How an Adder Became an Arrow
Battle Kennings in Old English Poetry
Mariëlle Bosch, S1495852 Supervisor: Dr. K. Olsen
30 June 2011 Word count: 13.501
1 Table of Contents
Introduction ………..……… p. 2 Chapter 1: The Kenning ……….………. p. 3 Kenning Formation ……….. p. 5 Kenning‘s Meaning ………. p. 7 Chapter 2: Theories on Kennings ………... p. 10
2 Introduction
For poets it was, and still is, important to impress their audience. In order to
accomplish this they needed to show off their skills. The Scandinavian skalds1 as well as their Anglo-Saxon counterparts used a poetic device that was known as the kenning. They started to paint pictures in words, often enhancing the words by depicting scenes more tangible for their Medieval audience. The kenning still intrigues today‘s audience. Perhaps not all of the kennings are as easy to understand now as they were then. However, the image-enhancing quality of this poetic device withstands time.
Battles were a regular subject for the Anglo-Saxon poets. In order to show their audience the great battles and the brave warriors, Anglo-Saxon poets created kennings to denote the many concepts relating to battle like fighting, weapons, blood, etc.
In this dissertation, I will investigate the ‗what, how and why‘ about kennings. I will first examine already existing theories and then provide my own analysis of battle-related kennings. The main question I hope to answer is: Is it possible to find a formula in regards to the use of kennings? Did poets use a particular set of rules which told them when and where in a text it was ‗allowed‘ to use a kenning, or did they have a creative liberty — a poetic license, and thus, allowing the use of kennings wherever and whenever they wanted?
1 ―Scandinavian poet and singer of medieval times,‖ from O.N. skald ―skald, poet‖ (9c.), of unknown origin,
3 Chapter 1: The Kenning
Firstly, I will look at the word ‗kenning‘ itself before discussing its function. The etymology of the word might give some more information on what is meant by it. The English word ‗kenning‘ stems from the Old Norse verb kenna. This verb can mean ‗know, make known,2
teach, preach, perceive or recognize.‘3 The basis of this Old Norse verb can be traced back to the Proto-Germanic verb *kunnan, meaning ‗to be mentally able, to have learned‘.4
Known from Old Norse are the expressions kenna við and kenna til. The meaning of the word ‗kenning‘ probably derives from these expressions. Kenna við means ‗to name after‘5
or, as Roberta Frank describes it, ―to express … one thing by means of another.‖6
Kenna til means ‗to qualify by, making into a kenning by adding.‘7 Especially the latter expression shows that the kenning elements replace the name of a concept or thing; their literal meaning yields to the meaning of the concept the kenning is to represent.
The definition of a kenning is not easily found. Nevertheless, in order to give a general idea of what a kenning is, I looked at the definition given by the Oxford English Dictionary. According to the Oxford English Dictionary, a kenning is ―[o]ne of the periphrastic
expressions used instead of the simple name of a thing, characteristic of Old Teutonic, and esp. Old Norse, poetry. … The term is adopted from the Medieval Icelandic treatises on poetics, and is derived from the idiomatic use of kenna við or til, ‗to name after‘.‖8 Thus, a kenning is a kind of wordy replacement for the name of a concept or thing.
2 Online Etymology Dictionary, s.v. ―can,‖ accessed June 2011,
<http://www.etymonline.com/index.php?search=can&searchmode=none>.
3 Anthony Faulkes, Skáldskaparmál: 2 Glossary and Index of Names (London: Viking Society for Northern
Research, 1998), 334.
4
Online Etymology Dictionary, s.v. ―can.‖
5 Faulkes, Skáldskaparmál 2, 334.
6 Roberta Frank, Old Norse Court Poetry: The Dróttkvætt Stanza (London: Cornell UP, 1978), 42. 7
Faulkes, Skáldskaparmál 2, 334.
4 These explanations of the origin of the word ‗kenning‘ and the meaning of the word ‗kenning‘ itself already point to its form and function; a kenning is not always literally whatever is written down. The kenning gūðwine9 literally means ‗war-friend‘. However, this kenning does not represent a person who is someone‘s friend in war, but represents the concept ‗sword‘. It almost seems as if the poet first wants the reader to solve some kind of puzzle before the meaning of the kenning is known.10 The kenning has an almost riddle-like characteristic — a cryptic description of a concept. Wilhelm Bode describes this characteristic in Die Kenningar in der angelsächsischen Dichtung as follows:
Eine Kenning ist demnach eine etwas entlegene, seltene Bezeichnung für einen Begriff, für den ein allgemein gebrauchter Terminus vorhanden ist, ein Terminus, der so nahe liegt, dass man ihn erst umgehen muss, um zu der Kenning zu gelangen.11
The kenning is ―usually regarded as the prime feature of skaldic art[.]‖12
It is a poetic device that is found in many, if not all, Old Norse and Old English texts, such as the Old Norse treatise on poetics Skáldskaparmál and poems like Glymdrápa and Þórsdrápa and the Old English poems Beowulf and Judith. Skalds were very fond of using a lot of kennings in their poetic works. Just as their skaldic counterparts, Anglo-Saxon poets also wrote poetry using them. The kenning can share characteristics with metaphor but is not identical with the latter. A kenning is a circumlocution: concepts or persons are described by means of several words instead of just one. It is a way in which a poet describes a concept more extensively than is really necessary. As Roberta Frank notes: ―The Old Norse kenning is a periphrasis,
9
Beowulf, ed. Michael Swanton, line 1810.
10 For more on kennings versus riddles, see: Ann Harleman Steward, ―Kenning and Riddle in Old English,‖
Papers in Language and Literature 15 (1979): 115-36.
11
Wilhelm Bode, Die Kenningar in der angelsächsischen Dichtung (Darmstadt: Zernin, 1886), 7.
5 consisting of two or more substantive members, which takes the place of a noun.‖13
Where the Old Norse kenning can exist of two or more elements, the Old English kenning exists of only two. It is a more simplified version of the sometimes rather intricate kenning formed by the skalds.14
A further, in-depth discussion of the different theories on the nature of the kenning will follow in the next chapter. However, to get a general idea of what kennings are, the structure and meaning of a kenning need to be examined first.
Kenning Formation
For the formation of a kenning one concept is selected and a circumlocution is formed to denote this concept. Thus, two or more words make up a kenning for a single concept. The individual words do not denote concepts similar to the one that is described by the
circumlocution, though there has to be some kind of relationship between the original concept the kenning represents and the kenning elements. This relationship between the kenning and the concept or person denoted can be formed through using a ―statement of family
relationship.‖15
A kenning can refer to what a person has done; it identifies deeds to describe the person who performed them. These are nomen agentis kennings. ―In these … types, the noun in the nominative case literally designates the person or concept signified by the whole.‖16 An example of a nomen agentis kenning is bēag-gyfa, ‗ring-giver‘, which means ‗king‘. This kenning refers to a king‘s proper behavior towards his warriors after victory in battle; he
13 Frank, Old Norse Court Poetry, 42. 14 See Chapter 2, p. 10.
15
Frank, Old Norse Court Poetry, 42.
6 repays them by giving rings. This example shows the metonymical quality that a kenning can have; the concept is substituted by something that is closely related to it. This can be an action, an object, an institution, etc.17 In this case the king is represented by one of the actions he does.
