This interface will soon cease to be publicly available. Use the new interface instead. Click here to switch over now.
Consideration of the effects of kennings and the definition of what a kenning is have been closely connected in the scholarly literature, and opinions have differed about both. Heusler (1922) opposed Meissner’s definition of the kenning (see §5.1.1.A) primarily for aesthetic reasons. He wanted to limit the scope of the term ‘kenning’ to metaphorical kennings, for which he coined the phrase Metapher mit Ablenkung ‘metaphor with diversion’, which he defined by the following expression: A = B : C, d. h. A (der auszudrückende Begriff) ist zwar kein B (Grundwort), aber im Hinblick auf C (die Bestimmung) hat es B-Eigenschaften, es vertritt bei C das B ‘A = B : C, that is, A (the notion to be expressed), though not a B (base-word), nonetheless has B-qualities with regard to C (the determinant); it represents the B with respect to C’. Referring to the kennings ‘Áss <god> of the mountains [giant]’, ‘crucible-snow [silver]’ and ‘snake of the bowstring [arrow]’, he explains (Heusler 1922, 130): Der Riese spielt auf dem Felsengebirge die Rolle des Asen; das Silber ist der Schnee, der mit dem Schmelztiegel zu schaffen hat, der Pfeil ist das Schlangenähnliche, an das man bei der Bogensehne denkt ‘The giant plays the role of the Áss in the mountains; the silver is snow that has to face the crucible, the arrow is the snake-like object brought to mind by the bowstring’. He refers to all non-metaphorical kennings as two-part heiti. Although Heusler’s narrow definition is motivated by a justifiable concern to capture something of the effect of the kenning, only a few scholars adopted this definition (Trost 1933; Gardner 1972; Brodeur 1952). Most took up the broader definition offered by Meissner (Wolff 1923; Mohr 1933; Jón Helgason 1953; Lie 1963b; E. A. Kock 1938; Einar Ól. Sveinsson 1947; Hallberg 1962a; Fidjestøl 1974; Marold 1983).
Even if the broader definition of the kenning as a paraphrasing tool following various patterns is now accepted, it nonetheless makes sense to recognise Heusler’s concern with capturing the aesthetic effect of the kenning, because, from an aesthetic perspective, Meissner’s very formal definition brings together stylistic devices that have very different effects. It is therefore desirable from the standpoint of poetic style to distinguish between the several types of kenning. Attempts have repeatedly been made to subdivide kenning patterns into groups. Various classification principles have been applied, sometimes based upon the categories expressed in the kenning (e.g. kinship or deeds), sometimes upon its psychological background, and sometimes upon its contextual role. The commonest division distinguishes between metaphorical and non-metaphorical kennings (Noreen 1921; Jón Helgason 1953; Meissner; Einar Ól. Sveinsson 1947; Hallberg 1962a). This classification can be supplemented with a further class based on the category of the paraphrased concept, that is, on whether it is an appellative (a common noun) or a personage who can be identified either as a representative of a certain type or as an individual. A preliminary division distinguishes appellative kennings from personal kennings. This distinction corresponds essentially to the division between metaphorical and non-metaphorical kennings, even though there are also base-words with metaphorical character among the personal kennings. These include numerous man-kennings with the base-word ‘tree’, such as ‘tree of gold’. A closer examination of their structure suggests that they should be assigned to the personal kennings category.
