Abstract: The article addresses the basic valency orientation of Gothic. The data is

analyzed mainly qualitatively, but, differently from previous similar studies, an insight is given into a quantitative model further supporting the findings of the first

analysis. Whereas the only transitivizing strategy found is causativization through

Go. -ja- (< PIE *-éi̯e/o-), three detransitivizing strategies are employed: voice alternation, reflexivization, and na-derivation. The last-mentioned strategy pairs with ja-derivation to make up one of the two non-oriented means of encoding the anticausative alternation, the other being suppletion. A summary of the key findings is as follows. The passive voice appears to play a role, albeit marginal, as the noncausal element of the anticausative alternation. Out of the other two detransitivizing strategies, na-verbs are productive in East and North Germanic. Based on the Gothic evidence, na-verbs with specialized nonagentive meaning are formed primarily to roots whose semantics preferably selects for inanimate S/O’s. Following the extension to verbal roots of opposite alignment, the formation becomes a competitor to the reflexive construction. As a result of rise in productivity, na-verbs are selected to encode the noncausal end of the anticausative alternation, eventually ousting the reflexive construction as a competitor to the middle.

Abstract: This paper investigates Gothic valency patterns and alternations applying

the methodology of the ValPaL project. Our data consists of 87 verb meanings based

on those in the ValPaL corpus, which we supplemented due to gaps in coverage.

Valency alternations are divided into uncoded and coded patterns. The uncoded

group includes partitive, null object, external possessor, cognate/kindred object,

applicative, (marginally) causal : non causal, and (marginally) reflexive alternations.

The coded group comprises the applicative, passive, causal : noncausal,

reflexive, and reciprocal alternations. The nature of the Gothic corpus (extension,

genre, and nature of the text) influenced the results: some constructions were less

frequently attested than expected; genre influenced the frequency of some verbal

usages; some marginally attested constructions are owed to the Greek translational

source. Most remarkably, we showed that: (i) in respect to the debated issue

regarding the interpretation of Gothic passives, these can be employed to express

noncausal situations; (ii) within the causal : noncausal alternation, whereas the

causal domain is tied to ja-suffixation, the noncausal domain can be expressed by a

variety of means, including na-verbs. Our results confirmed Ottósson’s (2013) view

that the noncausal domain was systematized at a later stage than the causal one.