Nominal and Verbal Derivation in Old English

Nominal and Verbal Derivation in Old English* : The Case of Ge-words 1)

Yoo-kang Kim (Hankuk University of Foreign Studies)

Yoo-kang Kim. 2007. Nominal and Verbal Derivation in Old English: The Case of Ge-words. Language and Linguistics 40, 21-45. This paper is concerned with morphological derivation in Old English, specifically nominal and verbal derivation without affixation. Non-affixational derivation, which is traditionally called conversion, has been controversial in Old English word-formation. Morphologists' claims differ in the issues of the directionality of derivation and the postulation of a zero-morpheme. For example, Kastovsky (1968, 1992, 1996) argues for zero-derivation, postulating a zero-morpheme in Old English morphology while DaltonPuffer (1992, 1993), Ritt (1993) and Lass (1993) give evidence against the assumption of zero-derivation in favor of affixless derivation (cf. Plag 2003). Concerning the directionality of affixless derivation, any principles or criteria for determining the directionality has not been offered in Old English morphology. Aiming to give alternative solutions to these two morphological problems, this article provides a morphological analysis of Old English ge-words. It is shown that the derivation of various derivative forms is accounted for by the interplay of WordClass Marking and the Zero-Constraint without having the directionality problem. Furthermore, the non-occurrence of some derivatives is explained straightforwardly. Keywords: Old English, Prefix, Word-formation, Conversion, Zero-morpheme, Derivation

* This work was supported by the 2006 Research Grant from Hankuk University of Foreign Studies.

22 언어와 언어학 제40집

1. Introduction Old English (OE) word-formation has been examined by a number of articles and monographs (Marchand 1969, Fisiak 1980, Kastovsky 1968,

1985,

others).1) variety

1992,

1996,

2002,

Colman

1985,

Anderson

1998,

among

In particular, Kastovsky has published many papers on a of OE　 morphological phenomena (inflection,

derivation and

compounding). Based on Marchand's (1969) extensive empirical study of Present-day English (PE) word-formation and his morphological theory, Kastovsky postulates a zero-morpheme (ø) in order to account for

non-affixational

word-formation

conversion):

(traditionally-called,

V > [V+ø]N: [gief+an]V 'to give' vs. [gief+ø+a]N 'someone who gives'. The postulation of a zero-morpheme has been controversial in morphological theory. For example, Plag (2003) argues against the assumption of zero-derivation in favor of non-affixational conversion. Dalton-Puffer (1992, 1993), Ritt (2003) and Lass (1993) also provide evidence against the postulation of a zero-morpheme in OE derivation, arguing for zero-less derivation. Another

controversial

issue

involved

in

derivation

without

overt

marking is the directionality of derivation. Since there is no overt nominal or verbal affixes attached with a base, there is no apparent way of determining which one is the base of the other between two derivatives: a noun from a verb or a verb from a noun. Kastovsky has not suggested any principles or criteria for the determination of base. He just stipulates that morphological base can be chosen based on the historical or the semantic relationship between derivatives. Addressing these two derivational issues (the postulation of a zeromorpheme and the directionality of derivation), this paper aims to provide a morphological analysis of ge-nouns and their related derivative forms. 1) The following abbreviations are used throughout the paper: Indo-European (IE), Old English (OE), Middle English (ME), Present-day English (PE).

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 23

I argue that the directionality problem can be avoided by assuming that a base is marked for word-category by Word-Class Marking. In addition, it is shown that the presence or the absence of derivatives related with ge-nouns can be predicted and explained by the ZeroConstraint which requires a base to be affixed either by a zero- morpheme or by an overt affix in order to appear as a word at the surface. The organization of this paper is as follows. Section 2 provides a historical survey of typological change in OE word-formation and discusses main controversial issues. In section 3, I present 6 types of ge-nouns and pose main questions concerning their derivation. The following section gives main morphological algorithm and attempts to offer a morphological analysis of the data. Section 5 concludes this paper.

2. Typological Change in OE Word-formation and Controversial Issues In this section, I discuss some important issues in OE word-formation with special attention to morphological derivation of words. The first half of the section begins with a brief historical survey of typological change in word-formation from IE to OE because the structure of OE derivational morphology cannot be understood without some knowledge of its IE and Germanic antecedents. Then, I turn to the main issues of this paper, namely, the directionality of derivation and the postulation of a zero-morpheme in OE, reviewing how these issues have been handled by the previous literature.

