when our forefathers so set forth that my father was in tow to yours? Never in this whole wide world! And that is how it must be with you and me’ 195 . Here ’s gloss, laggard, is quite misleading, and it would seem that Jones has followed either Zoega, who improves on with a boat in tow…, [someone’s] inferior, or Fritzner, who offers for the figurative sense vera eptirbátr e-s: staa tilbage for en saa at man lader ham faa Forrangen stand behind somebody so that he is allowed to take precedence. glosses with greater attention to the metaphor: person som følger (i ngns kølvand), person som står tilbage (for ngn), ringere person (end ngn) // person who follows (in sby’s wake), person who is inferior (to another).

Another instance where helps us to understand the fuller semantic range of a word is búsýslumaðr. It offers orvalter af (arbejde på) gårdhusholdning // manager (of the work) on a farm as a first sub-sense, followed by a second sub-sense, marked with query: ?hushovmester // ?steward. has house-holder, husbandman and Fritzner Mand, som ved Iver og Duelighed styrer sin Husholdning man who with keenness and energy manages his household. To test these glosses against a well-known attestation in , ch. 1 : Svá er sagt, at Úlfr var búsýslumaðr mikill; var þat siðr hans at rísa upp árdegis ok ganga þá um sýslur manna eða þar er smiðir váru ok sjá yfir fénað sinn ok akra, en stundum var hann á tali við menn, þá er ráða hans þurftu; kunni hann til alls góð ráð at leggja, því at hann var forvitri. 1933 4 Bernard Scudder translates: Ulf is said to have been a very clever farmer. He made a habit of getting up early to inspect what his farmhands or craftsmen were doing and to keep an eye on his cattle and cornfields. Sometimes he would talk to people who were in need of his advice, for he was shrewd and always ready to make useful suggestions. 2001 8 For Scudder’s a very clever farmer Hermann Pálsson and Paul Edwards have an exceptionally able farmer 1976 21 and the glossary to Bjarni Einarsson’s edition has a very industrious farmer 2003 200 . Christine Fell’s version, It is said that Ulf was very efficient in the management of his estates 1975 1 , is clearly superior to all of these, because the word farmer in modern English is too limited in scope to fit the context. In Anglo-Saxon terms, it would be as if Kveld-Úlfr is acting as his own gerefa reeve . Here Fell’s characteristically painstaking analysis of terminology has enabled her to anticipate the semantic range indicated in .

A lexical item of great subtlety, semantically and culturally, is drengr. For this the primary sub-senses supplied in are as follows: 1) person af høj integritet, person af ære, brav person, modig/tapper person // person of integrity, person of honour, stalwart, courageous/brave person ; 2) ung mand (som ikke endnu har stiftet bu) // young man (who has not yet set up house) ; 3) (i tiltale til en ofte yngre mand // when addressing an often younger male) ; 4) (slags) menneske, fyr, skabning // (sort of) person, fellow, creature ; 5) (om mand i ngns tjeneste) svend, dreng, tjener // (of a male in sby’s service) boy, servant ; 6) (om barn) dreng // (of a child) boy ; 7) reb…// rope… Looking at the predecessor dictionaries, we find that ’s article suffers from being structured around a presumed etymology in drangr, whereas Fritzner (Vol. 1) is much better, methodically distinguishing six sub-senses: 1) Menneske der er som det bør være,” person who is as one ought to be ; 2) =drengskapr, 3) Karl i Almindelighed eller endogsaa i ufordelagtig Betydning, man in general or also in a pejorative sense; 4) ung, løs og ledig Mandsperson young, unattached, and unemployed male person; 5) Tjener, servant; 6) tyk Stok thick stick (the latter corrected in Vol. 4 to festetau).

The word drengr has recently been the subject of a sustained and valuable examination by Judith Jesch (216-232 et passim). If her study is not cited in , that is understandable in view of its publication date and scope (it confines itself to runic and skaldic material). Her methodology, however, seems worthy of emulation. The problem we face is that Snorri’s definitions of drengr, as of other key cultural words in his , probably embody prescriptive lexicography, based on the canonical work of the hǫfuðskáld classic skalds, and formulated for the benefit of novice skalds. To that extent, his explanations of vocabulary need to be treated with caution, as not necessarily reflecting contemporary prose usage. Sub-senses 1, 2, and 5 in all put citations from Snorri in pride of place and sub-sense 2, young man, depends on Snorri, since the non-Snorronian attestations do not appear to be decisive. To be fair, other citations are plentifully supplied in and can serve, where necessary, as a corrective.

