In consequence of the familiar High German consonant shift, the voiceless stops /p t k/ became the affricates /pf ts kx/, simplifying later to the geminate fricatives /ff ss (‘ƷƷ’) xx/ in certain environments. It has long been assumed that the affricates developed only from aspirated allophones of /p t k/, thus accounting for the retention of plain stops in /s/-clusters and a few other positions, but it is not obvious why particularly aspiration, ensuing from a laryngeal gesture of spread glottis, should have resulted in the oral property of affrication. Neither is it clear how etymologically simplex segments, upon deaffrication in certain environments, might have resulted in geminates without further stipulation. Proceeding from essentially traditional assumptions, and at the same time addressing important questions about Germanic phonetics raised by Vennemann (1984), the present paper attributes the triggering of the affrication event to an early factoring out, or segmentalization, of the feature for aspiration, i.e., pre-OHG [ph th kh] → [ph th kh]. Explanations are then proposed for the development of affricates out of these now disegmental sequences via place assimilation from the stops, and of the affricates into geminate fricatives (postvocalically and in some postconsonantal environments) via assimilatory weakening.

The operation and suspension of resolution in Old English meter, generalized as Kaluza's Law, is shown to be based on mora counting; -VC(C), -VV(C), and -VCV count as bimoric, while -VCVC(C) and -VCVV(C) are trimoric. The maximum moric value of a syllable is two. The proposed mora-based analysis leads to an improved account of a number of phonological processes in pre- and Early Old English.

This paper introduces a cognitive framework for perfect auxiliary selection (HAVE versus BE) in Germanic based on transitive (HAVE) and mutative (BE) prototypes as affected by lexical aspect and transitivity parameters (Hopper and Thompson 1980). The phenomenon of “HAVE-switch” is exemplified in the history of several Germanic languages. Here numerous modal and aspectual factors shift the perfect auxiliary with mutatives from the customary BE to HAVE. This shift is then explained in terms of the proposed model. The contexts in question are all seen to reduce the mutativity of a clause (the effective attainment of the resultant state in the patient subject) and hence the motivation for using BE. Several direct parallels between HAVE-switch in Germanic and aspectual usage in Russian are discussed and their motivation in terms of this approach shown.

The Swedish dialectal words nårot ‘worn-out’ and narg ‘a piece of wornout cloth’ have not been etymologically treated so far. I here connect them with a family of Balto-Finnic words, likewise still lacking an etymology: Finn. naarmu ˜ narvas (<*narwa−), Est. näru (<*narwo-) and narvas ˜ narmas, etc. In these Balto-Finnic words, the -v- has in most attested forms undergone a sound substitution to -m-, known to have taken place in some Germanic loan words in Balto-Finnic. I propose that the above dialectal Swedish words can be explained as old wō/wa–stems, related to West Germanic words like Ger. Narbe, OE nearwian ‘to constrain’, etc. Thus, it appears that Finn, narvas-, naarmu, Est. näru, etc. are Proto-Germanic loan words in Balto-Finnic, and that the words later died out in the Germanic languages, only to be retrievable through dialectal Norwegian and North Swedish data and reconstructable through comparison with the Balto-Finnic loan words.

OE ge- functions synchronically as a proclitic, contrary to widely held views. This morphological status is an integral part of a broader diachronic pattern in Germanic, namely grammaticalization of an old particle to a prefix. OE g e- provides a clear example of grammaticalization. Following its path from Proto-Indo-European adverbial to Old English proclitic, and from its near grammaticalization as a past participle marker and subsequent loss, one can see easily the cline of grammaticalization.

In this paper it is proposed that Germanic locative adverbs in *-on- bear an etymological relationship to Hittite ablatives in -anza and therefore yield additional evidence for the archaic nature of the Germanic dialect group.

In this article the development of German können, dürfen, sollen, mögen, müssen, and wollen is examined in the light of grammaticalization theory. It is assumed that these verbs form a class (or subclass) of modal auxiliaries. Some of the assertions made on the development of modal verbs across languages are examined. For example, do the modal meanings develop from nonmodal meanings? Do the epistemic uses of these auxiliaries evolve from deontic ones? Can such an evolution be described in terms of metaphor and can the development of these verbs be described as following a distinct “path” that is “unidirectional”? Some of these questions are answered in the affirmative, but new question arise in the course of the investigation.

After Martin Luther first translated and published the New Testament in 1522, he immediately began the work of revision—work that would last through his lifetime and beyond. Working with a group of biblical scholars, he made thousands of changes to the text, continuing until his death in 1546. Although some critics have seen Luther's earlier language as vulgar and coarse—particularly in the Gospels— and have suggested that he refined his language over time, others suggest that a more differentiated view is necessary. This article examines the lexical differences in the Gospel of Matthew between the Septembertestament of 1522 and the last Bible published during Luther's lifetime, in 1545. Major lexical changes are compared with the Greek source text, and assigned to three major classes: (I) changes that bring the translation closer to the original Greek meaning; (II) changes that diverge from a close rendering of the source text, for comprehension or esthetic reasons; and (III) changes that are neutral with regard to the source, originating from target language (German) considerations. Most major changes arise from either the source text or understandability considerations. The original lexical choices in the 1522 version are not as coarse or extreme as some have suggested.

A significant aspect of the changes in Martin Luther's style over the years can be characterized linguistically in terms of clause linkage. In order to analyze such diachronies incisively, we need a framework which can deal with the shifting and overlapping iconicities of morphology, syntax, lexicon, and discourse. Lehmann's (1988) clause linkage typology construct provides such a systematic framework. With it we are able to characterize rigorously the varying clause linkage strategies in two versions of Luther's Romans. Categorization of some 60 sentence pairs from these corpora yields incontrovertible evidence favoring change toward semanto-syntactic compression, and it suggests a typological shift toward more compressed structures. In turn, Lehmann's typological framework is further strengthened by its application to the ENHG Luther data.