Based on the structure of the kennings, a distinction between several types of kennings is found in Old Norse and Old English texts. In order to explain these types I will here refer to E.O.G. Turville-Petre‘s descriptions in his book Scaldic Poetry. As mentioned before,
detailed descriptions and discussions of theories on kennings will follow in the next chapter. Turville-Petre mentions two structural types of kennings: The simple kenning and the extended kenning. Simple kennings are circumlocutions consisting of ―two nouns, or of a name combined with a noun in the genitive case or forming a compound with it.‖18
Thus, these kennings consist of two parts, the ‗base word‘ and the ‗determinant‘. The base word and the determinant separately do not have the same meaning as when they are used together. In this case, the ―meaning is never that of the base word.‖19
An example of a simple kenning is
sig-Tý, ‗victory Týr‘, which is a kenning for Odin.
Extended kennings are kennings that consist of more than two parts, often having a part in it which itself is a kenning as well. An example is hyrjar hrannbrjótr ‗breaker of the fire of the wave‘. ‗Fire of the wave‘ is a kenning for gold, and thus a ‗breaker of gold‘ is a kenning for a generous man.20
Turville-Petre refers to Snorri‘s Skáldskaparmál to explain the different groups: There are kennings, tvíkennt ‗double‘ and rekit ‗chased‘ types. ―It is a kenning to call battle fleinbrak
17 Oxford English Dictionary, s.v. ―metonymy,‖ accessed June 2011, <http://www.oed.com/view/Entry/117628>. 18 E.O.G. Turville-Petre, Scaldic Poetry (Oxford: Clarendon Press, 1976), xlv.
19
Frank, Old Norse Court Poetry, 42.
7 (clash of spears), but it is tvíkennt to call the sword fleinbraks fúrr (fire of the clash of spears), and if the expression is longer it is rekit.‖21
Kenning’s Meaning
My short analyses have illustrated that, based on their structure, there are several types of kennings. This variety of kenning types also causes a kenning not always to be recognized for what it is, and it is also a cause for the different theories on kennings that scholars have proposed. As mentioned before, the kenning — although it is a literal expression — cannot be taken quite literally: Before the reader understands the meaning of a kenning, he has to solve a kind of puzzle first. The words that are used to form kennings have also been briefly
mentioned. Combinations of nouns — an important group being names and, in cases of skaldic poems, references to myths22 — form the circumlocutions. In addition to the structure of kennings, the semantic level of the kennings requires more attention.
There are several semantic groups to which Old English and Old Norse kennings can be assigned, ranging from kennings for the elements (e.g. fire, water, etc.) to kennings for warriors and battle. The list of kennings in Rudolf Meissner‘s Die Kenningar der Skalden:
Ein Beitrag zur skaldischen Poetik gives us an overview of the circumlocutions found in Old
Norse skaldic poetry.23 The list contains 106 headwords for which kennings were formed. A lot of these kennings, or similar ones, are also found in Old English. I have placed the
headwords into groups that share semantic characteristics. However, I have also searched for further shared characteristics between these 106 words and found that there are several groups that belong to the same general semantic group. There are kennings for:
21 Turville-Petre, Scaldic Poetry, xlviii. 22
See Chapter 2, p. 10.
8 - Persons in general (lēod-gebyrgea, ‗protector of a people‘ for a king)
- One specific person (Hangatýr, ‗Lord of the hanged‘ for Odin) - Parts of the human body (ennitungl, ‗brow-stars‘ for eyes) - Objects related to battle (beado-lēoma, ‗battle light‘ for a sword) - Animals (hæð-stapa, ‗heath-stalker‘ for a stag)
- Natural/meteorological elements (rodores candel, ‗heaven‘s candle‘ for the sun) - Concepts related to religion (ǣrendgāst, ‗message spirit‘ for angel)
Each kenning has its own meaning, though within these semantic groups there is often more than one kenning to denote a single concept or person. A good example is the concept ‗ship‘. Meissner fills about fifteen pages with kennings for this concept, ranging from kennings with the word horse/steed (e.g. gjalfrmarr ‗sea steed‘)24 to those with the word bear (e.g. flóðs
bjǫrn25
‗flood‘s bear‘).
Turville-Petre mentions another semantic group of kennings, namely the mythological and legendary kennings. In mythological and legendary kennings, a part of the kenning refers to a myth, which must be known in order to understand the true meaning of the kenning. For some, only the very basic knowledge is necessary to understand the kenning. Other kennings need a ―deeper knowledge of religious belief or legend.‖26
Accordingly, it can happen that the meaning of this type of kenning is lost simply because the myth that the kenning refers to is not known anymore. An example of a kenning based on a myth is Freyju tár, ‗Freyja‘s tears‘. This is a kenning for gold. The meaning of this kenning can only be derived because the myth about Freyja‘s tears is known: Freyja cried tears of gold when Óðr, her husband, left her.27
24 Meissner, Die Kenningar der Skalden: Ein Beitrag zur skaldischen Poetik (Bonn: Schroeder, 1921), 211. 25 Meissner, Kenningar der Skalden, 218.
26
Turville-Petre, Scaldic Poetry, xlix.
9 Deciphering the meaning of kennings is very important for the understanding of the poetry wherein they are used. The poets used their poetical freedom in those instances and expected their audience to get the meaning of the circumlocutions. Unlike the poets‘ contemporaries, readers nowadays, and especially those of the Old Norse poetry, can
experience difficulties in finding the concept that a kenning denotes. Old English kennings are more easily understandable because of their structure. Most are made up out of two elements28 and this in combination with the position of the kenning in the poetic line helps to find the intended concept. Also, there are no mythological kennings in Anglo-Saxon works, and thus, there is no danger of not being able to understand a kenning because the original myth it derived from is lost.29
28 Turville-Petre, Scaldic Poetry, xlviii. 29
10 Chapter 2: Theories on Kennings
In this chapter I will discuss seven theories on kennings. I will do this chronologically, starting with Snorri Sturluson‘s Skáldskaparmál, a 13th
-century work on poetic language and one of the oldest works discussing the kenning. This will be followed by Wilhelm Bode‘s Die
Kenningar in der angelsächsischen Dichtung, a work dating back to 1886 which specifically
focuses on the Anglo-Saxon kenning. After this Rudolf Meissner‘s Die Kenningar der
Skalden, a work dealing with the Old Norse kenning on which many later scholars based their
own theories. The next work is Scaldic Poetry, a theory on kennings by the Icelandic scholar Gabriel (E.O.G.) Turville-Petre ―born of a lifetime‘s study of and delight in this literature.‖30 Roberta Frank‘s Old Norse Court Poetry and Edith Marold‘s Kenningkunst — both focusing more on Old Norse kennings — are also discussed, and finally Gary Holland‘s Kennings,
Metaphors, and Semantic Formulae in Norse Dróttkvætt. After these seven theories I will
shortly discuss my view on them.
Snorri Sturluson, Skáldskaparmál (13th century)
In Skáldskaparmál ‗language of poetry‘, Snorri Sturluson gives insight into the diction of skaldic poetry through a dialog between the sea-giant Ӕgir and Bragi, the god of poetry. Ӕgir visits Asgard where he talks with the áss Bragi about the Ӕsir. Bragi is the one who starts to explain the origin of poetry: ―Then spoke Ӕgir: ‗This seems to me a very good way to
30
11 conceal it [gold] in secret language.‘ And Ӕgir went on: ‗How did this craft that you call poetry originate?‘‖31
Bragi explains that a man, named Kvasir, was made out of a vat of spit. He knew the answer to every question. The dwarfs Fialar and Galar killed him and mixed his blood with honey and made mead of it. Anyone who would drink of this becomes a poet or scholar. This is the reason why poetry is called ‗Kvasir‘s blood‘ or ‗dwarf‘s drink‘.