As an additional example of the methodological issue here, consider the headword dolgr, which is demarcated into two sub-senses in . Under sub-sense 1, enemy, opponent, we find that citations are limited to three, one from the analysis of kennings in , where Þórr is termed the dolgr of the Miðgarðsormr, and two from . Here, once more, the citation can scarcely be accepted as unequivocally indicative of prose usage, since in context it appears to represent Snorri’s free citation from verse texts. The preponderance of citations is grouped under sub-sense 2, monster, ogre, devil, and it is these that seem to indicate normal prose usage, whereas sub-sense 1 apparently represents poetic, archaic, or otherwise heightened diction. Accordingly, chooses fiend as his main gloss and Fritzner (Vol. 1) reaches essentially the same solution, though he attempts a more elaborate categorization into devils, trolls, dragons, giants, ghosts, and human beings who behave like them.

Returning to drengr, and this time considering the analysis of the derived adverb drengiliga, we find a three-fold distinction of sub-senses in : 1) som det sømmer sig for en ‘drengr’, ærefuld, brav, loyal, tapper // in a manner befitting a ‘drengr’, honourably, nobly, fairly, loyally, courageously, bravely ; 2) (i kristen sammenhæng // in a Christian context) ; 3) mandig // manly; 4) flot // imposing. glosses as brave, bravely, without distinction of other sub-senses. ch. 2, offers another instance where translators will be helped by the more thorough analysis and documentation in . Bjǫrn, sonr Ketils, svarar: ‘Skjótt mun ek birta minn vilja. Ek vil gera at dœmum gǫfugra manna ok flýja land þetta; þykkjumk ek ekki af því vaxa, þótt ek bíða heiman þræla Haralds konungs, ok elti þeir oss af eignum várum, eða þiggja af þeim dauða með ǫllu.’ At þessu var gǫrr góðr rómr, ok þótti þetta drengiliga talat. 4 Keneva Kunz translates as follows: Ketil’s son Bjorn answered: ‘I can tell you at once what I want to do. I want to follow the example of other worthy men and flee this country. I see little honour to be gained in sitting at home waiting for King Harald’s henchmen to chase us off our lands, or even in meeting death at their hands.’ They applauded his words as being boldly spoken. 177 Kunz’s rendering of drengiliga, boldly, is in line with , but does it do full justice to the word in its context? Bjǫrn seems to be staking out a position whose comradely and consensual attributes are quite as important as its courage and spirit, and the analysis of this headword puts us in a position to appreciate those broader connotations.

In passages such as that above, a variety of nuances can be seen to be simultaneously active, making a formal distinction of sub-senses potentially counter-productive, and where a word is of that kind quite often abstains from listing sub-senses. Consider the treatment of drengskapr: handlemåde der sømmer sig for en ‘drengr,’ integritet, ridderlig opførsel, ære, ædelmodighed, storsindethed, redelighed, mandighed, bravhed, tapperhed // behaviour fitting a ‘drengr’, integrity, chivalrous behaviour, honour, nobility, noble-mindedness, high-mindedness, decency, goodness/worthiness, prowess, bravery. The absence of formally demarcated sub-senses is in line with the entries in previous dictionaries. , here, as often, playing the minimalist, offers simply courage, highmindedness, and Fritzner’s more general definition also eschews elaboration: Tænkemaade, Opførsel, der gjør en til et saadant Menneske, som han bør være mode of thought or behaviour that makes somebody into the sort of person he ought to be. Once again, has supplied by far the most forthcoming of the entries, as we can judge by considering a well-known passage in ch. 60: Þá segir Arinbjǫrn: ‘Ef þú, konungr, ok þit Gunnhildr hafið þat einráðit, at Egill skal hér enga sætt fá, þá er þat drengskapr at gefa honum frest ok fararleyfi um viku sakar, at hann forði sér; þó hefir hann at sjálfvilja sínum farit hingat á fund yðvarn ok vænti sér af því friðar; fara þá enn skipti yðr sem verða má þaðan frá.’ 184 Fell translates : If you, Sir, and you, Gunnhild, have decided between you that Egil is to get no reconciliation here, then the decent thing to do is to allow him truce and permission to travel for the length of a week, so that he may save himself. Yet he has of his own free will come here to meet you, and hoped for peace for himself from this. Your dealings from then on can go as they will. 1975 107 Neither of the sub-senses registered by fit here and Fritzner is indicative in only the broadest possible fashion, whereas ’s nobility and decency are both directly applicable, as is illustrated by the independent decisions taken by translators: Fell’s the decent thing compares closely with Hermann Pálsson and Paul Edward’s a matter of decency 1976 157 and with the English glossary to Bjarni Einarsson’s edition of , which offers decent behaviour, honourable behaviour. Scudder translates as the noble course of action 2001 113 , which seems appropriate in an address to royalty.