2.1 Typological Change in Word-formation from IE to OE According to the morphological status of the input to the morphological processes, namely, the status of the base form, derivational morphology is

24 언어와 언어학 제40집

typologically

subdivided

into

three

types:

word-based,

stem-based,

or root-based morphology (Kastovsky 2006: 157, cf. Kastovsky 1996, 1992).2) In word-based morphology the inputs to the morphological processes are free lexical items, which have at least one word-form that can appear without any inflectional endings in an utterance (e.g.,

boy vs. boys). This is typically the case in PE. Stem-based morphology is characterized by the fact that the input to the morphological processes is always bound so it cannot occur as a word in an utterance unless it is accompanied by an inflectional ending (OE ner- in nerjan 'to save', oxin oxa 'ox, singular, nominative'). This is true of OE verb morphology, and partly also of OE noun morphology. Finally, root-based morphology begins from an abstract lexical element which may or may not be wordclass-specific, and which had to first undergo certain stem-formative processes in order to be combined with the inflectional system proper.3) IE morphology is root-based: the root usually being represented as a consonantal skeleton and the vowel being supplied by ablaut alternations (e.g., *Vd- 'eat' (cf. OE etan), *mVd- 'measure' (cf. OE

metan), *trV- 'tremble' (cf. OE þrēat)) (Kuryłowicz 1968: 200ff, Szemerényi 1996: 102ff).4) The actual paradigms are derived by first 2) "A word in this sense is basically a free form, and can occur in an utterance without additional material such as inflectional or derivational morpheme" (e.g., cat in cats); "a stem is a bound, word-class-specific lexeme representation stripped of any inflectional endings, but potentially containing derivational affixes or stem-formatives, which determine the inflectional category of the lexeme in question" (e.g., scient- in scientist); "a root is the element that is left over when all derivational, stem-forming, and inflectional elements are stripped away." (e.g., IE *Vd-, OE etan 'to eat') (Kastovsky 2006: 157) 3) Bauer (1983: 16) defines a stem formative as a "distributional segment of a word-form independent of whether or not it is also a morph." This term sometimes called as theme is used to denote an element which, when added to a root, forms a stem to which inflections may be added. Thus, Germanic *luf-ōj-an 'love' consists of root + stem formative (theme) + inflectional ending. See Colman (1985) for the discussion of some morphological formatives in OE.

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 25

adding stem-formatives and then inflectional endings: Root + Stemformative

+

Inflectional

endings.

Since

word-class-specific

properties

are added to the roots by the various morphological processes which derived word-class-specific stems, many of IE roots are word-classneutral (Kastovsky 1996: 110). The addition of stem formatives and of the appropriate endings (case, number, aspect, person) produces primary nouns or primary verbs. This means that at this stage there is no direct derivational connection between a verb and its derived noun or a noun and its derived verb; both are only related via their common root (Kastovsky 2006: 162). Put differently, the stem-formatives are added to roots, producing verb-stems or noun-stems, to which finally the appropriate inflectional endings are added. The evolution of the Germanic language family is characterized by a transformation of its phonological and morphological

system.

In

Germanic, stress came to be fixed on the first syllable and the stress shift eventually led to a weakening and ultimate loss of medial and final unstressed syllable, which in turn resulted in a growing loss of stemformatives and inflectional endings. This development brings about a gradual shift from root-based to stem-based morphology in OE (Kastovsky 2006: 163). This typological change is exemplified below in (1). (1) The basic structure of weak verbs in Germanic and OE (Hogg 1992: 157, Kastovsky 1996: 104, Kastovsky 2006: 164) Germanic Infinitive Present

*trum + j + an-az *trum + j + is

Preterite

*trum + i + d + a

OE trymm + ø + an trym + ø + is trym + ed + e

As shown in (1), the Germanic root *trum- is followed by a stemformative *-j-/*-i- by which the root is derived into a causative verb 4) V indicates the ablaut vowel. Ablaut is a variation in the root vowel according to tense and number. The symbol "*" indicates an unattested form.

26 언어와 언어학 제40집

and then inflectional endings occur. However, in OE the stem formative is no longer segmentable because they have either fused with the inflectional endings, or they syncopated leaving the i-umlaut trace in the stem (*trumj- > *trümj- (i-umlaut) > *trüm (syncopation of *-j-)> OE trym-).5) With nouns, the development is similar. In Germanic the original stem-formatives have begun to merge with the case/number exponents, losing their class-marking function: e.g., a-stem: *dag+a+z

> OE

dag+as (plural nominative), *dag+a+ns > OE dag+as (plural accusative). The stem-formative *-a- became a thing of the past, and the form (root (*dag) + stem-formative (*-a-) + inflectional ending (*-z)) came to be reanalyzed as stem (dag-) + case/number endings (-as) in OE morphology. Clearly this implies a shift from root-based to stem-based inflection. The next stage of typological change is the loss of some inflectional endings in OE. For example, the nominative/accusative singular ending *-as of a-stem nouns is lost, resulting in uninflected forms (*dagas > OE dag). Consequently, OE noun morphology comes to allow some uninflected forms in morphological paradigms and the uninflected forms (words) can be interpreted as unmarked base forms. This is the case where we speak of word-based morphology. Namely, the OE endingless nominatives/accusatives like cyning, stān, function as unmarked baseforms with word-status. Here, inflection and derivation become wordbased, and it is this type that will eventually prevail in ME. Since there are also many weak nouns and strong feminines like gum+a and tung+e which are still stem-based, it can be stated that OE is "in a stage of transition from stem-based to word-based inflection and derivation" (Kastovsky 1992: 397). 5) In Germanic, there are four different stem-formatives for the formation of weak verbs: Class 1: -j-/-i- (Gothic satjan, OE settan), Class 2: -ōi-/-ō- (Gothic salbôn, OE sealfian), Class 3: -ē(j)- (Gothic haban, OE habban), Class 4: -nō- (Gothic fullnan, OE beorhtnian) (Kastovsky 1996: 102). With the exception of Class 3, these are still recognizable in Gothic but for OE only the first two classes are relevant.

