Volume 21, Issue 2017 - 21st Anniversary Issue 2017

Basic Valence Orientation has been suggested as a typological parameter by Nichols, Peterson, & Barnes (2004). Generalising over the entire lexicon, the idea is that languages can be distinguished as transitivising or detransitivising, depending on whether their verbs are basically intransitive or transitive and the opposite valency values require some means of derivation, such as causativisation or decausativisation. Whereas derivedness among valency opposites is assessed through easy-to-spot overt segmental morphological or syntactic markers by Nichols et . (2004), we argue that phonological alternations, on their own or attendant upon conjugation class switches between intransitives and transitives, can be as directed as derivations are which are implemented through adding segmental markers. Illustrating from German, we show that stem vowel patterns in strong and corresponding weak verbs (with the former expressing inflectional categories through ablaut) as well as umlaut alternations in verb derivation are systematically involved in valency oppositions and are both directed. Thus, German emerges as being typologically mixed, being strongly transitivising on the grounds of such asymmetric formal patterns, while also showing (as observed by Nichols et . 2004) a detransitivising or indeterminate disposition through syntactic “middle” marking or verb lability. This typological result is also of diachronic significance, insofar as the older transitivising inclination is seen to have been remarkably pertinacious, long surviving the loss of the affixal valencyincreasing morphology of Common Germanic and able to hold its own against more recent detransitivising competition.

Grammaticalised reciprocal markers in Germanic derive from combinations of a quantifier and the alterity word ‘other’, elaborating on a minimalist strategy of identical NP repetition suggesting rather than expressing reciprocity (‘earl[s] hated earl[s]’). Subserved by quantifier floating, they develop fromfree to tighter syntactic combinations and eventually intomorphological units, tending towards complete inflectional deactivation. Sooner or later in all Germanic languages, the quantifier part of the reciprocal gets inside prepositional phrases (‘earls fought each/one with other’> ‘earls fought with each/one other’). German continues this fusional theme by combining the reciprocal with prepositions in compounds; and in Bavarian it eventually gets reduced further to a bound stem limited to (partly lexicalised) combinations with a preposition, thus being barred from the direct object relation, unlike the reflexive. In tracing this overall diachronic scenario, the question is raised of the pronominality (or pro-NP-hood) of reciprocals in Germanic. It is argued that, regardless of their nominal and referential source, reciprocals here strongly incline towards becoming adverbs of attenuated, situational rather than personal reference, highlighting the relational (role reversal) rather than the (co-)referential component of reciprocity, as is common also elsewhere.

Being agglutinative or flexive are not properties of entire languages, nor are they simple properties. There is a whole ränge of simple properties, all logically independent ofeach other, prominently including those ofseparation/ cumulation and invariance/variance. They are all properties ofindividual wordforms, and again there is no logical necessityfor these to agree in their property sets. This creates a huge potential for heterogeneity within andfor diversity across languages, which, if realized to the fall, would render morphological typology unviable. However, an examination ofsplits between Separation and cumulation and between invariance and variance along the lines ofword-classes, ofsubsets within single word-classes, of morphological categories, and ofterms ofcategories suggests that mixtures between agglutination and flexion, though multifarious, are not random. Ifgrammars arefound to be less heterogeneous, and languages less diverse, than they could be, this can be due to universal, timeless principles or to regularities ofchange. Both play a wie in shaping morphological Systems.

Throughout its history, the hope has always been cherished that typology is holistic, and holism entails that there is systematic co-variation not only within levels or modules of grammar but also between them. Accordingly, numerous claims have been made that phonology does not vary across languages independently of morphology and syntax, and vice versa. The variables that are allegedly interrelated pertain to segment inventories, the shapes of syllables, morphemes, and words, phonological or morphonological rules, tones and accents, and rhythmic or prosodic patterns on the one hand and to analytic or (poly-)synthetic grammar, Separatist or cumulative morphological exponence, the complexity of grammatical units, and their linear order on the other. These claims are cataloguedin thispaper. To substantiate them and to accommodate those that are found valid in theories of the Interface between phonology, morphology, and syntax remain äs tasks for the future.