Another case where a fine teasing out of sub-senses appears to be resisted is einræði (neuter form), where has egenrådighed, selvbestemmelse, egen beslutning // wilfulness, intransigence, own decision. emphasizes the pejorative connotations with self-will, obstinacy. Fritzner’s gloss is more neutral: Beslutning, som grunder sig alene paa den besluttendes egen Mening eller Vilje decision that is based solely on the decision-maker’s own intention or wish. The following passage from , ch. 2, not cited by , illustrates how in actual usage affirmative and pejorative connotations can be nicely poised: Sannspurðan hefi ek fjándskap Haralds konungs til vár; sýnisk mér svá, at vér munum eigi þaðan trausts bíða; lízk mér svá, sem oss sé tveir kostir gǫrvir, at flýja land eða vera drepnir hverr í sínu rúmi. Em ek ok þess fúsari, at hafa slíkan dauðdaga sem frændr mínir, en eigi vil ek yðr leiða í svá mikit vandkvæði með einræði mínu, því at mér er kunnigt skaplyndi frænda minna ok vina, at þér vilið eigi við oss skiljask, þótt mannraun sé í nǫkkur at fylgja mér. 4 Kunz opts for the affirmative meaning, translating as follows: Of King Harald’s animosity to us there is proof enough; it seems to me we should expect little friendship from that direction. We seem to have two choices before us: to flee the country or to be killed off, one by one. Although I would prefer to meet my death as my kinsmen have done, I do not wish to make a decision on my own which will make things difficult for all of you. I know only too well the character of my kinsmen and friends: you would not want us to go our separate ways despite the trials that following me would involve. 176-177 While Kunz’s rendering decision on my own is good, it is conceivable that simultaneously the speaker is citing a potential criticism of his judgement as exhibiting wilfulness; in other words, both affirmative and negative connotations may be active in the context.

As Anatoly Liberman observes, speaking of lexicography in general, division into senses is always partly arbitrary, save for the most obvious cases 117 , and it might be that just occasionally distinguishes sub-senses too minutely. In the case of the compound dómhringr, the division into two sub-senses, one meaning a ring of stones or posts within which the judges are stationed, and the second meaning the ring formed by the judges themselves, is undermined by the lack of firm attestations to demonstrate the latter sub-sense (a problem duly acknowledged by ). and Fritzner give only the former.

The simplex dómr gets divided into altogether thirteen sub-senses, special idioms that pertain to individual sub-senses being noted within these categories. The sub-senses are as follows: 1) dom, kendelse, afgørelse, voldgiftsafgørelse // judgement, verdict, ruling, finding, award/decision by arbitration ; 2) (om den yderste dom på dommedag // of the Last Judgement on Doomsday) ; 3) (Guds) bestemmelse // (God’s) ordinance ; 4) straf // punishment/sentence ; 5) idømt) bøde // (imposed) fine/penalty ; 6) beslutning // decision ; 7) bedømmelse, vurdering, mening // assessment, opinion ; 8) domstol, ret, retsafhandlinger, rettergang, proces, retsmøde, sted/tid for retshandlinger // court, legal proceedings, court session/hearing, place/time for legal proceedings ; 9) nævn af domsmænd, dommerpanel // panel of judges ; 10) domsmyndighed, jurisdiktion // judicial authority, jurisdiction ; 11) styre, regering // rule, reign ; 12) tilstand, ?skæbne // condition, ?destiny ; and 13) ?(værdifuld) genstand // ?(valuable) object. These discriminations are exceedingly fine, and the editors signal difficulties in distinguishing sub-senses 8 and 9.

One or two queries about the sequencing of information in the entry on dómr also surface. Ian McDougall comments on the general system in in the following terms: It is surprising to see stated in the introductory matter that ‘no attempt is made to arrange the senses in a semantically orientated hierarchical structure’…The editors state that the ‘meaning…regarded as primary is as far as possible given first’. Although it is not always clear what the editors consider a ‘primary meaning’, most entries are presented, as one would expect, with senses arranged from concrete to abstract, from the most general to the more specialized. 2003 96 This is the arrangement we see with the verb drepa, for instance, yet the entry on dómr is sequenced so that some generic sub-senses, such as decision and assessment, follow more specific ones, such as (God’s) ordinance. In the case of sub-sense 13 it might be helpful if the undoubted attestations could be given first and the dubious example (the sole attestation not accompanied by the adjective heilagr) following them. Otherwise the initial question mark could be taken as reflecting doubt on the validity of the entire sub-sense.