Appellative kennings as a group have predominantly metaphorical base-words, but here we also find base-words one would describe as metonymic or synecdochic rather than metaphorical. The so-called metaphorical kennings should thus be referred to more precisely as ‘tropic kennings’ (corresponding to the rhetorical notion of a trope: a figurative, non-literal expression). A further subdivision within this group could be made according to the various kinds of trope used: in metaphorical kennings, the base-word relies on a comparison, e.g. hjól meyjar Hǫgna ‘wheel of the maid of Hǫgni <legendary king> [= Hildr > shield]’ (Bragi Rdr 2/3, 4III). The wheel and the shield are comparable in that both are round. If the base-word and the referent fall under a common generic concept, this constitutes a synecdochic kenning. An example is dǫgg hræva ‘dew of corpses [blood]’ (Bragi Rdr 4/3III). Here, the base-word ‘dew’ and the referent ‘blood’ are of the same generic type: ‘liquid’. If the base-word expresses a function or a trait of the referent, this can be called a metonymical kenning. An example would be vegr mána ‘path of the moon [sky/heaven]’ (Þjóð Haustl 14/8III), where the base-word ‘path’ denotes a function of the referent ‘sky’, across which the moon wanders.
Personal kennings paraphrase personages either according to type (e.g. warrior, ruler, or giant) or as individuals, such as Haraldr harðráði ‘Hard-rule’ or Óðinn. This distinction results on the one hand in typifying kennings, i.e. those that typify groups, and, on the other, in so-called antonomastic kennings. Both groups can be further broken down according to the specific circumlocutory device employed. By means of such a classification one can distinguish four groups of personal kennings. Persons can be paraphrased according to: 1) their deeds and accomplishments; 2) their characteristic connections to various objects; 3) their relationships to various other people or mythical beings; or 4) their kinship with others. Various techniques are used with regard to the base-word of this type of kenning. It can express either a deed, a possession, or the relationship itself (often through nomina agentis); or it can use a personal name, frequently a god’s name, or tree-name as a kind of metaphorical device.
A circumlocution based on deeds can be formed in one of the following ways: 1) through the use of an agentive noun (nomen agentis), as in stillir lýða ‘controller of men [ruler]’ (Bragi Rdr 10/1III); 2) through forming the base-word from a god’s name, with a verbal element appearing in the form of a stem compounded with the name, as in felli-Njǫrðr flótta ‘the slaying-Njǫrðr <god> of the fleeing ones [warrior]’ (i.e. the Njǫrðr who slays those who flee) (Eskál Vell 29/1, 2I); 3), through use of a word for ‘tree’ as base-word, with the verbal element likewise appearing as a stem, here compounded with the tree-name, as in hleypimeiðr hlunnviggja ‘the impelling-tree of roller-steeds [ships > seafarer]’ (i.e. the tree that makes the steeds of the roller gallop) (Hfr Óldr 2/7, 8I). The technique is similar for characterisations based on objects, as in njótr hafra ‘user of goats [= Þórr]’ (ÚlfrU Húsdr 5/3III) with a nomen agentis, Viðrir Hveðru brynju ‘the Viðrir <= Óðinn> of the Hveðra <troll-woman> of the mail-shirt [axe > warrior]’ (Bragi Rdr 11/2III) or meiðr morðteins ‘tree of the murder-twig [sword > warrior]’ (KormǪ Sigdr 2/1, 2III).
Early in the skaldic tradition, the use of a god’s name as the base-word of a personal kenning probably served an honorific, perhaps even a religious purpose, although the putative religious significance cannot be established with certainty. The virtual abandonment of this type of kenning during the first few decades after Christianization can be seen as evidence that a god’s name evoked a connection with the religious sphere. Later, these kennings were conventionalised and used mostly without regard for their religious meaning.
The group of very common kennings with tree-names as base-words may have arisen by way of metaphors, even though the tree : man relationship also appears in myths (e.g. the creation of mankind from tree trunks in Vsp 17-18). Derivation from an ofljóst process (see below), as Snorri suggested in Skáldskaparmál (SnE 1998, I, 40), is unlikely: he argued that reynir ‘the tester’ is homonymous with reynir ‘rowan’, which he then claimed was varied with other tree-names. It seems that kennings with tree-names as base-words appear only in the group of typifying personal kennings.
