Dieter Wunderlich

Minimalist morphology: the role of paradigms

In this paper,1 I will give a short summary of the basic assumptions and principles of Minimalist Morphology (MM) proposed in Wunderlich (1992a) and Wunderlich and Fabri (1993). I will then consider the construction of paradigms more precisely and finally turn to two questions raised in the recent literature: how inflectional classes are made up, and whether rules of referral are needed.

Argument Extension by Lexical Adjunction

This paper proposes a unified account of the semantic and syntactic properties of certain particle and prefix verbs in German as well as of constructions involving resultative and depictive predicates. I will argue that the secondary predicates, considered to be adjuncts, are lexically integrated into the verb by an operation called Argument Extension (ARG), a general option of the modification rule which states that the head accommodates to the adjunct, in contrast to the traditional view, according to which the adjunct accommodates to the head. Section 1 motivates the theoretical possibility of the operation ARG within Categorial Grammar, while section 2 briefly introduces the framework of Lexical Decomposition Grammar, in which the following analyses are situated. Section 3 shows that the operation ARG allows a satisfying analysis of prefixes and particles that introduce new arguments to the verb. This analysis will then be extended to resultative and depictive secondary predicates in sections 4 and 5, and will be compared to some of the syntactically based alternatives that have been proposed in the literature

On the interaction of structural and semantic case

Abstract We consider two variants of morphological case: structural case (such as accusative, ergative, or dative), which is encoded by abstract case features reflecting the semantic ranking of arguments, and semantic case (such as instrumental or directional), which encodes an additional semantic relation to be licensed by the meaning of the verb. Individual verbs may be lexically marked for either structural or semantic case. We show how a correspondence-theoretic approach can successfully describe the various patterns of structural case found in German. We then discuss instances of double object where structural case and semantic case compete with each other: under certain circumstances semantic case is favored, while under other circumstances structural case is favored. We mainly consider two scenarios, described by the respective rankings of S tructural C ase (which requires all objects to be realized by a structural case) and U niqueness (which restricts a structural case to occur at most once in a clause). In particular, we analyze the different options for structural vs. semantic case in the dialects of Quechua.

Argument linking in Basque

Abstract In this paper we aim at a comprehensive but minimal description of verb classes in Basque in terms of their argument structure. We assume the principles of Lexical Decomposition Grammar, according to which lexically decomposed semantic representations determine the argument hierarchy and a common feature system encodes both the argument hierarchy and the set of structural linkers. Basque has three structural agreement markers in the finite verb, together with an elaborate case system. We will investigate the verb classes of Basque with respect to their behaviour under causativization and affective alternation; both operations are used here as a key to argument structure.

Methodological Remarks on Speech Act Theory

I regard the notion of ‘speech act’ as one of the most fruitful notions of contemporary linguistic theorizing. It orients our scientfic endeavours towards the function of language in human communication. In doing so, it allows for a combination of different methods and fields of linguistic, as well as of philosophical, investigation, such as, e.g., the theory of grammar, the theory of meaning and the theory of discourse.

Transitivity alternations in Yucatec, and the correlation between aspect and argument roles

Abstract This paper analyzes the argument structure of Yucatec Mayan. In line with traditional assumptions but in contrast to more recent proposals, we show that Yucatec has ergative morphology but not a split-ergative system. For intransitive verbs there is a strict correlation between aspect and the type of linker: they take ergative clitics in the imperfect, but nominative suffixes in the perfect. Moreover, each intransitive verb is either inherently perfective or inherently imperfective, and only the opposite aspect is marked. We account for these data by two basic assumptions: (i) verbal (sub)predicates may characterize the beginning or the end of the situation, and (ii) imperfect selects an initial phase but perfect the end. In addition, Yucatec has an overt system of transitivity alternations: intransitive verbs may add a subject or an object, and transitive verbs may demote the subject or the object. We show that all these alternations fully conform to our basic assumptions. Finally, we show how three-place verbal concepts are expressed in a language with only two structural arguments by the interaction of detransitivization and transitivization.