Despite these incidental demurrals, the treatment of dómr represents a great advance on the lexicography of , which again adopts a minimalist approach to the distinction of sub-senses: 1) a court of judgement, the body of judges, or the court itself, 2) doom, judgement, sentence. The invocation of the word doom, testimony to a fatal fascination with etymology on the part of Guðbrandur, is likely to prove misleading, semantically speaking. Fritzner is more empirical and methodical, distinguishing seven sub-senses: 1) Beslutning, Bestemmelse, Erklæring, der er Udtrykket af ens Vilje eller Forgodtbefindelse decision, determination, declaration, that is the expression of someone’s wish or judgement; 2) Dom, Retskjendelse der afgives af den eller dem, som have at dømme, stifte Ret i en Sag judgement, verdict delivered by a person or persons who have the task of adjudicating or establishing the law in a case; 3) Retslig Behandling af en Sag og dennes Afgjørelse ved Dom legal handling of a case or its determination with a judgement; 4) Ret, Dommere der have at afgjøre en Sag, et Tvistemaal court, judges who have the task of deciding a case, a dispute; 5) Stedet hvor Ret holdes, de dømmende have sit Sæde the place where a court is held, the judges have their seat; 6) en Tings Væsen, Stilling, Tilstand med hvad dertil hører the essence, situation, condition of a thing, with what pertains to it; 7) Ting, Gjenstand, i Forbindelsen heilagr dómr thing, object, in the collocation heilagr dómr. To this Vol. 4 adds further citations under sub-senses 2, 3, and 4 but no new sub-senses. It is notable that the equivalents of Last Judgement and God’s ordinance are not given the status of separate sub-senses in Fritzner’s treatment.

’s glossing policy, insofar as it favours the translation equivalent over the encyclopaedic explication, leads to cryptic results with certain lexical items relating to mythology, law, technology, and other specialist fields. An example is alfr, which is defined simply as alf // elf, without discussion of any kind as to what mythological entity or entities might be so designated. Neither is there any attempt to distinguish native Scandinavian from foreign matter in the arrangement of citations. This is admirably non-encyclopaedic, but it does rather invite users to fill the gap with whatever personal knowledge of elves they might already possess. If we contrast the two major predecessors, we see that ’s entry is definitely in the encyclopaedic mould. It expatiates on cognates of the word in English, German, and other languages and on distinctions between ljósálfar and dǫkkálfar, álfar and dvergar; some of the writing is highly speculative, without being signalled as such. The assurance is offered, for instance, that while in old writers the Elves are rarely mentioned, it is clear that the same tales were told as at present. Modern Icelandic lexis (e.g., álfaskapr) is freely introduced, and conversely a medieval term such as alfa-vǫlkun illness inflicted by elves does not appear, either here or in the Supplements; alfa-vǫlkun is also missing from Fritzner (Vol. 1), but it appears in the supplement (Vol. 4), as also in . Perhaps surprisingly, Fritzner is fully as encyclopaedic as . He begins by explaining what an elf is, viz, in English translation, a being that was regarded as distinct from human yet at the same time partially of a human type, so that elves not merely appeared on occasion in human guise but also human beings could father children on elfish women. He equates medieval conceptions of elves with those current in the Norway of his times, noting that elves can sometimes be seen riding as a company, that elf-woman seek help from human women when close to childbirth, that they own livestock as humans do, and that they can make themselves visible or invisible at will. That elves could be regarded as the dead, or vice versa, is supported with reference to and .

Rather than overtly stepping into the role of an encyclopaedia in that way, sometimes incorporates explanations and commentary that occur in the primary texts themselves. Consider the handling of two legal terms. The term duradómr or dyradómr is glossed as ret afholdt ved ngns dør // court held at sby’s door, and placed in prefatory fashion within the Danish part of this gloss is an illustrative quotation from the Gulaþing law that explains the concept even though it does not contain the headword: nu scal dom setia firi durum verianda. en eigi a bak husi now a court must be set in front of the defendant’s doors, and not at the back of the house. This quotation, combined with citations that do contain the headword, enables the editors to avoid the kind of interpretive gloss used by , court at the door of the defendant. In the case of the term dómhringr, the corresponding explanatory quotation explains how such a ring was set up, preparatory to the holding of the court. Whether the placement of these quotations, awkwardly sandwiched between the Danish and the English glosses, is ideal is another matter. Maybe they would be better positioned after both the glosses and before the citations proper. Translating terms such as these into natural English (or Danish) often remains an insoluble problem, and the editors’ translation equivalent obviously does not purport to be a mot juste capable of direct use by an actual translator. For dómhringr has court circle, which hardly seems felicitous. For duradómr Hermann Pálsson and Paul Edwards, aided by context in , translate simply if unidiomatically as door-court 1973 51 .