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 27

2.2 Controversial Issues in OE Word-formation In the previous section 2.1, I have provided a brief historical survey of typological change in word-formation from IE to OE. The typological change results in a shift of morphological patterns in OE synchronic morphology,

which,

in turn,

causes a

morphological

problem. The

following section discusses the relevant problems. 2.2.1 The Directionality of Morphological Derivation

As described above, OE morphology became stem-based due to the loss of stem-formatives or the fusion of stem-formatives with the following inflectional endings. The loss of stem-formatives which produced word-specific stems (e.g., noun-stems, verb-stems) in the previous periods results in the indeterminacy of directionality of morphological derivation. The relevant examples are given below in (2). (2) The derivational relationship between nouns and strong verbs in OE (Kastovsky 1992: 392) nouns

verbs

cum+a (-an, -ena, -um) 'guest' gief+end 'someone who gives'

cum+an (-e, -est-, -eþ) 'to come' gief+an 'to give'

brec+ung 'act of breaking'

brec+an 'to break'

As shown in (2), there is no overt stem-formative between a root and the following inflectional ending so the base is unmarked for wordclass. For example, the base stems cum-, gief-, brec- are classneutral because they do not have any word-class markers. Their wordcategories are overtly determined only after the following inflection endings occur. Historically, the stems cum-, gief-, brec- of the strong verbs in (2) are verbal so the nouns (cuma, giefend, brecung) may be viewed as

28 언어와 언어학 제40집

derived nouns (verb

→

noun). Kastovsky (1968: 84-89) claims that strong

verbs are always basic, and nouns related to them with or without ablaut alternations must be regarded as deverbal derivatives. However, in a synchronic analysis of OE morphology this can no longer be maintained as a general principle because from a purely synchronic point of view, there is no overt marker for word-class in the input (stems in OE) to morphological operations. Plag (2003: 108-111) presents four possible ways of determining the directionality of conversion. The first is to look at the history of the language and see which word was first. However, simply looking at earliest

attestations

does

not

solve

the

directionality

problem.

As

described above, according to Kastovsky (1996: 99, cf. Kastovsky 1968), strong verbs are always basic and the nouns related to strong verbs are typically deverbal. However, these deverbal nouns, in turn, often served as the basis for secondary verbal derivatives (e.g., faran 'to go' (strong verb) > fōr 'going, journey' (derived noun) > fēran 'go, come, depart') (weak verb). Consequently, from a synchronic point of view, the directionality problem remains unsolved: the base stem fVr (where V indicates an ablaut vowel) can be either nominal (considering the historical relationship between fōr and fēran) or verbal (considering the relationship between faran and fōr. The second criterion is to investigate the semantic dependency between a base and its derived word. Plag (2003) states that, in general, derived words are semantically more complex than their base. Kastovsky (1996: 95) also argues for the basis of semantic dependency, demonstrating that the item which is required for the definition of the other pair is regarded as the basis (cf. Marchand 1969): PE ring (noun) > ring (verv) 'to provide with a ring', OE huntian 'to hunt' > hunta 'one who hunts'. However, the semantic information alone cannot solve the directional problem because there are many cases where the semantic relationship between base forms and derived forms cannot be clearly

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 29

determined (e.g., hwistlian 'to whistle' vs. hwistle 'whistle', cnyllan 'to strike, knock, ring a bell' vs. cnyll 'clang, stroke of a bell').6) The next criterion proposed by Plag is the inflectional behavior of forms. He states that the regularly inflecting word is derived from the irregularly inflecting word. For instance, the irregular inflectional behavior of verbs like to drink, to hit, to shake, or to sleep, he claims, is a strong argument for the deverbal nature of the nouns such as

drink, hit, shake, and sleep. However, this criterion cannot be used for derivational relationship between OE nouns and verbs because in OE noun morphology there is no way of determining regularity of nouns. Since OE has several noun declensions (a-stem, o-stem, n-stem, z -stem, etc.) which contain various inflected forms according to class, number and gender, it is not possible to tell which one is more regular than the others. The last property relevant for the determination of directionality is frequency of occurrence. Plag states that there is a strong tendency for derived words being less frequently used than their base words. This criterion may also end up with difficult cases in OE because there are many derived words in OE whose base words are less frequently words than them or even lost. According to Kastovsky (1996: 99), there are many instances where the nominal base was lost and only the verbal derivative survived in OE. I have discussed the issue regarding the determination of directionality by citing Plag's four criteria. The criteria are not successful for OE derivational morphology because many equivocal cases occur in OE. I argue that this directionality problem can be avoided by assuming 6) Even in PE, there are many cases where the semantic criterion does not lead to a clear result. For example, forms such as love (N.) and love (V.) are hard to decide upon because both have existed since OE times, and that neither of them seems to be semantically primary. Namely, to love could be paraphrased as 'being in a state of love', which would make the noun primary. However, the opposite direction can also be argued for because the noun could be paraphrased as 'state of loving', which makes the verb primary (cf. Plag 2003: 111).