German noun plural reconsidered

In this paper I discuss German noun classes as organized in an inheritance tree which simultaneously satisfies requirements of DATR accounts and of Minimalist Morphology. It is argued that -s and -n are the only plural suffixes, that the plural ending -r is always lexically fixed (which is also the case for other plurals and for the umlauting feature), but that schwa-plurals and unmarked plurals follow from a prosodic constraint, in line with constraint-based proposals. Moreover, the German noun plural system exhibits at least two properties that are problematic for a simple ruleor fact-based analysis: (i) umlaut plural takes precedence over n-plural in the feminines, and (ii) many plural forms of nonfeminines vacillate between schwaand s-ending. The latter fact indicates that German is still in the process of adopting s-plurals as the default, a situation that can best be described by means of changes in the constraint ranking. Therefore, I sketch an OT analysis that incorporates the morphological facts in the form of constraints with a partially changing order, and which accounts for the two above-mentioned properties. This analysis shows that the plural of feminines is rather robust, even if the plural of nonfeminines is subject to drastic changes.

Advances in the theory of the lexicon

The book investigates the interface structure of the lexicon from various perspectives, including typology and processing. It surveys work on verb classes, verb-noun similarities, semantic representations, concepts and constructions of polysynthetic languages, research on the processing of inflectional and derivational elements, and new work on inheritance-based network models. The book will be of interest to researchers and advanced students in all fields of linguistics and in the cognitive sciences.

How Gaps and substitutions can become optimal: the pronominal affix paradigms of Yimas

In order to describe the possible and impossible pronominal prefix combinations in the Yimas verb, a set of partially ordered constraints is proposed within the correspondence format of Optimality Theory (OT). These constraints simultaneously explain the split in the inventory, the order of prefixes, why gaps appear in some category combinations, and why less specified prefixes substitute for fully specified ones in other combinations. Moreover, I will show that the Yimas inventory of ergative and accusative morphemes is rather stabile under the given constraint ranking. A scenario in which Yimas shifts to a pure accusative system is more imaginable than a scenario in which it shifts to a pure ergative system, despite the fact that the ergative morphemes outnumber the accusative ones. Thus, an OT account may result in evaluations of the inventory that differ from more intuitive judgments.

Person marking in Quechua. A constraint-based minimalist analysis

Abstract Quechua, which is spoken from the southern parts of Colombia to the northern parts of Chile and Argentina, exhibits a rich verbal morphology. The various dialects of Quechua can be ranked in terms of developmental stages, according to the way in which the inflectional system is organized. Some Peruvian dialects (such as those of Ancash and Ayacucho, belonging to different dialect groups) exhibit remarkable asymmetries in the marking of objects : 1st person objects are marked transparently with separate object and subject affixes, whereas the marking of 2nd person objects is somewhat opaque. We account for this asymmetry by the assumption that the person hierarchy in Quechua is 1 ⪢ 2 ⪢ 3 and that the marking of objects is constrained by the Object-Subject-Constraint (OSC). ‘The object may be marked separately from the subject only if it refers to a person that is higher on the hierarchy of person than the person to which the subject refers’. We will then show that the dialect of Cuzco (Peru) has lost any independent morpheme for 2nd person objects, hence, OSC has become irrelevant. It is this stage that allows reanalysis: the most recent dialects such as those of Potosi (Bolivia) and Santiago del Estero (Argentina) have reinvented a 2nd person object morpheme and have given up OSC; in these dialects a symmetric system of object marking has evolved. We will discuss our findings in terms of two possible types of inflectional systems: one type has a small set of morphemes but additional constraints that restrict the actual combinations of morphemes — and so induce asymmetries, whereas the other type has a greater and more finely differentiated set of morphemes but does not need additional constraints — and so allows more transparency and symmetry in the system. We will also show how the development of the Quechua person system is intermingled with that of the number system. Only if the plural suffix follows the person affixes, as happens in the Quechua II dialects, can plural allomorphs arise that are sensitive to person or that fuse plural and person into one single morpheme, facts that have considerable consequences for the whole system.