At times, on the other hand, allows the encyclopaedic element freer play, an example being the entry for dís. As sub-sense 1 offers dis, overjordisk kvindeligt væsen (med egen kult, cf. Mundal 1974 81), (ngns) skytsgydinde, fylgje” // “supernatural female being (with its own cult), (sby’s) guardian spirit, fetch. Sub-sense 2 is as follows: heiti før kvinde (søster?) // heiti for a woman (sister?), with a reference to Faulkes 1998 257. The glosses for sub-sense 1, reinforced by the reference to Else Mundal, expand into explanatory mode, but within carefully maintained constraints, marking a clear difference in policy from previous dictionaries. To the sub-sense a goddess or priestess(?), a female guardian angel, appends a further explanation, who follows every man from his birth, and only leaves him in the hour of death, and continues expansively with references to and other texts. It is left unclear whether these passages truly exemplify the dís, as distinct from related figures such as the fylgja and hamingja. Fritzner (Vol. 1) glosses dís as Kvinde i Almindelighed woman in general, before moving to the more specific sense of Kvinde som kommer fra en anden Verden, hvori hun ved Døden er indgangen woman who comes from another world, into which she has entered upon death, expatiating from there into a detailed and somewhat speculative analysis of the different functions of dísir. The supplement to Fritzner continues in the encyclopaedic tradition: med et fellesnavn kalles valkyrjer, norner, hamingjer, fylgjer dísir, undertiden også enkelte gudinner, særlig Frøya (Vanadís); egentlig visst navn på en egen gudeflokk valkyries, norns, hamingjur, and fylgjur are referred to under the common designation of dísir, as also on occasion individual goddesses, especially Freyja (Vanadís); properly it was certainly the name of a specific group of gods.

Although, as I say, maintains higher standards of accuracy than its predecessors, there are a few entries that puzzle, one of them being that for dísarsalr. Included in the citations is this one from : Aþils konvngr var at disablóti ok reið hesti vm disarsalinn (cf. 57-58) King Aðils was at the sacrifice to the dísir and rode his horse round the hall of the dís. cites only a form with genitive plural, dísasalr, glossing the temple of the dísir. Fritzner Vol. 1 cites the form actually attested in the manuscripts, dísarsalr, with a carefully agnostic gloss: Bygning med en vis særegen Bestemmelse eller af en særegen Beskaffenhed, hvorom det er vanskeligt at komme til sikker Kundskab, uagtet derved paa de fleste Steder synes ment en Fruerstue, Kvindestue building with a certain special purpose or of a special character, concerning which it is difficult to obtain definite information, although in most contexts a ladies’ or women’s room appears to be meant. In the Fritzner supplement (Vol. 4), the gloss is altered to the more decisive sal der disene, kanskje særlig Frøya, ble dyrket hall where the dísir, perhaps particularly Freyja, were worshipped. In we seem to have a continuation of the Fritzner supplement, though Freyja is left out of the picture: disesal, sal til dyrkelse af diserne // hall used for worship of the dísir. But it is not clear how the distinctively singular form of the genitive-case noun dísar can be disregarded, unless the various editors have been influenced by the genitive plural in dísablót. The singular form should be retained, as hinting at the existence of a particular, tutelary dís who presides over the space and over other dísir.

The treatment of berserkr is another instance in where the usual editorial standards appear to falter. The categorization into sub-senses is as follows: 1) bersærk, kriger med særlige egenskaber (cf. berserksgangr sb. m.) // berserk, warrior with special qualities. (Should qualities read characteristics or attributes?) 2) kriger, stridsmand, kæmpe // warrior, champion. 3) saracener, person tilhørende fremmed hedensk folke-slag (cf. serkr sb. m.) // saracen, person of foreign heathen descent. The citations used to support sub-sense 1 consist of unequivocal descriptions of classically berserkr behaviours or accoutrements. Other less decisive attestations are relegated to sub-sense 2, but that is questionable, in that some of them would fit happily under sub-sense 1, since evidently they too refer to the classic berserkr type, and the only difference is that they do so in less detail. Sub-sense 3, for its part, seems over-specific: as Liberman remarks, usually the third sense is merged with the second 117 . It does not figure in Fritzner, and the suggestion of an association with serkr in the sense of Saracen appears dubious.