According to Snorri, there are two categories into which poetry can be divided, language and verse-forms. The category ‗language‘ can further be divided into several subcategories: ―To call everything by its name; the second category is the one called substitution; and the third category of language is what is called kenning [description][:]‖32
[A]nd this category is constructed in this way, that we speak of Odin or Thor or Tyr or one of the Ӕsir or elves, in such a way that with each of those that I mention, I add a term for the attribute of another As or make mention of one or other of his deeds. Then the latter becomes the one referred to, and not the one that was named; for instance when we speak of Victor-Tyr or Hanged-Tyr or Cargo-Tyr, these are expressions for Odin, and these we call periphrastic terms; similarly if one speaks of Chariot-Tyr [i.e. Thor].33
This last category is what Snorri calls kennd heiti, the kenning.34
Other groups that Snorri refers to when discussing kenning types are called fornofn (substitutions), viðkenningar (circumlocutions) and sannkenningar (true descriptions). Snorri seems to prefer to give an explanation for the kennings from an etymological point of view,
31
Snorri Sturluson, Edda, trans. Anthony Faulkes (London: Dent, 1987), 61.
32 Snorri, Edda, 64. 33 Snorri, Edda, 64. 34
Gary Holland, ―Kennings, Metaphors, and Semantic Formulae in Norse Dróttkvætt,‖ Arkiv för Nordisk
12 rather than looking at its metaphorical characteristics. When discussing the kenning and its types, Snorri focuses more on where they come from — their origin, and not so much on what they do. An example of this can be found in Snorri‘s discussion of ‗poetry‘:
Poetry is called sea or liquid of the dwarfs, because the liquid in Odrerir was Kvasir‘s blood before the mead was made, and it was made in that cauldron, and hence it is called Odin‘s pot-liquid, as in the poem of Eyvind quoted above:
…while his descent in pot-liquid of gallows-cargo we trace to gods.
Poetry is also called the dwarf‘s vessel or lid. Lid is a word for ale and lid is a word for ships. This is the origin of the expression whereby poetry is now as a result called dwarf‘s ship, as it says here:
I have ready both swollen wind of rock-earl‘s [giant‘s] bride [troll-wife‘s wind is ‗thought‘] and unforgettable dwarfs‘ ship to send the same way.35
Here, Snorri gives an explanation for the creation of kennings for the concept ‗poetry‘ and not so much the use and function of these kennings. As Clunies Ross argues: ―Snorri resists metaphorical explanations of kenning types, even when it seems obvious to a modern analysis that they are based on metaphorical equivalences.‖36
Structural types of the kenning that Snorri also mentions are that of the kenning, the
tvíkennt and the rekit. The tvíkennt is a kenning that consists of a double kenning; a kenning
within the kenning. The example Snorri gives is fleinbrak ‗clash of spears‘, this is a kenning for the concept ‗battle‘. If another part — another kenning — is added the kenning turns into a tvíkennt: fleinbraks fúrr ‗fire of the clash of spears‘, meaning ‗sword‘. If another kenning is added to the tvíkennt, it is called a rekit, a kenning consisting of multiple elements. A rekit for
35
Snorri, Edda, 72.
13 the concept ‗blood‘ could, for instance, become a kenning along the lines of ‗sweat of the fire of the clash of spears‘.
Most of the text of Skáldskaparmál contains lists of kennings and an explanation of the ways that they were formed. One of the ways to denote ‗man‘ is by referring to a man‘s deeds:
He shall be referred to by his actions, what he gives or receives or does. He can also be referred to by his property, what he owns and also if he gives it away; also by the family lines he is descended from, also those that have descended from him. … [Call] him achiever or performer of his expeditions or activities, of killings or voyages or hunting, or with weapons or ships.37
An example of this is calling a king a ‗ring-giver‘. In this case ‗giver‘ is the base word and ‗ring‘ is the determinant. Thus, the determinant is the most important of the two because it assigns the meaning to the kenning. According to Snorri, the king‘s deed of giving his retainers rings is what forms the kenning. For a man, a name of the Ӕsir may also be used as base word, as well as names of giants, though these are mostly used satirically or critically. The use of elves‘ names in a kenning for ‗man‘ is seen as a compliment. In the determinant of a kenning a woman is referred to by the drinks she gives, or the jewels she wears. For the base word trees can be used: ―It is proper to refer to her by calling her dealer (selia) or consumer (lóg) of what she hands out, but selia [willow] and log are trees. Hence woman is called in kennings by all feminine tree-names.‖38 Names of Asyniur, valkyries, norns or disir and by referring to the activities or possessions or descent of the woman are also possible base words
37
Snorri, Edda, 94.
14 for the kenning ‗woman‘. Snorri does not discuss whether any of these base words for a ‗woman‘ kenning are seen as negative or positive.
Without periphrasis, a poet can only call ―everything by its own name‖ according to Snorri.39 Calling ―everything by its own name‖ is what Snorri refers to as the heiti. This is one of the three categories that he mentions. The group heiti is for Snorri separated from the ‗real‘ kenning in that it is not a circumlocution and therefore cannot be seen as a kenning but as a synonym. However, he notes that there are similarities: One word is substituted by another, more poetic, variety, as for example, the word ‗horse‘ by ‗steed‘.40 This similarity is reflected in the names Snorri gives to the categories; kennd heiti and (ókend) heiti.41
Snorri also refers to the poets‘ playing with the double meaning of certain words, making it harder to understand the meaning of the kenning ―if it is a different distinction of meaning that has to be taken than the previous line seemed before to indicate[:]‖42
Læti means two things. Noise is called læti, disposition (ædi) is called læti, and ædi
also means fury. Reidi also has two meanings. It is called reidi [wrath] when a man is in a bad temper, the gear (fargervi) of a ship or horse is called reidi. Far also has two meanings. Far is anger, far is a ship.43
Snorri calls this word-play ofliost and states that this was used to ―compose with concealed meaning[.]‖44
39 Snorri, Edda, 132.
40 That this works similarly in other languages can be derived from the Dutch example ‗ros‘, another, more
poetic version of the word ‗horse‘. However, what must be noted is that the Dutch language shares the same Indo-European language origin as Old English and Old Norse.
41
―Ókend heiti: names, appellations, designations, terms without periphrasis, without qualifiers or attributives (determinants).‖ Faulkes, Skáldskaparmál 1, lxxii.
42 Snorri, Edda, 155. 43
Snorri, Edda, 155.
15 Nowadays, all the work that scholars do in the field of kennings finds its base in
Snorri‘s Skáldskaparmál, ―in part because of Snorri‘s insight and first-hand knowledge of the tradition, but in part too because of the insistence on the importance of taxonomy and the difficulties of interpretation by modern authorities.‖45
Wilhelm Bode, Die Kenningar in der angelsächsischen Dichtung (1886)
Wilhelm Bode focuses on the kenning in Anglo-Saxon works, which has a less complex structure compared to the Old Norse kenning. In his work Die Kenningar in der
angelsächsischen Dichtung, Bode starts off by giving his definition of a kenning: ―Eine
Kenning ist demnach eine etwas entlegene, settene Bezeichnung für einen Begriff, für den ein allgemein gebrauchter Terminus vorhanden ist, ein Terminus, der so nahe liegt, dass man ihn erst umgehen muss, um zu der Kenning zu gelangen.‖46
For Bode, a kenning often has the same syntactic role as a pronoun, and makes repetition of the same words unnecessary. Kennings often occur instead of a person‘s name, or as apposition, for example: Bēowulf
maϸelode, bearn Ecgϸēowes.47
About the structure of the Anglo-Saxon kenning Bode states: ―Die Kenningar sind zumeist Verbindungen zweier Nomina, sei es, dass sich dieselben zu einem
zusammengesetzten Worte vereinigen, sei es, dass sie getrennt nebeneinanderstehen, das eine vom andern abhängig: lîffruma, lîfes fruma.‖48 A few of the examples he gives are:
ϸeodcyninӡ ‗king of a people‘, ϸeodfruma ‗people‘s chief‘ and leodsceaða ‗people‘s enemy‘.