30 언어와 언어학 제40집

that OE base is not marked for word-category, thus being word-class neutral at the initial stage of morphological derivation. Word-category can be specified by a separate morphological process before inflection or derivation takes place. I will turn to this issue again in section 3. 2.2.2 Conversion or Zero-Affixation?

Conversion is a morphological process referring to "the change in form class of a form without any corresponding change of form" (Bauer 1983: 32). For instance, the form napalm, which had been used exclusively as a noun, came to be used as a verb (They decided to napalm the

town) is a case of conversion. Conversion is frequently called zeroderivation, a term which many scholars prefer (Jespersen 1942, Marchand 1969). Kastovsky (1968, 1992, 1996, 2006) also strongly argue for the role of a zero-morpheme in the OE derivational morphology. Some relevant examples are given below in (3). (3) OE derived Class 2 weak verbs (Kastovsky 1996: 93)

ādlian 'to become ill' < ādl 'illness' endian 'to end' < end 'end' fiscian 'to fish' < fisc 'fish'

According to Kastovsky, the infinitive ending -ian of the derived weak verbs functions as an inflectional ending just as inflectional suffixes (-ie, -aþ, -ode) in fiscie, fisciaþ or fiscode so the denominal derivation involved in such instances was affixless due to the existence of a zero morpheme (e.g., end (base) + ø (denominal derivational morpheme) + ian (infinitive inflectional morpheme). The postulation of a zero-morpheme has been controversial and the question remains open in which particular cases it is justified to postulate a zero-morpheme. Most morphologists think that a zero-form can be justified only in those cases where there is also an overt (non-zero)

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 31

form that expresses exactly the same meaning or function as a zeromorpheme does (Sanders 1988, Plag 1999, cf. Plag 2003: 112-113). This constraint has also been called the overt analogue criterion. This means that for each type of conversion we would have to find at least one affix that expresses exactly the same range of meanings as conversion. If so, we can safely assume the existence of a zero-morpheme, if not, we have to reject it. Let us take an example of OE deverbal suffixation to see what types of semantic types they represent and to determine whether they can express the same range of meanings as deverbal conversion. Kastovsky (1985) examines the semantic types of OE deverbal suffixes including those of deverbal conversion. According to him, the following suffixes are involved in the derivation of OE deverbal nouns: -(e)d/-(o)þ/-t,

-(e)l/-ol, -els, -en, -end, -ere, -estre, -ett, -icge, -ing/-ung, -ness, -ø. The semantic types he considers include action, agentive, experiencer, objective,

factitive,

goal/benefactive,

instrumental,

and

locative.

He

argues that a ø-derivative is the most productive among the deverbal nouns and that it has a wide range of semantic types: action, agentive, objective, factitive, goal, instrumental, and locative. However, he adds that overt suffixes also produce a variety of semantic types of nouns. For example, the nominal suffix -end, according to him, produces a deverbal nouns whose semantic types include action, agentive, experiencer, objective, goal, and instrumental. All semantic types represented by a zero-morpheme are also expressed by other overt affixes even though there is the relative strength of each suffix in terms of their productivity (cf. Kastovsky 1985: 254). This means that in OE there are overt suffixes which can express the same range of semantic meanings as a zero-morpheme so it can be said that the overt analogue criterion is satisfied in OE deverbal noun morphology and that the postulation of a zero-morpheme in OE can be justified. In conclusion, in line with Kastovsky's argument for the postulation

32 언어와 언어학 제40집

of a zero-morpheme, I assume that a zero-morpheme exists in OE. Furthermore, I claim that the zero-morpheme plays an active role in OE morphology. It is shown in section 4 that whether a derivative form appears at the surface or not (i.e. attested or unattested) depends on the presence or the absence of the zero-morpheme. In sum, this section has discussed the two issues concerning OE derivational morphology, providing a critical review of some relevant previous works: the directionality of derivation and the postulation of a zero-morpheme. To handle these two issues and to provide a morphological analysis of derived words in OE, I examine the derivation of OE ge-nouns. The following section presents some relevant data and poses main research questions.

3. Data and Main Questions In order to investigate OE derivational process, I chose the OE prefix

ge- because the prefix is very productive both when preverbal and when prenominal.7) In particular, ge- is affixed both to nouns and to verbs having derived from the same base, resulting in various morphological forms: ge-noun (ge+feoht 'battle'), underived noun (feoht 'fight, battle'),

ge-verb (ge+feohtan 'to fight'), underived verb (feohtan 'to fight'). As a way of providing an answer to the two morphological questions (the directionality of derivation and conversion), I attempt to show how the affixation of ge- occurs, and how and when the word-category of the base is marked in OE. As a point of departure, I collected 38 ge-nouns found in Beowulf using Mitchell & Robinson's (1998) edition. The ge-nouns are listed 7) The prefix ge - is also used before pronouns (e.g, gehwer 'whoever'), adjectives (e.g, gecynde 'natural'), and adverbs (e.g, genoh 'enough') even though it is much less productive than before verbs and nouns.