The editors’ decision to supply references to relevant secondary literature within entries like this one, with the Fritzner supplement (Vol. 4) as a precedent, has led to some discussion among reviewers. Liberman (117) argues, using berserkr as an example, that references should have been left out altogether except for those pertaining to the form, transmission, and meaning of a word, insofar as the semantic analysis is not trivial. On the other side of the argument, Ian McDougall contends that dictionaries without such references deprive readers of easy access to more detailed study of words and their contexts, and condemn the unfortunate entry-writers who work on them to the always thankless and often impossible task of reducing to a few words of definition arguments and explanations which other commentators have needed the space of one or more articles to treat in any adequate way. 2003 96 While one can only sympathize, a practical qualifier on this score is that reliance on secondary references to clarify glosses will reduce the utility of the dictionary in less-resourced centres of study. Many libraries have experienced reductions in funding for periodicals and reference works over the last few decades, and publications like , to which, as we have seen, often refers, have never been particularly easy to access outside the Scandinavian world.

The handling of technological terms also contains its small share of vagueness and inconsistency. This can be illustrated by the entry for eggvǫlr, where the English gloss is bulge alongside a cutting edge, in relation to an axe. The term sharpening bevel, bevel face might be more in keeping with specialist craft vocabulary. We can compare McDougall’s comments (2003 97) on the entry for brunnvaka.

The term blóðrefill is an interesting case. gives the point of a sword, Fritzner Sværdspidse. is more expansive, with sværdspids, sværdblad (damasceret = forsynet med mønster som ‘refill’? cf. Liestøl 1951 75-77 // point of a sword, sword-blade (pattern-welded = decorated with a pattern like a refill), but the relevance of the words in parentheses is not spelt out and could be formulated more clearly. If we consider the simplex refill, its dominant set of attestations relates to hanging tapestries. A rare compound tannrefill denotes a toothed file. The compound refilstígr, contextually translated as hidden, mysterious path , s.v. , appears to mean literally an entrenched path, path along a shallow dip in the terrain, comparable to the English toponymic hollow way. From these attestations, the core meaning of refill may be deduced as strip, stripe, groove. In swords of the Viking Age and later a groove (technically, a fuller) is seen on each side of the blade. The purpose of the fuller was to reduce weight and enhance the strength and elasticity of the weapon (Pedersen 593). The central flat or fullered face of the blade ran from the base adjacent to the hilt to within a few centimetres of the tip. It might feature pattern-welding, popular motifs being straight herringbone and twisted bands alternating with straight areas (Jones 7-11). Blood would naturally flow along such a groove. Thus the word blóðrefill would mean literally blood-groove or blood-strip(e). It would, strictly speaking, if this explanation is correct, denote neither blade nor tip per se, although from the core sense a vaguer sense of flat of the blade (virtually all the way to the tip) might well have developed.

Amongst the literary attestations, a reasonably clear example of the sense I am suggesting appears in , ch. 2: Þorsteinn tók við sverðinu ok brá þegar, tók blóðrefilinn ok dró saman milli handa sér, svá at uppi lá blóðrefilinn við hjǫltin; þá lét hann aptr hlaupa, ok var þá ór allr staðrinn 132 Þorsteinn took the sword and drew it immediately, took the blóðrefill and brought it together between his hands so that the blóðrefill lay against the hilts. Then he let it spring back and it had lost all its elasticity. Despite Jónas Kristjánsson’s note ( 132, n. 2), which identifies the blóðrefill as the tip of the sword, the mention of staðr elasticity will only make sense if it is the fuller or perhaps the entire flat of the blade that is meant here. Indicative in the same direction is a more figurative attestation in describing a comet whose blóðrefill points to Jerusalem; identifies this blóðrefill as the tail of the comet, no doubt meaning the ion tail, which (exemplified by those of Halley or Hale-Bopp) has a rayed appearance and sometimes even patterns of ropes, knots and streamers, all formations that might resemble the inlay patterns in a blóðrefill.