45 Holland, Kennings, 131. 46 Bode, Die Kenningar, 7. 47
Beowulf, ed. Michael Swanton, line 529.
16 The poet plays an important role in kenning formation. Bodes notes that ―eine kräftige Phantasie und ein leicht und tief erregbares Gemütsleben die Kenningar gedeihen lassen, dass dagegen eine strenge Schulung im logischen und sachangemässen Denken gegen die
Entwikklung und Blüte eines reichen Stiles wirken.‖49
This is a logic conclusion. After all, without any fantasy in the poet‘s mind, how could he be able to write poetry? ―Zu diesen Vermögen gehört besonders die Phantasie zu den Mitteln, die Phantasie zu lebendiger Thätigkeit zu veranlassen, die Benennung der Begriffe durch Kenningar.‖50
Bode gives the term malende kennings.51 A kenning invokes an image in the audience‘s mind. He proceeds by giving five different groups into which kennings can be divided:
1. ―geradeswegs malende angelsächsischer Kunstnamen‖ (eoforcumbol; ‗boar on sign, helmet‘)
2. ―festhaltende Kenningar‖ (sweordpleӡa; ‗sword-play, battle‘) 3. ―durch Vergleichung malende Kenningar‖
o Belebte Wesen werden untereinander verglichen (hildewulf ‗war wolf‘ for warrior)
o Belebte Wesen werden mit leblosen Dingen verglichen (leohtes leoht ‗light‘s light‘ for God)
o Leblose Dinge werden mit lebendigen Wesen verglichen (personification;
brimhengest ‗sea-horse‘ for ship)
o Leblose Dinge unter einander verglichen (merehûs ‗sea-house‘ for ship) o Abstracta werden mit konkreten verglichen (wyrmsele ‗serpent‘s hall‘ for
hell)
4. ―definierende Kenningar‖ (enӡla eard ‗angel‘s earth‘ for heaven)
49 Bode, Die Kenningar, 12. 50
Bode, Die Kenningar, 16.
17 5. ―episodische Kenningar‖ (Wêlandes ӡeweorc ‗Weland‘s work‘ for sword)
According to Bode, all kennings have their own group they belong to, though many kennings fit in more than one of the groups.
Bode calls kennings ―indirekte und allermeistens konventionelle, formelhaft Benennungen.‖52
Their indirectness and mostly conventional and formulaic naming is what sets them apart as a poetic device, and they seemed to be a popular device as well. Bode states that, on average, in every nine or ten Anglo Saxon verses a kenning can be found.53 He also mentions that most kennings found in heroic epics are about battle, life and death.
Bode argues that alliteration is one of the reasons why we have kennings: ―Kenningar und Appositionen … dienen namentlich zum Flicker, zur Ausfüllung der zweiten Halbzeile, zum Weiterkommen.‖54
They are a decoration in the second halfline of a poem. According to Bode, if it had not been for the alliteration that we find in Anglo-Saxon poetry, the number of kennings and appositions would be about a third of what we have now.
Bode suggests that not only poets play a part in the forming of kennings: ―[N]icht etwa nur Dichter … sondern oft auch Prosaiker, Gelerthe.‖55 The riddle-like qualities of kennings seem to be a reason for him to argue this. He ends with listing kennings that have been found for fifty-four concepts.
52 Bode, Die Kenningar, 9. 53 Bode, Die Kenningar, 11. 54
Bode, Die Kenningar, 14.
18
Rudolf Meissner, Die Kenningar der Skalden (1921)
Rudolf Meissner‘s Die Kenningar der Skalden is one of the most important works on kennings. Though many scholars do not agree with his opinion on kennings, his work is the foundation of many discussions on this poetic device. According to Meissner, there is no general definition known from Snorri‘s texts. He argues that Snorri merely gives examples and no good definition for the poetic device. Snorri‘s examples are also of one type, while Meissner agrees with Bode that there are more types of kennings. Meissner discusses several types of kennings, one of which he calls ―Kenning im engeren Sinne‖:
Wenn die Umschreibung im Grundwort einen Ausdruck hat, der nicht unmittelbar seinem Wortsinne nach mit der Bestimmung verbunden werden kann, sondern, erst durch eine Hülfsvorstellung, ein Vergleichen, ein Übertragen, eine bildliche
Auffassung, dann ist ein von dem vorher getrenden sich merklich unterscheidender Typus vorhanden[.]56
The type of kenning Meissner seems to refer to here is a metaphorical kenning.
Meissner also looks at the structure of the kenning, which is a replacement for a noun consisting of two parts: ―Die einfache Kenning ist also ein zweigliedriger Ersatz für ein Substantiv der gewöhnlichen Rede.‖57
Thus, he starts off by looking at the structure of the kenning, which is a two-element replacement for a noun. Later on, he specifies that for a kenning it is important that it is seen as replacement and as such has a universal, typical quality. He also mentions that the word ‗kenning‘ and its meaning derives from the use of the
56
Meissner, Kenningar der Skalden, 27.
19 verb kenna ‗to know, teach, recognize‘ in the Edda.58 Meissner looks at the structure of the kennings when forming his definition, while Bode looked at the function of the kennings. Only later on in his work does Bode look at the structure more in depth.
Meissner gives the two elements of a kenning the names Grundwort and Bestimmung. The Grundwort, the base word, takes over the role of a noun in the sentence; this base word itself is also a noun. The Bestimmung, the determinant, is what gives a different connotation to the base word, and so, changes the meaning of the entire term. For example, in hreinbraut ‗path of the reindeer; earth‘ braut is the base word, hreinn the determinant. A characteristic of most of the kennings is that the determinant is only formed out of nouns although, as
Meissner continues, ―Umschreibungen mit anderer Bestimmung könnte man gewiß an sich auch Kenningar nennen.‖59
This rather small group is, according to Meissner, not part of the kenning system. Here his comment does not help with finding the right definition, for first he specifies a characteristic, then states that ‗Umschreibungen‘ without the characteristic can also be called kennings. However, Meissner does recommend to only categorize the descriptions built up out of two nouns as actual kennings. A rule Meissner mentions about forming a kenning is that the meaning of the entire kenning cannot coincide with the meaning of the
Grundwort.
A great number of kennings have a nomen agentis as base word. A nomen agentis (agent noun) is a noun that identifies an entity through a typical activity. In the case of a kenning this would for instance be líðendr ‗wayfarers; men‘. A nomen agentis occurs very often in man-kennings; however, it is also used for animals and objects: By using
personification in the kenning, the created image becomes more vivid. For these
personifications, masculine agent nouns are used. Kennings for ‗woman‘ constructed out of a
58
Meissner, Kenningar der Skalden, 2.
20 base word that is an agent noun are very rare, for example kveldriða ‗evening rider; witch, hag‘.
