The article deals with segmentation and stress in early Germanic. An attempt is made to isolate minimal prosodic units of Germanic, not traditional syllables. These units are then analyzed in terms of morae and phonemes. Alternations by the Akzentumsprung and Verner's law are interpreted as indicating that early Germanic had sentence rather than word stress.

The formation of the Pennsylvania German dependent infinitive construction varies across Pennsylvania German speech communities. The use of zu to mark infinitives which do not complement modals is receding in all communities and has been replaced variably by far or far…zu. Younger nonnative speakers have grasped the (far)…zu construction as template for producing discourse. Sectarians (Amish and Mennonites) mark dependent infinitives less frequently than nonsectarians, and Amish speakers often avoid the construction by the progressive aspect. The influence of English on these developments is indirect. The loss of zu, the use of (far)…zu, and the sectarian omission of far and zu represent generalizations of different rules. The contact seems to have promoted generalization, but has influenced each group to generalize differently

Traditionally, noun modifiers in German fall into two classes, the so-called der- and ein-words, and descriptive or attributive adjectives. In the noun phrase, members of these word classes are inflected by the addition of one or another of two sets of endings, i.e., the primary, strong, or pronominal endings, or the secondary, weak, or nominal endings, in highly predictable combinations. In the data collected by Ljungerud (1955), however, sequences of endings in noun phrases containing nine modifiers occur with noticeable departures from the norm, i.e., folgend, sämtlich, ander-, einig-, viel-, manch-, welch-, irgendwelch-, and solch-. It is concluded that the anomalous sequences of endings in phrases containing these modifers are motivated by the growing association of certain sequences of endings with the feature +/-Generality as a component of the meaning of the noun phrase as a whole.

This paper represents a return to the theme of Germanic umlaut, umlaut in general, and sound change in general. It invokes first the concept of the parameter, then the concept of parametric inertia. That is, since all parameters are produced by physical organs, they are governed by inertia, and this allows a (phonetic) explanation of assimilation (including umlaut) and dissimilation. Further, the concept of a hierarchy of parameters based on a phonetically most similar notion is developed to replace the unworkable phonetically similar criterion in classical phonemics.

Anatoly Liberman's recent article in this journal (“The phonetic organization of Early Germanic” 2,1:1990) contains a number of potentially important claims about the nature of early Germanic accentuation. (1) His major conclusion appears to be twofold: first, the syllable struture of early Germanic disyllabics must have normally been CVC.V (using “.” to represent a syllable boundary), because CV (and by extension therefore CV.CV) was not a possible structure of a Gothic word. Second, early Germanic possessed no lexical stress (“word stress”), but instead only phrasal stress. (2) These conclusions are at least provocative to those familiar with the massive literature on Germanic accentology, and this was no doubt their intent. In this brief piece, I would like to note several implications and possible extensions of the points Liberman raises and the conclusions he draws, but also to disagree on some points. My perspective here differs from Liberman's primarily in focusing not on early Germanic in relation to its attested daughter languages, but rather on Germanic vis-à-vis Indo-European and some data from the languages of the world, that is, typological data.

Notker's short five-section treatise De musica was written exclusively in Old High German, except for the retention of technical terms in Latin and pseudo-Greek, with De mensura fistularum organicarum being the last. In De musica Notker abandoned his practice of giving the Latin text before each part of the German translation, for the discourse was no doubt intended as a manual for organ builders, of whom Latin could not be required. In this study relevant passages from the Latin texts alternate with Notker's translation. Then the four manuscripts preserving the fifth section in full or in part are compared in their fidelity to Notker's accentuation and allophones of initial consonants. Finally, the twelve terms retained in Latin and pseudo-Greek are enumerated.

On the basis of certain compositional features introduced into the Tristan story by Thomas of Brittany, it has been established that the many Tristan reminiscences in Old Norse literature stem from Tristrams saga (Schach 1965). In this paper it will be argued that Tristrams saga provided the structural model for Egils saga.

All scholars think that they express themselves clearly; they also expect that their arguments will carry immediate conviction. Polemical notes show how mistaken they are. In this rejoinder I will attempt to clarify my position and add a few new arguments.