If we bear in mind these arguably decisive cases, other candidate instances are available for consideration. Two are contained in . I take first the famous scene in ch. 55 where Egill lays out for recompense from Aðalsteinn after Þórólfr’s death in battle: Aðalsteinn konungr sat í hásæti; hann lagði ok sverð um kné sér, ok er þeir sátu svá um hríð, þá dró konungr sverðit ór slíðrum ok tók gullhring af hendi sér, mikinn ok góðan, ok dró á blóðrefilinn, stóð upp ok gekk á gólfit ok rétti yfir eldinn til Egils. Egill stóð upp ok brá sverðinu ok gekk á gólfit; hann stakk sverðinu í bug hringinum ok dró at sér, gekk aptr til rúms síns; konungr settisk í hásæti. 1933 144 Scudder translates as follows: King Athelstan was sitting in the high seat, with his sword laid across his knees too. And after they had been sitting there like that for a while, the king unsheathed his sword, took a fine, large ring from his arm and slipped it over the point of his sword, then stood up and walked across the floor and handed it over the fire to Egil. Egil stood up, drew his sword and walked out on to the floor. He put his sword through the ring and pulled it towards him, then went back to his place. The king sat down in his high seat. 2001 90 Scudder coincides with Fell ( 1975 84) and Faulkes’s glossary to the Bjarni Einarsson edition in the interpretation sword-point. But a translation of blóðrefill that allows the reader to imagine the ring sliding beyond the tip, on to the area of inlaid groove, could bring out the idea, crucial to the context, that this is a large arm-ring and therefore munificent compensation. It might similarly serve the purposes of dramatic hyperbole in ch. 44 of : Egill kastar horninu, en greip sverðit ok brá; myrkt var í forstofunni; hann lagði sverðinu á Bárði miðjum, svá at blóðrefillinn hljóp út um bakit; fell hann dauðr niðr, en blóð hljóp ór undinni. 1933 110 Scudder translates as follows: Egil tossed away the horn, grabbed hold of his sword and drew it. It was dark in the doorway; he thrust the sword so deep into Bard’s stomach that the point came out through his back. Bard fell down dead, blood pouring from the wound. 2001 68 His interpretation of blóðrefill as the point is in line with Sigurður Nordal ( 1933 n. 1; cf. Hermann Pálsson and Paul Edwards 1976 129 and Fell 1975 84), but perhaps instead the sword-stroke is imagined as so forceful that not merely does the tip of the sword emerge from Bárðr’s back but also the fuller. In light of these two attestations in association with Egill Skalla-Grímsson, it would seem interesting, at the very least, also to register the occurrence of blóðrefill in the Egill attribution , where it is used as an apparent pars pro toto for sword. The policy of excluding the poetic corpus has had the effect, however inadvertent, of placing this attestation out of visibility.

This general policy seems regrettable on a variety of grounds. As McDougall (2003 95-96) observes, a specialized sense of a word well attested in prose might occasionally be best supported by a citation from poetry. It can also be the case that one or other aspect of the meaning of a word happens to occur in poetry but not in prose, without its necessarily being a purely poetic usage. We see this in the handling of bautasteinn, mindesten // memorial stone. While it is true that the entry for this word in contains a lengthy series of citations, none of them includes the key point that bautasteinar were customarily erected alongside roads. For that aspect is the classic source, but this text, because of its poetic form, lies outside the scope of . The same information can of course be gleaned from the supplemental references, notably Richard Perkins’s recent contribution (1999), but it would be nice to have documentation in the shape of an excerpt from a primary text, since that fits so well with ’s methods elsewhere.

A comparable case is that of einfǿtingr. glosses as person med én fod, person med én ben // uniped, person with one foot/leg, and here the English seems more precise than the Danish. It is also more precise than or Zoega, who offer simply one-legged man. Here, once again, Fritzner is better: menneske som kun har et Ben…især om sådanne, til hvis Natur det hørte a human being who has only one leg…especially one to whose nature this pertained. It is good that has also captured this distinction, albeit not formally divided into sub-senses, but it is strange not to have citation of the best-known attestation of the sub-sense uniped, the famous passage from the Vínland episode in , ch. 12. Instead, this passage is merely noted, in an example of what the editors term a blank reference. The text is as follows: Þat var einn morgin, er þeir Karlsefni sá fyrir ofan rjóðrit flekk nǫkkurn sem glitraði við þeim, ok œpðu þeir á þat. Þat hrœrðisk, ok var þat einfœtingr ok skauzk ofan á þann árbakkann, sem þeir lágu við. Þorvaldr Eiríksson rauða sat við stýri, ok skaut einfœtingr ǫr í smáþarma honum. Þorvaldr dró út ǫrina ok mælti: “Feitt er um ístruna. Gott land hǫfu vér fengit kostum, en þó megu vér varla njóta.” Þorvaldr dó af sári þessu litlu síðar. Þá hleypr einfœtingr á braut ok suðr aptr. Þeir Karlsefni fóru eptir honum ok sá hann stundum. Þat sá þeir síðast til hans, at hann hljóp á vág nǫkkurn. Þá hurfu þeir Karlsefni aptr. Þá kvað einn maðr kviðling þenna: ‘Eltu seggir, / allsatt vas þat, / einn einfœting / ofan til strandar, / en kynligr maðr / kostaði rásar / hart of stopir, / heyrðu, Karlsefni.’ Þeir fóru þá í brott ok norðr aptr ok þóttusk sjá Einfœtingaland. 1935 231-232 It happened one morning that Karlsefni and his men could see over a clearing a kind of speck which glittered back at them, and they shouted at it. It moved – it was a uniped – and hopped down to the river-bank off which they were lying. Thorvald Eirik the Red’s son was sitting by the rudder, and the uniped shot an arrow into his small guts. He drew out the arrow, exclaiming, ‘There is a good coat of fat round my paunch! We have won a fine country, though our time to enjoy it proves short.’ Thorvald died of this wound a little later. The uniped skipped off back north, and Karlsefni and his men gave chase, catching sight of him every now and again. The last glimpse they had of him, he was running towards some inlet or other. They turned back then, and one of them sang this ditty: ‘Men went chasing, / I tell you no lie, / a uniped racing / the seashore by. / But this man-wonder, / curst son of a trollop, / Karlsefni, pray ponder! / escape at a gallop.’ Then they set off back north, thinking they had sighted Einfætingaland. Jones 154-155