Meissner also mentions the structural types Snorri discusses; the tvíkennt and the rekit: ―Wird die Bestimmung einer Kenning wieder durch eine Kenning ausgedrückt, entsteht eine Doppelkenning (tvikennt), jede weitere Verlängerung durch neue Kenninger heißt rekit.‖60
Meissner refers to a group of ―zusammengesetzten Umschreibungen‖61
ending in –nir. This group does not occur often in skaldic poetry. However, in the Þulur62 heiti and names are found that are partly constructed like kennings.63 Meissner continues by stating that: ―Der zweite Bestandteil [the base word] kann eine deutliche verbale Vorstellung enthalten und ein entsprechendes Verbum neben sich haben; ist der erste Bestandteil [the determinant] ein Subts., so liegt eine normale Kenning vor.‖64 One of the examples Meissner provides here is
hreggskornir ‗storm cutter; eagle‘. Here, the base word is skornir, a noun that is derived from
the verb skorinn ‗to cut‘. The base word contains a verbal, or verb-related image. The determinant is the noun hregg ‗storm‘ which in combination with the base word forms a
normal kenning.65
Finally, Meissner states that not every nominal compound, even if its elements form one image, is a kenning. If both elements can be taken in a literal sense, they do not form a kenning; still, the semantic context of the compound may make a literal interpretation
unlikely. An example Meissner gives is that of the King of Norway. Meissner argues that the King of Norway is hardly ever represented as the ruler of the entire country in a kenning. He is represented as, for instance, Lista þengill, ‗Lista-prince‘, Lista being a place at the southern
60 Meissner, Kenningar der Skalden, 40. 61 Meissner, Kenningar der Skalden, 10.
62 ―[N]ow used for strings of rhymes running on without strophic division and of rhymed or alliterative formulas.
The þulur include wide areas of poetic vocabulary and give names of gods, dwarfs, giants, valkyries, sea kings as well as synonyms for man, woman, weapon, battle, sea, ship etc.‖ Dictionary of Northern Mythology, comp. Rudolf Simek (Rochester: University of Rochester Press, 1993), s.v. ―þulur.‖
63 See Finnur Jónsson, Arkiv för Nordisk Filologi, vol. 35 (1948): 302. 64
Meissner, Kenningar der Skalden, 10.
21 coast of Norway.66 ‗Lista-prince‘ should not be taken literally. The base word, in this case a base word for Norway, is replaced by a Teilbegriff of this intended base word.
Meissner also looks at the kenning‘s function within the text. Poets do not repeat kennings but use new, similar words instead in order to change a dull, flat expression into a lively, more colorful one.67 Furthermore, Meissner notes the (obvious) playful use of a word‘s ambiguity. ―Wird bei einem mehrdeutigen Worte der Sinn, der zum Zusammenhange nicht paßt, durch eine Kenning ausgedrückt, so ist die Kenning zum Rätsel geworden.‖68
This characteristic gives the kenning its riddle-like quality. An example of this play on ambiguity is heit in kǫld heit.69 In this case heit means ‗vow‘ but the combination of kǫld and heit also evokes the contradiction of ‗hot‘ heitr and ‗cold‘ kaldr in the audience‘s minds.
Meissner proceeds his work by giving a long list of concepts and its kennings.70
E.O.G. Turville-Petre, Scaldic Poetry (1976)
Turville-Petre starts off by giving Snorri‘s definition of a simple kenning as given on page eleven of this dissertation. According to him, this is the clearest definition. Turville-Petre argues that often a kenning is not literally the concept it substitutes, but that it acts similarly, or has the same function as the concept. An example of this is brimhengest ‗surf-horse‘: A ―ship is not a horse, but it sails over the sea as if it were a horse moving over land.‖71
Like Meissner, Turville-Petre refers to Snorri‘s three groups: kennings, tvíkennt ‗doubled‘ and rekit ‗chased, driven‘.
66 Meissner, Kenningar der Skalden, 14. 67
Meissner, Kenningar der Skalden, 25.
68 Meissner, Kenningar der Skalden, 83. 69 Meissner, Kenningar der Skalden, 83. 70
See Appendix A.
22 Furthermore, Turville-Petre argues is that it is necessary to have knowledge of the beliefs and legends of the Medieval Scandinavian people.72 For some kennings it is not always easy to decipher whether they derive from a myth or whether the poets actually believed the story behind the kenning to be true. An example of this is, for instance, a gold kenning like fjarðeldr ‗fjord-fire‘. The poet could be referring to a myth or he may actually believe that there is gold in the water, though, the latter reasoning seems rather naïve to me. In addition, Turville-Petre states: ―It is doubtful whether we could understand kennings like these if Snorri had not explained them in the Skáldskaparmál.‖73
Turville-Petre also mentions an unusual type of kenning called ofljóst ‗over-light, obvious‘, ―although such kennings are exceptionally obscure, since they depend on word-substitution.‖74
These kennings use word play, ―often by substituting homonyms when one is a proper name, i.e. using a proper name or a kenning for it for the common noun
equivalent.‖75 An example Turville-Petre give is Hergauts vina ‗Hergautr‘s mistress‘. Hregautr is another name for Oðinn and one of his mistresses was Jorð. Here, the intended, literal concept ‗earth‘ is replaced, using the goddess‘ name.
There are also half-kennings; however, these are very difficult to explain. They might be abbreviations of at that time well-known kennings and seem closely related to heiti. A good example is hauks holms Jorð ‗goddess of the hawk‘s island‘. This kenning probably means ‗goddess of the arm, woman‘. However, here the determinant is a half-kenning: ‗goddess of the fire of the hawk‘s island (of gold)‘.76
Eddaic poets developed the kenning close to the way the skalds formed them, more so than Old English poets. They drew heavily on legends or myths; furthermore, Turville-Petre
72 Turville-Petre, Scaldic Poetry, xlix. 73 Turville-Petre, Scaldic Poetry, l. 74
Turville-Petre, Scaldic Poetry, li.
75 Faulkes, Skáldskaparmál 1: Introduction, Text and Notes (London: Viking Society for Northern Research,
1998), lxxii.
76
23 observes that ―when they [the poets] used extended kennings, it was probably under influence of the scalds.‖77
He states that the kenning is not always successful in bringing across its meaning. However, when kennings are used skillfully, a rich mental image is created. In this way, the kennings become a kind of miniature myth/legend. Nevertheless, even kennings without allusion can still create a very vivid image.78 When referring to ‗battle‘ as hjǫrregn ‗rain of swords‘ the kenning does not use allusion to convey the meaning but still provides a lively picture for the audience. However, there is also a danger to this mythological kenning type: They can become so ordinary that they become stereotypes: ―When poets combined names of gods with words for ‗battle, weapon‘, to denote simply ‗man, warrior‘, we may wonder whether they had any clear conception of the god named.‖79
In these cases it might have been the metrical rules that, as it were, determined the poet‘s kenning choice.
Turville-Petre also discusses the development of the kenning. He states that, though the kenning was already fully defined in Bragi‘s works, later poets have definitely elaborated on it and made more use of the kenning. Still, we can never be certain: ―Any generalizations which we may make about the development of the kenning in the tenth and eleventh centuries are suspect: although much skaldic poetry is preserved, we see no more than the tip of an iceberg.‖80
Finally, Turville-Petre discusses the relation of kennings with pagan religion, and argues that this relationship prevented certain skalds from using many of these ‗pagan
elements‘ in their works: ―Great proportion of the surviving skaldic poetry from late tenth and early eleventh centuries was dedicated to fanatical Christian kings.‖81 Later on in the twelfth century, it seems that poets are allowed to use these pagan kennings again.
77
Turville-Petre, Scaldic Poetry, lv.
78 Turville-Petre, Scaldic Poetry, lvii. 79 Turville-Petre, Scaldic Poetry, lviii. 80
Turville-Petre, Scaldic Poetry, lviii.
