A particular understanding of clause structure has emerged in recent years in which a large number of highly specialized functional heads occur in a fixed order, universally. The clearest and most detailed articulation of this position is Cinque's 1999 book. Cinque argues, among other things, that there is an epistemic modality projection (hosting, e.g., "probably") above a modality of necessity projection (hosting "necessarily") above a frequentative aspect projection (for "often") above a perfective aspect projection ("always"). Cinque identifies at least thirty different rigidly organized categories, but interestingly negation is not among them. Instead, following Zanuttini, Cinque supposes that negation can appear in a variety of different places.

Thus, on this view, there is a fundamental difference between the rich inventory of modal, temporal, and aspectual operators on the one hand, and negation on the other. In this talk I examine the possibility that there are other features which are more like negation in this way, and discuss what it means for the worldview sketched above.

In this talk I will try to give an analysis of stylistic fronting that takes facts that have so far received little attention in the literature into consideration, namely that stylistic fronting has semantic effects and that stylistic fronting is possible in subordinate clauses with a weak subject pronoun. In particular, I will propose that stylistic fronting is a feature driven movement into FocusP in the sense of Rizzi (1997).

On various occations, stylistic fronting has been related to the loss of V°-to-I° movement in the Mainland Scandinavian languages. The reason is that stylistic fronting reestablishes the base-generated word order. Although Icelandic has V°-to-I° movement, Icelandic can have the same word order as Mainland Scandinavian if a negation or a medial adverb has undergone stylistic fronting.

Stylistic fronting of a negation:

(1) Allir sem  ekki höfðu borðað kæstan    hákarl veiktust
    All   that not  had   eaten  fermented shark  got.sick

This sentence is ambiguous with respect to both stylistic fronting and V°-to-I° movement. It can either be analysed as a sentence with both stylistic fronting and V°-to-I° movement or it can be analysed as a sentence with neither stylistic fronting nor V°-to-I° movement.

This only holds for clauses in which a negation or a medial adverb has undergone stylistic fronting. Stylistic fronting of participles and verb particles creates a word order that is ambiguous with respect to V°-to-I° movement but not with respect to stylistic fronting.

Stylistic fronting of a participle in a subject relative clause:

(2) Allir sem  borðað höfðu t kæstan    hákarl veiktust
    All   that eaten  had     fermented shark  got.sick

In this talk I will try to show that because of stylistic fronting of the negation and medial adverbs (i.e. because of sentences like (1)), V°-to-I° movement is no longer obligatory in subordinate clauses with no overt subject in Icelandic. I will also try to show that the possibility of not having V°-to-I° movement in subordinate clauses with no overt subject is spreading to subordinate clauses with an overt subject.

According to Zhang (1990), natural languages have two strategies for forming negative imperatives, employing either: (i) the language’s regular negative marker or (ii) a special imperative negative marker. A special negation marker may be in a different structural position to the regular negation and/or of a different status (either Xº or XP). This paper’s primary aim is to identify the strategy used in English, Danish, Norwegian and Swedish and to determine the status of the negative marker in each language. A second goal of this paper is to articulate the nature of the imperative clause structure. CP and IP analyses will be entertained.

The topic of this paper is suppletive imperatives – negative commands comprised of a negative marker plus an infinitive or subjunctive verb. After presenting the well-documented literature on Romance negative commands, I will turn to such constructions in other languages including Danish motherese constructions such as Ikke lege med maden. A variety of possible analyses will be entertained.

In most the Scandinavian languages, pronominal objects undergo obligatory object shift (OBJ-shift) when the main verb is in V2 position. In Swedish, however, this operation is optional. Unlike all the other Scandinavian languages, Icelandic also has OBJ-shift of full (non-pronominal) DP objects. Both pronominal and full-DP OBJ-shift are subject to a requirement on linear order of arguments stating that the indirect object should always precedes the direct object. In double-object constructions, OBJ-shift only applies to one or both objects as long as this requirement is satisfied. Thus, OBJ-shift can be suppressed in order to respect the word order of the objects.

When one of the objects in a double-object construction is a negative object, the picture changes. Sentential negation in the Scandinavian languages must be expressed within NEGP, either by direct insertion of a negative adverbial or by obligatory movement of a negative object (NEG-shift). This constraint on negation, known as the Negative Criterion, is not subject to the requirement that the main verb is in V2 position, but it must respect the requirement on word order. Thus, this has effects in compound tense which normally blocks OBJ-shift. Furthermore, the presence of a negative object leads to neutralization of optionality in Swedish.

I present a syntactic analysis of these phenomena within the framework of Optimality Theory which accounts for the linguistic variation in terms of differential rankings of general universal constraints.

In most Scandinavian varieties, the expression of sentential negation is obligatorily VP-external, though this is often masked by various kinds of movement rules which vacate the VP. The lack of transparency of the structures involved, coupled with the problem of negative evidence (in the original sense), create a learnability problem. Such problems are often evaded by postulating a universal rule; in this case, it would be assumed that sentential negation is universally expressed outside the VP. However, this is false, as languages like English allow VP-internal expression of sentential negation. I examine a couple of alternative solutions to the problem.