The following table offers an overview of the typological diversity of the kenning:
Þat eru nýgǫrvingar at kalla sverðit orm ok kenna rétt, en slíðrirnar gǫtur hans, en fetlana ok umgjǫrð hams hans. Þat heldr til ormsins náttúru at hann skríðr ór hamsi svá at hann skríðr mjǫk til vatns. Hér er svá sett nýgjǫrving at hann ferr leita blóðs bekkjar at þar er hann skríðr hugar stígu, þat eru brjóst manna. Þá þykkja nýgjǫrvingar vel kveðnar ef þat mál er upp er tekit haldi of alla vísulengð. *En *ef sverð *er ormr kallaðr, <en síðan> fiskr eða vǫndr eða annan veg breytt, þat kalla menn nykrat, ok þykkir þat spilla.
These are nýgǫrvingar to call the sword a snake and use an appropriate determinant, and call the scabbard its path and the straps and fittings its slough. It is in accordance with a snake’s nature that it glides out of its slough and then often glides to water. Here the nýgjǫrving is so constructed that it [the snake] goes to find the stream of blood where it glides along the path of thought, i.e. men’s breasts. Nýgjǫrvingar are held to be well composed if the idea that is taken up is maintained throughout the stanza. But if a sword is called a snake, and then a fish or a wand or varied in some other way, this is called nykrat [a monstrosity], and it is considered a defect (translation modified from Faulkes 1987, 170).
The two phenomena mentioned here, nýgerving and nykrat, have played an important role in research on the stylistic effect of the kenning, but they have not been used by researchers in quite the same sense as that in which Snorri uses them. Einar Ól. Sveinsson (1947) interprets nýgerving very generally as the harmony of a metaphoric kenning with its environment, something he takes to be a further development of original skaldic style which had been characterised by the semantic contrast between a kenning and the sentence elements surrounding it. Analogously, Lie (1952) sees in nykrat an attempt of the early skaldic tradition at a style that emphasised unatur, literally ‘unnaturalness’ in the sense of a streben etter øket uttrykksintensitet ‘striving toward heightened intensity of expression’ and apotropaic powers. Snorri, on the other hand, does not perceive in nykrat any attempt at achieving an expressive quality, but regards it as a defect. Thus Snorri is concerned not with the relationship of a kenning to its syntactic-semantic environment, but with the way the kennings used in a stanza relate to one another. Nykrat would then denote the use of different metaphorical kennings in a single stanza, leading to constantly shifting imagery. A study by Marold (1993b) shows that it was precisely the early skalds who avoided this device. Nýgerving, by contrast, can be thought of as a creative expansion of the imagery of a kenning: the originally selected field of imagery is maintained throughout, and new kennings are created along these lines, as Snorri demonstrates with his example of the sword-kenning with the base-word ‘snake’, from which new kennings for ‘scabbard’, ‘strap’ and so on are formed by referring back to the snake’s habitat.
Given the complexity of this circumlocutory language and its obvious lack of congruence within its semantic and situational contexts, the question of the function of the kenning arises, assuming one does not wish to join Borges (1966, 17) in condemning it as eine der unverfrorensten Verirrungen der Literaturgeschichte ‘one of the most brazen aberrations in the history of literature’. Snorra Edda makes no statements about the function of the kenning; only on the basis of a passage in Skáldskaparmál, that constitutes a sort of dedication to readers, can a few conclusions be drawn (SnE 1998, I, 5): En þetta er nú at segja ungum skáldum þeim er girnask at nema mál skáldskapar ok heyja sér orðfjǫlða með fornum heitum eða girnask þeir at kunna skilja þat er hulit er kveðit: þá skili hann þessa bók til fróðleiks ok skemtunar ‘But these things have now to be told to young poets who desire to learn the language of poetry and to furnish themselves with a wide vocabulary using traditional terms; or else they desire to be able to understand what is expressed obscurely. Then let such a one take this book as scholarly inquiry and entertainment’ (Faulkes 1987, 64). Two functions can be discerned from Snorri’s statement: the kenning serves to expand the body of synonyms at the poet’s disposal, which consists of simple appellatives, poetic terms and kennings; and it offers the poet a way of speaking obscurely.