24
Roberta Frank, Old Norse Court Poetry (1978)
In her book Old Norse Court Poetry — The Dróttkvætt Stanza, Roberta Frank first defines the Old Norse kenning, which clearly is similar to Meissner‘s definition; it ―is a periphrasis, consisting of two or more substantive members, which takes the place of a noun.‖82
Frank also discusses the formal and structural aspects of kennings. The base word of a modern kenning is more concrete while its determinant is rather abstract. Old kennings were copied, though the original was rarely repeated completely since both parts could be replaced by a heiti. Brimdýr ‗surf animal‘ could become brims blakkr ‗horse of the surf‘ or bóru fákr ‗horse of the wave‘.83
According to Frank, one of the most characteristic circumlocutions is a phrase of the type ‗ship of the desert‘. These circumlocutions are formed out of two terms. In this case, ‗ship‘ is the base word and ‗desert‘ ―is a noun to which the base word is made to relate.‖84
In Old Norse, the latter word is often a genitive, though the circumlocution can be a substantive compound as well, with the first part limiting the second part. ―Taken together, the nouns have a significance that neither has separately and this meaning is never that of the base word.‖85
Just like Meissner and Turville-Petre, Frank refers to the way in which kennings can be expanded by more elements. According to Frank, there is only a theoretical limitation to kenning length: The ―syllabic capacity of the half-stanza.‖86 Franks refers to Snorri‘s suggestion to not use more than five elements in one kenning ―to keep within the bounds of good taste.‖87
82 Frank, Old Norse Court Poetry, 42. 83
Frank, Old Norse Court Poetry, 45.
84 Frank, Old Norse Court Poetry, 42. 85 Frank, Old Norse Court Poetry, 42. 86
Frank, Old Norse Court Poetry, 45.
25 Frank lists the semantic groups from which the base words are selected to form
kennings for men: divinities, masculine tree names, verbal derivatives like ‗destroyer‘, ‗feeder‘, ‗terrifier‘. The determinant is often selected from the following groups: gold and silver, battle, weapons, armor and ships. Then Frank proceeds with the same type of list for women: goddess names, feminine tree names, feminine words for ‗earth‘ are used to form the base word of the kenning, and the determinant is chosen from the groups ‗clothes‘, ‗jewelry‘ and ‗servings‘.88
According to Frank, circumlocutions need a context to relate their meaning; ―their poetry does not come from laboratory analysis but from photosynthesis.‖89 Frank depicts how the meaning of a concept is changed by a kenning, that is how the kenning becomes more than just words replacing the original noun:
When an Icelandic woman is called ―Freyja of the stone necklace,‖ she is no longer just a simple farm girl but a cosmic force to be reckoned with; a living reality still, she becomes at the same time an emanation of something known to have existed since primeval times and to have continually influenced human destinies, an Aphrodite whose neck ornament inflamed and distracted the mightiest of gods.90
Frank argues that Christian skalds were not afraid to use these pagan influences in their kennings. Unlike Turville-Petre, it is unclear whether Frank refers to skalds that composed their poetry after the 11th century or not. However, the use of pagan influences in the kennings and the temporary prohibition of it could explain the difficulties the audience now sometimes
88 ‗Servings‘ are ‗beverages‘, but should also be read as ‗drinking vessels‘ and ‗the bed‘. Frank, Old Norse Court
Poetry, 43.
89
Frank, Old Norse Court Poetry, 44.
26 has when trying to understand a kenning: The myth it referred to could have been lost during the time that pagan influence in skalds‘ works was forbidden.
Furthermore, Frank discusses riddle-like kennings, ―metaphoric comparisons that alert the audience to hidden correspondence in the universe, to meaningful parallels between microcosm and macrocosm[.]‖91 An example of this is foldar bein, ‗bones of the earth‘ as a kenning for stones. Frank also mentions the importance of mythology/legends. These mythological kennings were a way to contain the knowledge of the past in ―an efficient, memorable form.‖ However, with this type of kenning it might occur that the ―full force‖ of it is gone because the myth is not known anymore.92
In addition, Frank argues that, although circumlocution was not limited to
Scandinavian poetry, the skalds developed ―entire systems of interlocked kennings‖ with its own ―syntax or language of metaphoric expression‖ which depended on both ―previous knowledge and training‖ as well as ―feeling for or observation of nature‖ to be fully understood.93 This system was passed on from one generation of skalds to the next.
Edith Marold, Kenningkunst (1983)
Like Frank, Marold mentions the importance of Meissner‘s work. She gives his definition of a kenning being a two-part replacement for a noun that follows a certain traditional pattern and variation to construct kennings: ―Die Kenning entsteht daher nicht wie das dichterische Bild aus einem individuellen Erfarungsprozeß des Künstlers, sondern durch Wahl aus einem durch die Tradition vorgegebenen Formelschatz.‖94
Marold identifies four different functions that a
91 Frank, Old Norse Court Poetry, 44. 92 Frank, Old Norse Court Poetry, 44. 93
Frank, Old Norse Court Poetry, 44-5.
27 kenning can have. These functions are: Sprachspiel (wordplay), Poetisierung (poetification),
Preis (praise) and Kontrast (contrast).95
Marold gives a schematic illustration, a kind of formula to express the definition of a kenning given by Heusler: ―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[.]‖96
An example of this is hildenædre ‗battle adder‘. In this case A is the concept ‗arrow‘, B is ‗adder‘ and C is ‗battle‘. A is not B; an arrow is no adder, but because of C it attains qualities of B.
Marold notes that there is still no general consensus on the definition of a kenning. She mentions several definitions among which Meissner‘s — which she thinks is too broad — and Bode‘s. The question she asks is whether a kenning should be regarded as a metaphor or as a description (Umschreibung). Skaldic poetry contains a lot of metaphorical kennings. In these kennings the Grundwort is a metaphor and is the kenning type Meissner refers to as ―Kenning im engeren Sinne.‖ She gives the example of the ‗horse of the sea‘ for ‗ship‘. This type of kenning played an important part in making the definition for kenning: ―Dieser Kenningtyp wurde allmählich als hauptsächlicher Vertreter der Kenning angesehen, begreiflicherweise, da man hier so etwas wie eine Bildwirkung erahnte.‖97
For a metaphor to be understood, the context wherein it is found is needed. Without such context it would not exist. Where others state that this is also the case with kennings, Marold argues the opposite: ―Die Kenning existiert vor und außerhalb des Gedichts, und es ist nicht der Satzkontext, der die Metapher in der Kenning „zur Metapher macht―, sondern der innere Kontext, das Bestimmungswort.‖98 For kennings it is not necessary that the kenning for
28 which the concept is made refers to the context, which is the case with fixed epithets in
epics.99
She tries to find characteristics with which she can divide the kennings into types. For this she uses two principles: The category of the described concept, and the kenning‘s base word and its relation with the concept. She mentions four types:
1) Metaphorischen Kenningar: The relation between base word and the original concept is similarity or analogy. (Hǫgna meyjar hjól, ‗wheel of the maid of Hogni, shield‘)
2) Synekdochischen Kenningar: The relation between base word and the original concept is that they both belong to a common concept. (hræva dǫgg, ‗dew of corpses, blood‘)
3) Metonymische Kenningar: The base word expresses a function or feature of the original concept. (marker mein þjófr, ‗hurtful thief of the fields, fire‘)
4) Abgeleitete Kenningar: No connection between base word and original concept. This type is part of the group of metaphorical kennings, because they derive from them. (lauf runna Leifa landa, ‗leaf of the trees of the land of Leifi, shield‘)100
Gary Holland, Kennings, Metaphors, and Semantic Formulae in Norse Dróttkvætt (2005)
Holland forms his theory by looking at the metaphorical characteristics of the kenning. The structure of the kenning is not that important to him. He starts off by stating that the discussion on kennings has mostly been on whether or not a kenning is a metaphor. The
99
Marold, Kenningkunst, 212.