The most unequivocal indication in this passage that we are talking about a true uniped (i.e., not somebody who has simply lost a leg or foot) lies in the place-name and the stanza, both of which are officially outside the scope of the dictionary. The name Einfœtingaland makes clear that we are dealing with an entire race of unipeds, and Fritzner includes it, glossing as Land som beboes af einfœtingar. Complementarily, the verse refers to the einfœtingr as a kynligr maðr peculiar/strange man, appropriate only if we are talking about a uniped. Whether the verse was intended by its composer as a literal description of a uniped is a question taken up by Ian McDougall (1997) but which I shall bypass here.

This passage reminds us how awkward it can be to distinguish two corpora, one of prose and the other of poetry, when so many saga texts are couched in prosimetric form. The consequences for practical lexicography can be seen again in ch. 9 in , where the word drápa is attested in both the prose narrative and the verse quotations. Although the entry on this headword is very full and informative, citation of the Sneglu-Halli episode (with the verse included alongside the prose) would have been useful, since it carries key information about drápa form and function (Gade 1991).

The sheer volume of what can be called poetological prose in Old Icelandic poses further awkward issues. A considerable amount of purely poetic lexis has to be included in de facto, because it is cited within the prose of and other treatises. Some of these items, and especially the heiti, are essentially intractable, given the paucity of evidence (see Elena Gurevich 1992a and 1992b), and it is admittedly difficult for to deal with them within its mandate. A case in point is drǫfn, glossed in as follows: (proprial.) (heiti; for bølge i kenning for havet? for havet? // for a wave in a kenning for the sea? for the sea?). This gloss, enclosed in parentheses to reflect the status of the word as a proper name, seems unclear and indecisive, compared with Faulkes’s treatment in (Snorri Sturluson 1998 452). The sole citations supplied in are from , where heiti for sea and the names of the daughters of Ægir and Rán are listed. The presence of the same noun in a series of heiti for river is noted by Faulkes but not by . Despite the evident wavering of attention that we detect here, it is commendable that has at least recorded such lexical items, since, as Gurevich observes some of these heiti…might originally be euphemisms or isolated names of some other kind, once used beyond the bounds of skaldic tradition 1992b 37 .

A few points to conclude. The lay-out and typography of the volumes published thus far is both elegant and astoundingly accurate, despite the challenging complexity of the contents. I have noticed very few typographical errors (one such is syllabe for syllable in the entry for belgdraga).

In addition to the rich resource made available in the published volumes, a number of electronic resources have been put in place on the dictionary’s website (). By consulting the Word-List it is possible to gain information on all headwords, both those that have already been treated in print in and those that will be incorporated in forthcoming volumes. Thus a search on the word refill elicits that 44 citations are held. The website also gives access to an electronic version of parts of the and to selected texts from the beginning of C.R. Unger’s edition of I (, , , , and .

Grimstad (2003 220) rightly remarks on the importance of this magnificent dictionary to all who work in Old Icelandic and Norwegian studies and expresses the hope that further volumes will appear regularly. Sadly, however, this expression of hope has been overtaken by history. A memo from the Commission dated 29 December 2004 notified subscribers that the next eight to ten years will be devoted to the production of a digitalized version of the entire body of the dictionary material and that further volumes will not be appearing in the interim. It is naturally disappointing and frustrating that for the next several decades must suffice for general, daily use, just as we were becoming accustomed to expecting better things (Cathey 1998 289). But it is possible that the that eventually emerges will itself be the better for the longer preparation time, and especially so if it can benefit from some other major editing, text-encoding, and lexicographical projects currently underway.