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 33

below in (4). (4) 38 Ge-nouns in Beowulf gebedda, gebrōþor, gebyrd, gedræg, gedryht, gefēa, gefeoht, geflit, gehygd, gehyld, gelād, gemēde, gemet, gemēting, gemynd, genip, gesacu, gescād, gescæphwīle, gesceaft, gesceap, geselda, gesīð, geslyht, getrum, geþinge, geþōht, geþonc, geþræc, geþring, geþyld, gewǣde, gewealc, geweald, geweorc, gewidre, gewif, gewrixle

Then, I collected nouns (without ge-) and verbs (both with and without ge-) which are morphologically related to the ge-nouns (e.g.,

feoht, gefeohtan, feohtan related to gefeoht) by referring to Toller's An Anglo-Saxon Dictionary (1973). Finally, the ge-nouns and their related words are classified according to 6 categories shown below in Table 1. 6 categories for morphological analysis

ge-noun

Type 1

Type 2 Type 3 Type 4 Type 5 Type 6

simplex noun

ge-verb

simplex verb

0 (gefea)

0 (feoh)

0 (gefeohan)

0 (feohan)

0 (gemet) 0 (geflit) 0 (gebedda) 0 (gedræg) 0 (gehygd)

X

0 (gemetan)

0 (metan) 0 (flitan)

0 (flit) 0 (bedd)

X

Examples gefea, gefeoht, gehyld, gesacu, getrum, geþinge, geþoht, geþonc, geþring, gewǣde, gewealc, geweald, geweorc, gewidre, gewrixle gemet, gemynd, genip, gesceap, gewif geflit, gesīð, geslyht, geþræc, geþyld gebedda, gebroþor, gelad, gesceaft, geselda

X

X

X

X

0 (dragan)

gedræg

X

X

X

gehygd

(0 = attested, X = unattested)

34 언어와 언어학 제40집

Each type differs according to the presence or the absence of attested forms: type 1 having all possible attested forms (ge-noun, simplex noun,

ge-verb, simplex verb) to type 6 having only ge-nouns. Historically speaking, the absence of attested forms may be considered as the results of the loss of the forms or simply as accidental gaps. From a synchronic point of view, the distribution of attested forms of each type

poses

a

number

of

questions

concerning

their

morphological

derivation. For example, in the case of type 6 where only ge-nouns are attested, what is their morphological source? Are they derived from unattested simplex nouns (or unattested simplex verbs) (Base

→ *N → ge + N / Base → *V → *N → ge + N) or directly from a base (Base → ge + Base)? In addition, in type 6 where all possible forms are attested, where do nominal derivatives (ge-nouns) come from? Are they derived from simplex nouns (Base (Base

→ V → ge-V → ge-N)?

→ N → ge-N) or ge-verbs

These questions regarding morphological

derivation are closely related with the morphological issues discussed in the previous section and have to be taken into consideration in a morphological study. Providing an answer to the questions can give an understanding of OE derivational morphology as well as provide a solution to the long-standing problems (the directionality of derivation and conversion). Main relevant questions are summarized below. (a) What is the morphological status of the input to morphological derivation of ge-nouns? (the status of base) (b) What is the morphological source of simplex nouns or verbs? Are they directly derived from base or from the other by zeroderivation? (zero-morpheme) (c) How can the occurrence of lexical gaps (unattested derivatives) be accounted for in a morphological theory? Is it possible to predict the presence or the absence of a specific morphological derivative?

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 35

4. A Morphological Analysis of OE Ge-nouns In this section, I attempt to provide a morphological analysis of each type of ge-nouns with the assumption that a zero-morpheme exists in OE morphology. Let us first present how OE morphological operations take place and then begin to analyze the data.

4.1 Assumptions and Morphological Framework The OE Morphological algorithm used in this paper is provided below in (5) with a sample derivation. (5) Morphological operations in OE8) feoht Base (B)

↓

↓

{feoht}N

{B}Word

↓ {feoht} +{ø} ↓ N

{ge}+[{feoht}N+{ø}]

↓

↤Word-Class Marking (obligatory)

Class (WCl)

↓ ↤Zero-derivation(optional) {B} +{ø} ↓ ↤Derivation(optional) WCl

9)

{ge}+[{Base}WCl+{ø}]

↓

↤Inflection(optional)

[{ge}+[{feoht}N+{ø}]]+{es} [{ge}+[{Base}WCl+{ø}]]

+Inflectional ending 8) The notation "{ }" indicates an individual morpheme; "[ ]" morphological structure; "+" a morphological boundary. 9) According to Lexical Phonology (Kiparsky 1982, Mohanan 1986), there are two levels or strata in English derivational morphology and affixes belong to one of the two. Affixes belonging to one level are distinguished from the affixes of the other level by a number of properties (e.g,, stress shift, morpho-phonological alternations). In addition, level 1 affixes are generally less productive than level 2 affixes. In the theory, irregular inflection occurs at level 1 while regular inflection takes place at level 2. In this paper, I do not treat this leveling issue because it is not directly relevant of the main purpose of this paper.