29 reason for kennings not to be metaphors can be aesthetic: ―[S]ince kennings do not have the same aesthetic effect as metaphors in the western poetic canon, they cannot be metaphors.‖101
However, according to Holland, the problem is the too narrow literary definition we have for a metaphor. Metaphors are found in everyday language as well as poetic language. However, there is a difference: ―Poetic thought uses the mechanisms of everyday thought, but it extends them, elaborates them, and combines them in ways that go beyond the ordinary.‖102
There is more to poetic language than everyday language. Holland states that typically the metaphors in kennings are thought to have been formed through culture and mythology, but that many scholars then conclude too hastily that there is no relation between these metaphors and ―ordinary linguistic structures[.]‖103
Holland argues that separating these two too strictly caused more difficulties with interpreting the kenning and ―an obscuring of the relationship between poetic language and ordinary language.‖104
Holland also argues that scholars have focused too much on ―specific properties of kennings‖ as well as on interpretation problems, and with that, they have overlooked an important function of the kenning. Looking at the way they are formed and their function within the sentence, they seem to be like other oral poetic formulae. There is a difference though:
The principal difference between kennings and other formulae is that because of the metrical requirements of skaldic verse (principally the alliterations and internal rhymes), kennings are formulaic solely on a semantic level rather than on both a semantic and a surface-syntactic level, as are the formulae in other traditions.105
30 Holland refers to Frank‘s example of a ship that has become a bear in one of Márkus
Skeggjason‘s poems. In the poem in question, there are four kennings that make this
identification: vetrliði rastar ‗bear of the current‘, íugtanni húna ‗bear of the masthead‘, bjǫrn
flóðs ‗bear of the flood‘ and bersi skúrǫrðigr ‗storm-breasting bear‘. All four words for ‗bear‘
can also be explained by themselves: vetrliði also means ‗one who has passed a winter‘, the origin of the word íugtanni is not clear, though it seems to have the word tooth in it and is according to Holland a sort of heiti for ‗bear‘. Bjǫrn is originally a color, making it a metonym for ‗bear‘. Bersi derives from bjǫrn. There is nothing that clearly indicates why ships and bears are alike, though there must be a metaphorical link to make this possible:
Presumably, the appropriateness of the bear-kennings in this context have to do with bears as large, fierce animals that break through obstructions and obstacles, here those characteristic of a sea voyage from Iceland to Norway. And of course there is always the possibility that the ship was named after a bear, that it was associated with a bear because of its color, or that it carried a bear-ornament.106
The end result of this kenning for ‗ship‘ also changes the kenning for ‗sea‘ that would be used in combination with the ‗ship‘ kenning: ―[W]hat we have is ‗the bear went over the mountain‘ transferred to a maritime context.‖107
According to Holland there has to be a metaphorical link between ‗bear‘ and ‗ship‘. Vertliði ‗one who has passed winter‘ shows this connection between ‗bear‘ and ‗ship‘; a bear searches cover from winter by going into hibernation, a ship is propped up on land and gets covered up. However, the way in which skaldic verse is unraveled into nominal phrases makes it lose its quality as poetry and this, according to Holland, caused most of the scholars to dismiss the metaphoric quality of kennings.
106
Holland, Kennings, 125.
31 Furthermore, Holland discusses Bjarne Fidjestøl‘s kenning-definition consisting of six parts:
1) Circumlocution (kenning replaces an ordinary noun) 2) Two components (Base word and determinant)
3) Semantic incongruity (parts of the kenning do not represent the whole it replaces) 4) Domain limitation (Meissner‘s list has 106 different concepts, but most are
assignable to fewer than twenty of these) 5) Kenning variation (Base word and determinant are open classes) 6) Rekit (The possibility of replacing the determinant by a kenning)
Holland argues that points four, five and six support his own argument in that they ―highlight the limitedness and the productivity of the kenning system.‖108
These three points show the distinction between kennings and metaphors and that they ―are to some extent
quantitative.‖109
Holland proceeds to show this limitedness by calculating the number of possible kennings for ‗gold‘ formulated as ‗fire of water‘: In Meissner‘s list of kennings there are 39 base words for ‗fire‘ and 64 determinants for ‗water‘, this results in 39 x 64 = 2496 possible kennings.110
Holland seems to be quite fixed on the kenning being a ―blend of metaphor and metonymy‖, as well as the argument that kennings were used in order to satisfy the metrical requirements of the poem.111 Holland argues that the kenning often contains ―a blend of metaphor (on the base-word level) and metonymy (on the determinant level).‖112 The kenning‘s metaphorical and metonymical qualities give this poetic device the formulaic
108
Holland, Kennings, 129.
109 Holland, Kennings, 129.
110 Holland here uses an example presented by Bjarne Fidjestøl‘s. Holland, Kennings, 129. 111
Holland, Kennings, 139.
32 characteristic Holland referred to earlier, especially when the limited number of concepts for which kennings are formed is taken into account: ―[I]t is the kennings, with their formulaic, fixed semantics and their metrically conditioned instantiations that bridge the gap between the requirements of a fixed text and oral composition.‖113
Holland uses Charles J. Fillmore‘s notion of semantic frames, adopting Fillmore‘s idea to kennings. Fillmore explains what he means with ‗frames‘: ―[A]ny system of concepts related in such a way that to understand any one of them you have to understand the whole structure in which it fits[.]‖114
When one of these concepts is used, others that belong to the same structure should automatically come to mind. For one scene, there are several elements that could come to mind to describe the concept. In a frame, a sort of ‗scene‘, different words that relate to the concept in different ways will be seen, but are all aspects of the frame. An example Holland gives is for the scene ‗commercial event‘. Words, or, as Holland calls them, elements that are aspects of this scene are buy, sell, pay, spend, cost, charge.115 Referring to the structured presentation of a scene as ‗frame‘, Fillmore argues that the meaning of the word is structured by the frame, ―and that the word ‗evokes‘ the frame[.]‖116
Holland argues that Fillmore‘s frame notion is easily applicable to kennings; if we replace ‗commercial event‘ with one of the concepts designated by kennings, it is easy to discern the right elements that belong to this new ‗scene‘. If we take for instance the scene ‗battle‘, elements that
immediately come to mind — and form the frame for this scene — are: clash, swords, arrows, blood, death.
Like Turville-Petre, Holland also mentions the importance of Snorri‘s Edda. It does not only help with interpretation of kennings, but also gives insight into the kenning itself: ―[A]ny reference to any part of a known mythological scene or a cultural topos automatically
33 calls up all the remaining aspects of the scene.‖117 Mythological works provide the scenes (Fillmore‘s frames) as well as information on how the kenning is formed. However, not only mythology but also real world knowledge is necessary. Holland calls the difference between mythological scenes and ordinary scenes ―a continuum[;]‖ both have ―cultural attitudes‖118
as well as real world knowledge contributing to the making of kennings. These scenes could be seen as two separate groups, but Holland sees them as a succession of one another. A king‘s generosity is something found in mythology but also in real life. A warrior‘s courage is a subject found in many mythological works yet also occurs in the real world.