The poet’s choice of a kenning may be guided by various factors. It certainly helps him meet the demands of metre and rhyme, primarily by virtue of its variability, but there are undoubtedly other functions. The external form promotes what is called copiosum in rhetoric, whereby the verbal richness of a complex kenning gives dominance to a particular notion in a half-stanza, in that the individual parts of the kenning are distributed across all its lines. For example, skalds have a predilection for placing extensive kennings for poetry in their poems’ introductory stanzas and this emphasises both their own cleverness as poets and their place in an intellectual tradition that goes back to the myth of the god Óðinn’s theft of the mead of poetry from giants and dwarfs.
The kenning generally serves its referential purpose in isolation, that is, independently of its context within the sentence. This aspect reveals the characteristic difference between modern poetic imagery and metaphor on the one hand and the kenning on the other. Research on metaphor has shown that it is closely connected to its context within the sentence, so closely, in fact, that it cannot exist as such outside it, because the meaning – insofar as it is not a conventionalised metaphor – is clearly established only by the context of the sentence itself. Weinrich (in Harweg et al. 1968, p. 100) takes account of this in his definition of the metaphor, describing it as a Wort in einem konterdeterminierenden Kontext ‘word in a counter-determining context’. This means that a word used metaphorically contradicts the expectation created by the semantic context, and it does this in such a way that there is no indication that a word is metaphorical at all until the context has been contradicted (cf. also Petöfi 1975, 290; Suerbaum (Harweg et al. 1968, p. 101); Hörmann 1971, 320-3; Mancaş 1969; Löwenberg 1975 and others). The comprehension of a metaphor comes about only in a process of understanding initiated by the contradictory context.
The kenning, even the metaphorical kenning, works differently. It exists prior to and independently of the poem – it is not the context of the sentence, but instead the ‘internal context’, the determinant, that identifies the metaphor contained in the kenning as a metaphor in the first place. Here is an example. The metaphoric use of ‑garðr ‘fence’ in the kenning ljósgarðr barða (Egill Skjalddr 1/2V) ‘the bright fence of the prow [shield]’ is not established or determined as such by the context mál es lofs at lýsa […] es þák ‘the matter at hand is to illuminate through praise [the shield] I received’, but instead by the determinant of the kenning, barða ‘of the prow’. The kenning thus maintains its independence from the sentence, since even outside of this context ljósgarðr barða would be a recognisable periphrasis for ‘shield’. This independence distinguishes kennings from metaphors, and it is for this reason that the kenning should be seen as a periphrasis and not as a metaphor (cf. Fidjestøl 1974 and 1979b).
Nevertheless the kenning stands in contrast to both its immediate verbal environment and its situational context. This applies particularly to tropic kennings, where the metaphorical or metonymic base is at odds with the semantics of the sentence. But even personal kennings, whether they be typifying or antonomastic, can contradict the situational context. It can even be said that this contextual incongruity is more or less typical of the functionality of the kenning. Yet there are also kennings that do match the context of the sentence, the so-called treffende ‘fitting’ kennings, and there are sometimes kennings that also become the basis for a metaphorical sentence arrangement.
Depending on the nature of its relationship to the contextual setting, a kenning can serve various functions. The context-independent effect of a kenning resembles that of the epitheta ornantia ‘ornamental epithets’ in epic poetry. Both are context-independent stereotypes that impart an appropriate decorativeness. The skalds regarded their compositions as gifts, and in fact they actually emphasized the craftsmanship of their poetry by means of comparisons with metal- and wood-working. Formal aspects of the poetry such as metre and artfulness of language, which arise to a large extent through the use of kennings, were held in high regard. Kennings, therefore, participate in and contribute to the function of praise.