36 언어와 언어학 제40집

As shown in (5), OE derivation begins with a base, which is not marked for word-class. The word-category of the base becomes specified at the next step by Word-Class Marking so derived stems (nominal or verbal) are produced, being ready for further morphological process (derivational or inflectional). This means that simplex nouns or verbs are directly derived from a word-class neutral base, not by conversion (from a noun to a derived verb or from a verb to a derived noun). As a consequence, the directionality problem of conversion can be avoided in this analysis because nouns and verbs directly come from a wordclass neutral base. Put differently, there is no need to determine which one is the base of the other. The word category of a noun or a verb is specified by Word-Class Marking at this point of morphological derivation. This class-marking process producing a word-class specified stem can be interpreted as a historical relic of the Germanic stemformative having been lost before OE (cf. section 2.1). The following step is the occurrence of a zero-morpheme, which is optional in morphological derivation. If an overt suffix occurs after the class-specified stem, the zero morpheme does not take place because there is no morphological motivation of the zero-derivation. The stem is already marked for word-class by Word-Class Marking and other morphological functions (person, case, number, tense) can be represented by the overt affix. By contrast, in cases where such an overt suffix does not occur, the zero-morpheme must take place in order to play a morphological

function

(e.g.,

{bāt}N+ø

(singular

nominative)).

If

the

zero-morpheme is not attached with a stem, the stem cannot surface, which means the form is not attested in OE (e.g., *{bāt}N). The prohibition of occurrence of affixless stems without having a zeromorpheme can be justified by the consideration of the typological stage of OE. As discussed in section 2, OE is at the transitional stage between stem-based morphology and word-based morphology.

This

ambivalent

typological status of OE can be well characterized by the postulation

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 37

of the zero-morpheme. Since OE is partly word-based, stems can occur as a word. However, the surface occurrence of OE stems must be restricted under a certain condition because OE morphology still remains stem-based. The OE morphological condition for the occurrence of affixless stems is to require the stems to acquire a zero-morpheme which can represent inflectional functions. In short, OE stems can appear as a word at the surface under one of the following two conditions: either with a zero-morpheme or with an overt affix. Therefore, it is assumed in this analysis that all uninflected forms (masculine/neuter

a-stem nominatives/accusatives (e.g.,

bāt 'boat'),

some

neuter

plural

nominatives/accusatives (e.g., word 'words')) are marked by a zeromorpheme at this stage of morphological derivation. The OE　 morphological requirement can be formalized as in (6). (6) OE Zero-Constraint OE stems can appear as a word at the surface only if either of the following conditions is satisfied: (a) an overt affix is attached to them (b) a zero-morpheme is attached to them

Finally, additional derivation and inflection take place in sequence, producing a final output. The application of these operations are also optional so that a number of affixless stems can appear (with a zeromorpheme) at the surface without being attached to any affixes. Now, we are ready for morphological analysis of each type of

ge-nouns. The following section attempts to provide a morphological analysis using the morphological algorithm presented in this section.

4.2 Morphological Analysis Let us begin with type 1 in which all possible derivatives are attested. The derivation of the derivative forms are exemplified below in (7).

38 언어와 언어학 제40집

(7) Type 1 (ge-noun/simplex noun/ge-verb/simplex verb) Nouns: gehyld 'protector', hyld 'protection', Verbs: gehyldan 'to keep', hyldan 'to recline' hyld

→ {hyld} → {hyld} ↘

N+{ø}

N

→ hyld

→ gehyld → [{hyld} ] {an} → hyldan ↘ {ge}+[{hyld}v] → [{ge}+[{hyld}v]]+{an} → gehyldan {ge}+[{hyld}N]

{hyld}V

V +

The base hyld becomes specified for word-category by Word-Class Marking in the beginning of the derivation so the nominal stem ({hyld}N) and the verbal stem ({hyld}V) are produced. Then, in order to appear as words at the surface, the stems must satisfy the OE Zero-Constraint formulated in (6), which requires them to be affixed either by an overt-affix (cf. (6a) or a zero-morpheme (cf. (6b)). The noun hyld becomes a word by being connected with a zero-morpheme and the other derivatives (gehyld, hyldan, gehyldan) by being attached with overt affixes (ge-, -an). In this way, the attestation and the derivation of all derivative forms can be explained without having directionality problem. Since the nominal stem and the verbal stem are separately derived by the base, not by their counterpart (noun from a verbal stem or verb from a nominal stem), there is no need to determine which one is the base of the other. In addition, the constraint in (6) shows why all the derivative forms are attested in this type. The stems of the words belonging to this type satisfy the derivational conditions, and thus can surface as words. Let us turn to type 2 where all derivatives forms are attested with the exception of simplex nouns. The sample derivation is provided below in (8).