Holland, like many others, also refers to the density of kennings in works, which is particularly high in Old Norse court poetry. Skalds used kennings more often than other poets. It seems that they could write about any subject; however, Holland does not think that is true and argues that there is a restriction as to which subjects were suitable to discuss in the skald‘s work. These restrictions do not seem to be imposed by the subject itself, but by the lack of appropriate language necessary to deal with the subject. An example is the story of King Haraldr Harðráði who saw a fight between a smith and a tanner. The king asked the poets that were with him to write a poem about it. The poets declined because they thought the subject was inappropriate. The king suggested to place the scene they had witnessed between the smith and the tanner as a fight between Sigurðr and Fáfnir. Holland argues that the poets objected because they did not have the proper language to write their poems. Only after the King suggested to place the fight into a ―North Germanic heroic tradition‖119 the poets were able to work the fight into a poem by being able to use the kennings coming ―from the traditional semantic and thematic stock.‖120
34
Sharing My Wordhord121
Personally, I feel that Meissner‘s definition of the kenning is, as others also stated, too broad. However, as Meissner argues, even though it is very tempting to qualify heiti and kenning as the same because of their shared characteristics, like their metaphorical quality, they
definitely a kenning exists of two or more elements, always containing a base word and a determinant. Neither the base word nor and determinant can never have the meaning of the concept the kenning replaces. Only when the elements are put together, the concept‘s meaning is formed. Structurally, there are the simple kennings, tvíkennt, and rekit.
Metrically, kennings were formed to maintain the required alliteration throughout the lines. However, I do not think that metrical consideration were the main reason for poets to use kennings. Obviously, poetry is not ‗average language‘; there is more to it. The riddle-like quality of the kenning testifies to that. Despite the fact that poets used kennings for metrical purpose, I do not agree that kennings only exist because of alliteration, even though the former is subjected to the latter; kennings would ‗work‘ even without following the line‘s alliteration.
I agree with Holland‘s argument that most kennings having metaphorical and metonymical qualities. The example of the bear/ship kennings is clear evidence for that. Marold‘s rather detailed kenning type system seems to me to be the best and most likely system that has until now been defined.122 It is very detailed with all its subdivisions and makes it easy to distinguish the kind of kenning found in a text.
Snorri‘s Skáldskaparmál is the source ‗closest‘ to the Old Norse language of poetry, giving insight into its origin, but it does not give the ideal definition for a kenning; instead, it
121
Wordhord is a kenning found in Beowulf (line 259) which means ‗store of words, knowledge, thought.‘
36 Chapter Three: Analysis of the Kenning
In this chapter I will have a look at several Old English texts to try and determine whether it is possible to connect kennings to specific rules, or possibly a formula, that poets used when they were composing their works. The characteristics that are known to be part of the kenning could possibly lead to the discovery of a fixed pattern that was used to determine when and where kennings were allowed to be used in the text.
For this analysis I focus on texts that deal with the theme ‗battle‘; the kennings I use here are, thus, all related to ‗battle‘ in some way. The Old English texts I use are:
- Beowulf
- The Battle of Brunanburh - Exodus
- The Fates of the Apostles - The Finnsburg Fragment - Genesis A/B
- Judith - Juliana
- The Battle of Maldon
With the theme ‗battle‘ in mind I have selected several concepts related to this theme. By looking at groups of kennings for related concepts I hope it will be easier to see whether there are distinct patterns that can be found regarding the use of the poetic device. The concepts I analyze are:
37 - Death - Helmet - Shield - Spear/arrow - Sword - Warrior - Wound
In order to find out more about the use of kennings, the analysis of these concepts could help to find answers to the following questions:
- Does the place of the kenning matter? (halfline A or B) - Does the place within the halfline matter?
- Does the case of the kenning matter?
- How does the kenning relate to the alliteration of the two halflines? - Are kennings, if at all, re-used/repeated?
- Is it possible to find a certain rule or formula regarding the use of a kenning in the two halflines?
In order to answer these questions I have looked at the kennings in different texts, but also at how the kennings occur within a single text. I used Beowulf to see how one poet handled the use of kennings.
The Place of the Kenning
38 kennings were used in halfline A or B, or whether it does not matter where the kenning is placed. Of the battle-related kennings I found about 73 percent placed in the first halfline. It seems that using a kenning in that halfline was favored. However, the 27 percent of kennings in halfline B do show that it is not the case that kennings are only used in one of the
halflines.123
This favoring of halfline A could be related to the oral tradition that was the basis to the works. It could be that having a kenning in the first halfline made it easier to remember. Unfortunately, this can in no way really be proven. In the end, although halfline A seems to be favored, it is clear that kennings are useable in either halfline, and so, there is no rule as to where the kenning is placed regarding the two halflines.
It could be that there is a specific place in the halfline that is reserved for kennings only. I started from the hypothesis that poets preferred to use kennings in certain places only, possibly in places where either they wanted the audience to remember the text better, or in places where they wanted to make a statement. In fact, the kenning seemed to be mostly placed at the end of the halfline, either as last word or last two words, or sometimes as only word of the halfline: about 26 percent of the kennings are placed as first word of the halfline, 74 percent has the kenning as the last or penultimate and last word, or as single word in the halfline.124 Again, there is not one exclusive place in the halfline reserved for the kenning, but certain places seem to be favored.
The Kenning and Alliteration
123
See Appendix C. For a list of all the kennings analyzed see appendix D.
39 As expected from what was discussed in chapter two, the kennings are almost always part of the alliteration of the line. Of the kennings analyzed, 77 percent start with the letter of the line‘s alliteration. However, in about 21 percent the alliteration of the line is unclear.125
In only 2 percent of the cases does the kenning start with a different letter than the alliterating one.126 Alliteration and kenning are clearly connected with one another. The alliteration that was part of the ways of the Old English poetic tradition was easier to follow by using kennings. Where using the original concept would break with the line‘s alliteration, the kenning itself can be formed in a way that it does start with the alliterating letter.
Repetition of Kennings
Of all the kennings there are only two that are the same: sweordbite, Fates l. 34, and sweordes
bite, Juliana l. 603, both kennings for the concept ‗wound‘. From the gathered data it is very
likely that, in general, kennings were not repeated.127 The poets had enough creativity to compose new kennings where needed. After all, repeating his kennings would not really testify to the poet‘s skills. However, the two parts the kenning is made up of are reused in different combinations. It does seem that certain nouns were preferred in kennings, the determinant re-occurring in combination with a different base word. This clearly shows the determinant is the most important of the two parts in making the actual kenning; it is the part that assigns the kenning to the concept it represents.128
125 In several lines the alliteration cannot be discerned because there is not one letter that occurs more often in an
alliterative pattern.
126
See Appendix C.
127 Even though the two kennings mentioned are similar, they do occur in different texts, which supports the
notion that kennings were never repeated. However, what must be noted is that both The Fates of the Apostles as well as Juliana are thought to have been written by Cynewulf.
40 All in all, the kennings are not repeated and the great variety of them suggests that it was almost a challenge for the poet to come up with his own kennings, a poetic struggle to make them fit his work perfectly by designing them himself.
The Kenning and Its Case
When focusing on the use of battle kennings in one text, in this instance Beowulf, the results are all the same as those for the use in general. However, with Beowulf, I also analyzed the case of the kenning, to see whether a kenning would occur in one particular case only. Again, unfortunately, there is not one definite answer to this question: The kennings are not written down in one single case. However, one of the options is favored again. About 41 percent of the kennings, both singular and plural, are written down in the accusative case. 28 percent is in the nominative case, 17 percent is dative and 15 percent is genitive.129
Having the largest group of kennings in the accusative case suggests that using the kenning for the direct object is favored. But why this case? It could be that a kenning in the accusative case contributes to the task of remembering the text as part of the oral tradition of medieval England. With the subject of the line written down in a normal noun form, and the direct object as a kenning, the message of the line is presented in a more dramatic way, painting a better picture of what is going on. After all, it is a poetic device that was used to create a more enchanting image. For example, the sentence ―The king travelled across the sea‖ is very standard, an image that we all know, but a very basic image at it. By calling the direct object (the sea) a ‗whale-road‘ the created image is immediately different. Now, the whale-road kenning creates the whole idea of a rough sea with huge whales coming up for air
