Human beings are able to modify the vocabulary and even the grammar of their native languages (in poetry this is called "poetic licence"). They can contaminate words (as in the Eng. word brunch , from breakfast and lunch ) or otherwise modify them (by adding or eliminating elements or by twisting them around). Most often this is done in order to obtain a comical effect. In English, the term pun is used for words or expressions used this way. In German there exists the term Wortspiel ("play with words"), which suggests that we always have to do with modified words. Since we can also modify phrases and sentences in order to achieve a comical effect, I use the term Sprachspiel ("play with language") as a general term, distinguishing between Wortspiel (as a pun in its narrow sense) and phrasales Sprachspiel (as a pun concerning phrases and sentences). In my article I undertake an attempt to classify the different techniques used in creating puns in German and their structures, being aware of the fact that often different techniques can be used to create the same kind of structure (e.g.a derived or compound word). I also try to show similarities between puns and speech errors. Sometimes, as in the case of Spottmarkt , it cannot be decided, whether the author produced a writing error– a slip of the pen as Victoria FROMKIN 1973 calls it– intending to write Spotmarkt (Eng. "spot market") or whether he wanted to make a joke ( Spott means "mockery"– but there is no real marker for mockery!).

In a large subclass of German particle verbs P+V, P can only fill the prefield if it is immediately followed by V. We argue that this unexpected and largely unnoticed phenomenon is due to parsing problems that arise when P is separated from V by intervening constituents. In addition, we point out two other major problems in the syntax of particle verbs that still await a solution.

In a production study on read German, we investigate the tonal, durational and articulatory parameters used in the marking of three different focus structures. Special attention is paid to target words in Second Occurrence Focus (SOF). Such words are both focussed and contextually given. Prominence in SOF items is compared to prominence in items which are focussed and new (First Occurrence Focus, FOF) and those which are in the Background. Results confirm previous findings that SOF is not marked by a pitch accent. However, we observe gradient but systematic adjustments of acoustic and articulatory parameters (word durations and temporal and spacial modifications of opening and closing gestures) leading to an increase in prominence from Background through SOF to FOF. The extent to which single speakers employ different strategies varies. Phonologically, we propose an analysis of SOF prominences as phrase accents.

This article deals with the underspecified semantics of the Imperfect in Romance. Our main objective is to identify and to describe the three essential "ingredients" of the abstract semantics of the verbal category: 1. the feature of indeterminacy due to its aspectual nature, 2. its anaphoric character and strong contextual dependency and 3. the share it has with the principle of typicity. We will show throughout the article that these three conceptual "essentials" underlie all major readings of the Imperfect in the Romance languages – irrespective of whether they are temporal or modal in nature. In the last part of the article we proceed to a formalization of the basic insights we have gained by analyzing the different readings of the Imperfect in Romance languages.