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 39

(8) Type 2 (ge-noun/*simplex noun/ge-verb/simplex verb) Noun: gemet 'measure of strength' Verbs: gemetan 'to measure', metan 'to measure' met

→ {met} (no zero-derivation) *met ↘ {ge}+[{met} ] → gemet {met} → [{met} ] {an} → metan ↘ {ge}+[{met}v] → [{ge}+[{met}v]]+{an} → gemetan N

N

V

V +

The only difference between type 1 and type 2 is that the simplex nominal form of type 2 words is not attested while all derivative forms appear in the case of type 1. The absence of simplex nouns in type 2 can be interpreted as the result of the violation of the ZeroConstraint in (6). As a zero-morpheme does not occur to the nominal stem ({met}N), the stem cannot appear as a word. Instead, the stem is affixed with the prefix ge- and becomes a part of the noun

gemet. Type 3 includes ge -nouns whose corresponding ge -verbs are not attested. Their derivation are exemplified below in (9). (9) Type 3 (ge-noun/simplex noun/*ge-verb/simplex verb) Nouns: geflit 'strife, discussion', flit 'scandal, strife' Verb: flitan 'to compete, to strive' flit

→ {flit} → {flit} ↘

→ flit {ge}+[{flit} ] → geflit

N

N+{ø}

N

→ [{flit} ] {an} → flitan ↘

{flit}V

V +

(no prefixation) *geflitan

40 언어와 언어학 제40집

The absence of ge-verbal form *geflitan can be simply seen as the result of no application of prefixaton. Let us move on to Type 4 where verbal forms are not attested at all. (10) Type 4 (ge-noun/simplex noun/*ge-verb/*simplex verb) Nouns: gelād 'a course, a path', lād 'passage, voyage' lād

→ {lād} → {lād} ↘

N+{ø}

N

{ge}+[{lād}N]

→ lād

→ gelād

No Verbal Marking (*lādan, *gelādan)

As shown in (10), the absence of verbal derivatives is due to the

non-application

of

Word-Class

Marking

(verb

marking).

By

contrast, the base is marked as a noun so nominal derivatives are produced. The choice of whether Word-Class Marking applies or not to a specific base in OE may be determined based on OE native speakers' morphological knowledge. Synchronically, verbal derivatives related to ge-nouns in this type do not exist in the OE lexicon while nominal derivatives do. This morphological information can be learned by OE speakers and their knowledge has an effect on morphological derivation. If speakers know there are derivatives from a certain word-class, Word-Class Marking applies to make the word-class stem.

Ge-nouns of the type-5 category have only affixless verbal forms. Their corresponding simplex nouns and ge-verbs are not attested in OE. Relevant sample derivation is given below in (11). (11) Type 5 (ge-noun/*simplex noun/*ge-verb/simplex verb) Noun: gedræg 'dragging, band' Verb: dragan 'to drag, to draw'

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 41

drag

→ {drag} (no zero-derivation) *drag ↘ {{dræg} } → {ge}+[{{dræg} }] gedræg {drag} → [{drag} ] {an} → dragan N

N

↘

V

N

V +

(No prefixation) *gedragan The simplex nominal form *drag and the ge-verbal form *gedragan cannot occur as words due to the absence of zero-derivation and prefixation, respectively. One thing I should mention about this case is the derivation of umlaut forms: {drag}

→

{{dæg}}. Historically speaking,

the phonological modification of roots was produced by i-umlaut which occurred in the pre-OE period. The OE residue of the phonological vocalic change is the variation of root vowels as shown in (11). From a synchronic point of view, this issue is involved in the interaction between morphological operations and phonological change. As this topic is not directly related with the purpose of this paper, I simply assume that a modified root form is represented by "{{ }}" (e.g., {drag}

→ {{dæg}}) without providing further relevant discussion.

Last, type 6 only consists of ge-nouns. Simplex nouns and verbal forms are not found in OE. The derivation of the type-6 ge-nouns is exemplified below in (12). (12) Type 6 (ge-noun/*simplex noun/*ge-verb/*simplex verb) Noun: gehygd 'thought, meditation' hygd

→ {hygd}

N

(no zero-derivation) *hygd

↘

{ge}+[{hygd}N}] gehygd No Verbal Marking (*gehygdan, *hygdan)

42 언어와 언어학 제40집

As shown in (12), the simplex nominal form cannot occur because its nominal stem is not attached with a zero-morpheme. The verbal forms also do not appear at the surface because the base is not marked as a verbal stem.

5. Conclusion In sum, this paper has provided a morphological analysis of 6 types of ge-nouns, addressing the controversial issues: the directionality of affixless

derivation

and

the

postulation

of

a

zero-morpheme.

The

directionality problem could be avoided in my analysis because unlike traditional morphological accounts assuming that a noun is derived by a verb or vice versa, a base is marked for word-category by WordClass

Marking.

