• Keine Ergebnisse gefunden

2. Previous morphological analyses

2.1. Withgott & Halvorsen (1984, 1988)

2.1.4. Classes of dominance

Withgott & Halvorsen’s (1984, 1988) analysis, as we have presented thus far, however, is not yet complete. We have only considered how it can account for accent assignment in mono- and bimorphemic words not multi-affixed words like be-stemm-else >bestemmelse1 ‘decision’.

Thus, we move onto Withgott & Halvorsen's analysis of words that are made up of more than one tone-bearing morpheme, e.g. 'omarbeidelse1 ‘reworking’, 'hederligst1 ‘most honest’.

Following Withgott & Halvorsen’s analysis, these words consist of three and two tone-bearing constituents respectively, i.e., {Lom}VERB {Harbeid} L{else} and {heder} {Hlig}{Lst}. In such polymorphemic words, which have more than one floating tone, the question is how to identify the tone that will ultimately win over the main stressed syllable and determine the tone of the whole derivation. Below in (14) we review Withgott & Halvorsen's analysis for 'hederligst1

‘most honest’.

(14) A multi-affixed word hederligst ‘most honest’ and accent assignment Stem Affix + stem Association

of floating H

Accent

assignment Gloss a. heder1 heder + Hlig > H'hederlig 'hederlig2 honest b. Hhederlig Hhederlig + Lst > H L'hederligst 'hederligst1 most honest

The H-inducing affix {-Hlig} succeeds in determining the accent in hederlig2 ‘honest’ (14a), but not in 'hederligst1 ‘most honest’ (14b). The floating L of the superlative suffix {-L(e)st}

apparently wins out over the floating H of the derivational suffix {-Hlig}. Thus, we see here that Withgott & Halvorsen (1984, 1988) need another set of classifications for morphemes to be able to predict which tone will triumph in such multi-derivations. They, consequently, posit that tone-bearing morphemes fall into two categories of dominance, namely weak or strong.

(15) Weakly dominant affixes

Weakly dominant affixes dominate over the tone of the stem but not over other affixes:

Affix Affix + stem Association of tone

Accent

assignment Gloss a. -Hlig latter1 + Hlig > H'latterlig > 'latterlig2 ridiculous b. -Helse kall + Helse > H'kallelse > 'kallelse2 callingNOUN

c. -Lsel kjør + Lsel > L'kjørsel > kjørsel1 drivingNOUN

Weakly dominant affixes are only able to override the tone of the stem as in for example (15a) latterlig2 where {-Hlig} induces Accent 2, i.e., it overrides the default L of latter1, resulting in latterlig2 ‘ridiculous’. The same is true of {-Helse} and {-Lsel} in (15b,c). We see that the suffix {-Lsel}, which Withgott & Halvorsen (1984, 1988) claim bears a floating L, induces Accent 1 in kjørsel1 ‘drivingNOUN’, and {-Helse} induces Accent 2 in kallelse2

‘callingNOUN’. However, these examples only show how H-inducing or L-inducing affixes can change a monosyllabic Accent 1 into Accent 2 or not change the accent at all respectively.

The L-inducing affix {-sel} that Withgott & Halvorsen list as weakly dominant attaches exclusively to monosyllabic stems as we saw in (10), which in effect remain Accent 1 after suffixation as well. The effect of all weakly dominant suffixes can at this point be summarised as only ever dominating over the accent of the stem if it has Accent 1. These affixes are not strong enough to determine the accent when another, stronger affix is added, as we see below in (16a).

(16) Strongly dominant affixes

Strongly dominant affixes dominate the tone of everything:

Affix Affix +stem Association of tone

Accent

assignment Gloss a. L/Hpå- Hpå + kjør + Lsel > H'påkjørsel > 'påkjørsel2 crash b. L/Hom- Lom + Harbeid + Helse > Lomarbeidelse > 'omarbeidelse1 reworking Although the semantics tell us that {- Lsel} is suffixed to the verb 'påkjøre1 ‘to run into’ as we saw in (13), for Withgott & Halvorsen's analysis it must be as follows. In (16a), the strongly dominant prefix {Hpå-} is added to the derived noun kjørsel1, and is therefore in its H-inducing capacity. This H overrides the L of the weakly dominant prefix {- Lsel}, resulting in Accent-2 'påkjørsel2 ‘crash’. According to Withgott & Halvorsen’s (1984, 1988) examples, strongly dominant affixes seem to be exclusively prefixes – stressed prefixes to be exact.

Perhaps the dominance of these affixes lies in the fact that they bear main stress as opposed to the weakly dominant affixes which all appear to be unstressed suffixes. These prefixes, like {Lom-}VERB, always determine the tone of a multi-affixed form, like'omarbeidelse1 ‘reworking’

(cf. (16b)). In fact, {L/Hom-} and {L/Hom-} always dominate regardless of whether they are L-inducing or H-inducing. However, there are strongly dominant affixes lacking main stress that Withgott & Halvorsen (1984, 1988) do not discuss, which induce Accent 1, e.g. {be-}, {for-}, as we saw in (1a). We give a few examples below in (17).

(17) Unstressed Accent-1 inducing prefixes:

Prefix Infinitive Gloss Prefix + infinitive Accent

assignment Gloss

Lbe- tenk-He to think Lbe + Htenke be'tenke1 to consider

Lfor- bedr-He to better Lfor + Hbedre for'bedre1 to improve Although Withgott & Halvorsen (1984, 1988) do not discuss these prefixes, following their reasoning, these prefixes should be included in their category of strongly dominant L-inducing affixes because they turn Accent-2 infinitives into Accent-1 infinitives. The floating L tones win out over the H-inducing infinitive suffixes. Thus, the strength of strongly dominant affixes cannot lie in their capacity of bearing main stress.

Withgott & Halvorsen's approach, however, still cannot sufficiently account for all possible scenarios of accent assignment. For example, it cannot account for the difference in accent of monosyllables vs. polysyllables in the plural, as we will see in (21). Withgott & Halvorsen

must integrate the rules of tonal behaviour of stems and affixes with the correct ordering of morphological and phonological processes for to correctly explain all accent distribution. As we will see, they attribute the differences in accent of nouns in the plural to the different levels of lexical phonology in East Norwegian and what processes belong to each level. It is peculiar that their classes of dominance do not follow from level ordering directly.