⁰ I am very grateful for the help that I have received, in the form of comments, data, judgements, and references, from Kersti Börjars, Joe Emonds, Elisabet Engdahl, Germen de Haan, Jack Hoeksema, Anthony Kroch, Line Mikkelsen, Marga Reis, Beatrice Santorini, Sten Vikner, and John Whitman. None of those named is however responsible for the use that has been made of their contribution.

¹The distribution of because and for in English is not complementary; rather for appears in a subset of the environments in which because is possible.

Although Wegener appears to claim that German weil+V2 and weil+V-final do not overlap in their interpretations, the native speakers with whom I consulted found that weil+V-final allowed the interpretations possible for weil+V2, as well as the interpretations that are excluded for weil+V2, suggesting that the same is true for these variants in German.

In his discussion of Swedish, Andersson (1975) mentions that därför att (because) can be followed by either V2 or non-V2 order. Därför att+V2 is necessarily outside the scope of e.g. negation in the clause to which it attaches; but he is less clear about the status of Därför att followed by subordinate word order, saying that a reading where takes low scope is "the natural one and for some people I have asked, the only one" (p. 24).

² The paratactic "so" construction in English might be analysable as a special case of complement preposing, given the following parallels:

• Backwards anaphora:

• Failure of binding in either direction:

• Impossibility of negation in final clause

Whether such an analysis could be extended to what Hoeksema & Napoli take to be an essentially identical construction in Dutch is less clear. Further, in German a V2 extent clause can occur in initial position, as shown in (61), but when the extent clause is not initial, it must be verb-final, and introduced by a complementiser (60a,b). For some discussion of the Dutch and German facts, see Section 5.2.3.1.

³ Hoeksema & Napoli also claim that the so/such-inversion in (6a) is a root phenomenon, and refer the reader to Hooper & Thompson 1973 and to Green 1976. However, there is no discussion in these earlier papers of this particular type of inversion, unless it is assumed to be an instance of "preposing around 'be'".

⁴ Hoeksema & Napoli do not discuss in any detail the claim that the order of the two clauses is fixed (p. 296). In fact it is not at all obvious that this claim is correct, given the grammaticality of the following examples from English:

Similar examples are also grammatical in Dutch (Jack Hoeksema, personal communication): see Section 5.2.3.1.

⁵ The force of Fabb's examples with "any" is somewhat weakened by the relative acceptability of some restrictive relatives on definites with "any" but without "only":

The same point, however, can be made with "ever" (Beatrice Santorini, personal communication):

⁶ Apparent counterexamples like (ia) are presumably the result of the same 'telescoping' phenomenon that allows the binding in (ib):

⁷ Although Hooper & Thompson do not give any examples, direct quote preposing may also cooccur with inversion as long as the subject is not a pronoun:

⁸ Of course, one could attempt to defend Hooper & Thompson's position by rejecting their own assumption that clefting and topicalization (or any other root transformation) are sufficiently similar to make a legitimate minimal pair.

⁹ In Dutch, when the extent clause is initial, it must be followed immediately by the zo+AP phrase unless the interpretation is simply one of "asyndetic conjunction" (Hoeksema & Napoli 1993: 293); in German this order is also preferred, but does not appear to be obligatory (Beatrice Santorini, personal communication).

¹⁰ Dutch is also like Frisian and unlike German in that the acceptability of embedded V2 is affected by negation in the matrix.