A further function of the context-independent kenning is that it poses riddles that yield pleasure and rational satisfaction when solved. The kenning is a stylistic figure that speaks to knowledge and reason. In order to appreciate these figures fully, one must delight in word-play and logical relationships. Poetry and intellectuality are not antagonistically disposed in this kind of poetry. Snorra Edda confirms this view of matters when it recommends itself to the reader til fróðleiks ok skemtunar ‘as scholarly inquiry and entertainment’ (Faulkes 1987, 64), or when it claims that everyone who partakes of Óðinn’s mead of poetry will become skáld eða frœðimaðr ‘a poet or a wise man’ (ibid.). In her interpretation of Snorra Edda, Clunies Ross (1987) goes a step further. She endeavours to comprehend Snorri’s presentation of kennings and heiti in Skáldskaparmál as a systematic illustration of poetic language based on a Platonic world-view, and thereby as a testimony to the pre-Christian religion.
Delight in riddles is clearly demonstrated by a phenomenon like ofljóst ‘excessively clear’, which can best be explained as an ironic term. In this case it is not the word itself that is replaced by a circumlocution, but its homonym. Snorri refers to this device in Skáldskaparmál (SnE 1998, I, 109). He gives a few pairs of terms such as reiði ‘wrath’ or ‘tackle’ or far ‘ship’ and fár ‘hostility’ and observes: Þvílík orðtǫk hafa menn mjǫk til þess at yrkja fólgit ok er þat kallat mjǫk ofljóst ‘People frequently use such vocabulary so as to compose with concealed meaning, and this is usually called word-play (ofliost [obvious])’ (Faulkes 1987, 155). Snorri’s example lið, líð, hlið, hlíð ‘limb, ale, gate, slope’ reveals that he does not limit the figure to strict homonyms. An example of such an ofljóst is found in a poem by Hallar-Steinn about Skáld-Helgi Þórðarson (HSt Frag 6/1, 2, 3III): sól bókar fló hvast of snertu ‘the sun of the book [colour (steinn ‘stone’)] flew sharply in the fight’. It derives from the homonym steinn, which means both ‘stain, colour’ and ‘stone’. In the present edition a kenning that depends on an ofljóst homonym for its understanding is formatted to show the Old Norse word that carries the pun, together with the alternative translation, in round brackets after the main referent, as in the example above. Delight in riddles is likewise evident in the so-called name-kennings: these are based upon the constituent and semantically transparent parts of personal names. Egill Skallagrímsson uses a circumlocution for his friend Arinbjǫrn’s name with bjóða-bjǫrn ótta birkis ‘bear of the table of the terror of the birch tree [(lit. ‘table-bear of the terror of the birch-tree’) FIRE > HEARTH (arin ‘Arin-’)]-bjǫrn’ (Egill Arkv 16/7-8V) (Eg 112). He replaces the first part of the two-part name (Arin-) by a kenning for the noun arinn ‘hearth’. Similarly Eyvindr skáldaspillir (Eyv Lv 14/3, 4I) calls the Icelanders álhimins lendingar ‘eel-heaven’s [ice]-landers’.
Each of the functions of a kenning mentioned so far also applies independently of the specific context in which it appears. However, kennings can nonetheless convey content by highlighting a dominant idea. They can deliver an ironic commentary or frame a parody. They constitute a sort of second level beyond the simple content domain of a stanza, which gives the poet the freedom to accentuate the plot, to give commentary, to look backward or forward in time and to intimate certain associations.