Namely,

nominal

or

verbal

stems

were

produced

directly from the base by means of the application of Word-Class Marking, not from the other stem by conversion. Furthermore, the Zero-Constraint in (6) successfully accounted for why some derivative forms were not attested in OE even though they were morphologically possible forms. It is very significant to attempt to explain the absence of some derivatives at a specific period of a language because there have been no formal treatments of this issue at all. Derivative forms have been simply considered as lexical gaps or as the result of historical loss. By contrast, this analysis clearly has showed how some derivatives appear at the surface while others do not. The presence or the absence of a derivative form were determined by the Zero-Constraint.

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 43

References Anderson, Fran. (1998) “A Core Morphology for Old English Verbs.”

English Language and Linguistics 2(2): 199-222. Bauer,

Laurie

(1983)

English Word-formation.

Cambridge:

Cambridge

University Press. Colman,

Fran.

(1985)

“On Some

Morphological

Formatives

in Old

English.” Folia Linguistica Historica 6(2): 267-283. Dalton-Puffer, Christiane (1992) “A View on Middle English Derivation: Verbs.” VIEWS 1(3): 1-15. Dalton-Puffer, Christiane

(1993)

“How

Distinct are

Inflection and

Derivation? Reply to Lass and Ritt.” VIEWS 2: 40-44. Fisiak, Jacek. (1980) Historical Morphology (ed.) (Trends in Linguistics.

Studies and Monographs 17). The Hague: Mouton. Hogg, Richard (1992) “Phonology and Morphology”. In R. Hogg (ed.) The

Cambridge History of the English Language. vol.1. The Beginnings to 1066, 67-167. Cambridge: Cambridge University Press. Jespersen, Otto (1942) A Modern English Grammar: On Historical

Principles. Part 6 Morphology. London: Allen and Unwin. Kiparsky, Paul (1982) “Lexical Morphology and Phonology”. In I.S. Yang (ed.) Linguistics in the Morning Calm, 3-91. Seoul: Hanshin. Kastovsky, Dieter (1968) Old English Deverbal Substantives Derived

by Means of a Zero Morpheme. Esslingen: Langer. Kastovsky, Dieter (1985) “Deverbal Nouns in Old and Modern English: From Stem-Formation to Word-Formation.” In J. Fisiak (ed.),

Historical

Semantics-Historical

Word-Formation,

221-261.

Berlin:

Mouton de Gruyter. Kastovsky, Dieter (1992) “Semantics and Vocabulary.” In R. Hogg (ed.)

The Cambridge History of the English Language. vol.1. The Beginnings to 1066, 290-407. Cambridge: Cambridge University Press.

44 언어와 언어학 제40집

Kastovsky, Dieter (1996) “Verbal Derivation in English: A Historical Survey.” Or: Much ado about nothing. In D. Britton (ed.), English

Historial Linguistics 1994, 93-117. Amsterdam: Benjamins. Kastovsky, Dieter (2002) “The Derivation of Ornative, Locative, Ablative, Privative and Reversative Verbs in English.” In T. Fanego et al. (ed.), English Historical Syntax and Morphology: Selected

Papers from 11 ICHEL, 99-109. Amsterdam: Benjamins. Kastovsky,

Dieter

(2006)

“Typological

Changes

in

Derivational

Morphology.” In A. van Kemenade et al. The Handbook of the

History of English, 151-176. Malden: Blackwell Publishing. Kuryłowicz, Jerzy (1968) Indogermanische Grammatik. Band 2: Akzent, ablaut. Heidelberg: Winter. Marchand, hans (1969) The Categories and Types of Present-Day

English Word-Formation. München: Beck. Mitchell, Bruce & Fred Robinson (1998) Beowulf: An Edition with

Relevant Shorter Texts. Oxford: Blackwell. Mohanan, Karuvannur P. (1986) The Theory of Lexical Phonology. Dordrecht: Reidel. Lass, Roger (1993) “Old English -ian: Inflectional or Derivational?”

VIEWS 2: 26-34. Plag, Ingo (1999) Morphological

Productivity: Structural Constraints

in English Derivation. Berlin: Mouton de Gruyter. Plag, Ingo (2003) Word-formation in English. Cambridge: Cambridge University Press. Ritt, Nikolaus (1993) “What Exactly is it that Makes OE -ian Derivational?” VIEW 2: 35-39. Sanders, Gerald (1988) “Zero Derivation and the Overt Analogon Criterion.” In M. Hammond et al. Theoretical Morphology, 155175. San Diego: Academic Press. Szemerényi,

Oswald (1996)

Introduction to Indo-European Linguistics.

[Translated from Einführung in die vergleichende Sprachwissenschaft,

Nominal and Verbal Derivation in Old English▪Yoo-kang Kim 45

1990]. Oxford: Clarendon Press. Toller, Northcote (1973) An Anglo-Saxon Dictionary: Based on the

Manuscript Collections of Joseph Bosworth. London: Oxford University Press.

[449-791] 경기도 용인시 모현면 왕산리 산 89 한국외국어대학교 영어학부 E-mail: [email protected]

논문접수 : 2007년 7월 15일 게재확정 : 2007년 9월 3일