The poet has access to a wide array of possibilities in emphasising specific aspects of an event. The tropic kennings (metaphor, synecdoche and metonymy) work by evoking associations with various domains. For example, when eyes are called ennitungl ‘forehead moons’, the figure is an expression of the referent’s unusual supernatural powers. Or, if the kenning pattern ‘liquid of the drinking horn [beer]’ is used, and beer is paraphrased by ‘wave of the bull’s spear [horn > beer]’ in a stanza about the drowning of the Yngling ruler Fjǫlnir in a beer keg (Þjóð Yt 1I), then the base-word ‘wave’ helps to highlight the aspect of drowning by evoking the image of a troubled sea. The highlighting of specific aspects can also be used to comment on a situation. When the woman-kenning ‘goddess of wealth’ is realised as Rindr mundar ‘Rindr <giantess> of the bride price’ (Bjhít Lv 10/3V), it is a purposeful variation on the pattern: the naming of Rindr draws on a myth in which Óðinn used magic to make this reluctant giantess amenable to his sexual advances, and the variation of ‘wealth’ with ‘bride price’ is directed at the situational context. By using this kenning, the poet, Bjǫrn Hítdœlakappi, levels an accusation at his rival: you purchased the reluctant woman from her father with your wealth!
The association of two domains through the metaphors within kennings also provides an often-used technique for intermingling the earthly and the mythical or heroic domains. This aspect has been described as an attempt to poeticise banal subjects, to lift an intrinsically banal, everyday event into the heroic and mythical sphere. This accusation was directed mostly at the use of gods’ names as base-words (Meyer 1889, 156-88; Mittner 1955, 13; von See 1964b, 9). However, earlier scholars had already (and rightly) pointed out that such kennings would probably have had a parodic effect if applied to inappropriate subjects, as is clear in two stanzas by Þjóðólfr Arnórsson (ÞjóðA Lv 5 and 6II) about an argument between a blacksmith and a tanner, whom he depicts once as Þórr and the Miðgarðsormr ‘World Serpent’ and another time as Sigurðr and the dragon Fáfnir. Much more frequently, however, the connection between the mythic or heroic and the earthly spheres is evoked in a spirit of eulogy rather than parody, with the implication that a ruler or his deeds re-enact or match mythic or heroic prehistory.
It is hardly possible to make definite statements about the origins of the kenning. It is attested in skaldic poetry from the earliest preserved examples onward as a fully-formed device, and must have originated significantly earlier. To an extent, this figure is rooted in a simple circumlocutory practice that was probably used everywhere, and even the metaphorical kenning can be paralleled in other literatures. Comparable devices are found not only in Old Irish, Sanskrit and Greek poetry (Krause 1925b), but also in poetry of the German Baroque period and occasionally even in Modern German expressions like Rebenblut ‘vine-blood’ for ‘wine’ or Wüstenschiff ‘desert-ship’ for ‘camel’. In Modern English ‘ship of the desert’ is also a camel, while a ‘clothes-horse’ is a rack for drying clothes. Beyond this, however, the question of where and when the system of circumlocutory types and their variation arose is one that cannot be answered.
Numerous scholars have attempted to derive the kenning from the magical-religious domain as a cryptic tabu reference (Olrik 1897; Portengen 1915; Moberg 1942; Ohlmarks 1943; Mittner 1951; Mittner 1955). While it is quite conceivable that noa-expressions (expressions used to avoid a tabu word) make use of periphrastic devices of great variety, and that they share this trait with the kenning, this theory fails to explain how this method of circumscription found its way into poetry, unless we assume, following Moberg, that the origins of skaldic poetry lay in funeral poems that systematically paraphrased tabu terms. However, this assumption cannot be proven.
Lie (1952; 1963b) attempted to explain the origins of the kenning style with a theory he proposed in the context of Bragi Boddason’s Ragnarsdrápa, according to which kenning style and skaldic poetry arose as a complement to and in imitation of shield painting. According to this theory of the kenning, its contrastive effect combining the most heterogeneous of domains is a counterpart to, and a linguistic recreation of, the supposed apotropaic-numinous effect of shield painting. The unnaturalness such a combination of heterogeneous objects would have invoked constituted a maktpotensering ‘increase in power’ in an apotropaic-numinous sense.