Examples (14a&b) show that verbs can occur in different positions in the main clause; finite verbs occupy a position in the left periphery of the clause, while participles and infinitives occupy a position further to the right. Work in the structuralist tradition, such as Haeseryn et al. (1997), often refers to the position of the finite verb as the first pole of the clause and the position of the non-finite verbs as the second pole of the clause. Example (14c) shows that the second pole can remain empty if there are no non-finite verbs to fill it.

Using the idea of two poles, we can divide main clauses into three subdomains. The first subdomain consists of the position before the finite verb. This position is often occupied by the subject, as in the examples in (14) above, but the primeless examples in (15) show that it can also be occupied by e.g. a questioned or topicalized direct object. The crucial observation, however, is that the finite verb can normally be preceded by only a single constituent; this will be clear from the fact that filling the position preceding the finite verb with a constituent other than the subject requires that the subject be placed after the finite verb; leaving the subject Jan in the position before the finite verb, as in the primed examples in (15), leads to unacceptable results.

Since the position preceding the finite verb can contain at most one constituent, this position is often called the clause-initial position; accordingly, the position of the finite verb is often called the second position of the clause, to contrast it with the clause-final position occupied by the non-finite verbs. The examples in (15) show that the term clause-final position can be somewhat misleading, since verbs in this position can be followed by other elements. The examples in (16) show that this is easily possible in the case of PP-complements, and even obligatory in the case of clausal complements, and in many cases we can find more than one constituent there. The positions following the clause-final verbs are called postverbal positions.

Since the clause-initial position is usually filled by some constituent in declarative clauses and wh-questions, the term verb-second position is quite appropriate for such cases. However, there are also cases where the clause-initial position remains empty so that the verb ends up in first position, as in the yes/no

questions in (17).

The examples in (18) show that an adverbial phrase in the form of a PP or a clause can also occur in a postverbal position. Note that clause adverbials differ from clausal complements in that they can occur both pre and postverbally.

The postverbal field is usually occupied by PPs and clauses, but this does not exhaust the possibilities: some adverbs may also occur postverbally. This is illustrated in (19a) for the modal adverb waarschijnlijkprobably.

Adverbial phrases indicating manner are special in that they categorically resist postverbal placement; the examples in (20) show that this is true not only for adjectival manner adverbs, but also for the corresponding prepositional forms.

Note in passing that the examples in (21) show that the ungrammatical orders in (20) improve considerably when the postverbal phrases are preceded by an intonation break and given an emphatic accent. In such cases, the adverbials function as afterthoughts, which are often considered to be external to the main clause and thus belong to the class of elements to be discussed in Chapter C37.

The area between the verbs in second and clause-final position is often called the middle field of the clause. This part of the clause can contain virtually all constituent parts of the clause, with the notable exception of clausal arguments; cf. (16a) above.

The most striking property of main clauses is that they usually require their finite verb to be in second position; the examples in (22) show that the embedded counterparts of the main clauses in (14) require the finite verb to be in clause-final position, just like non-finite verbs.

This means that, in general, the examples in (14) cannot be embedded as such; the number sign indicates that the examples in (23) can only be interpreted as direct (i.e. quoted) speech. That these examples cannot be interpreted as indirect speech is not a trivial fact, since this is possible in German and, to a lesser extent, in some eastern Dutch dialects; cf. Haider (1985/2010) and Barbiers (2008: §1.3.1.8).

The examples in (24) show that the cases in (23) do not improve when we add the complementizer datthat. Again, this is not a trivial fact, since this is the natural way to form embedded declarative clauses in English; cf. John said that John has sent Mary the book from China.

However, the requirement that the finite verb be clause-final is not absolute; there are a number of adverbial clauses that allow the verb to be in first/second position. The examples in (25) show, for instance, that conditional clauses can be introduced by the complementizer-like element alsif and have the finite verb in clause-final position, but they can also occur without als and then have the finite verb in first position. Such exceptions will be discussed in Section 10.3.

The discussion so far has shown that main and embedded clauses differ in the position of finite verbs: they appear in second position in main clauses, but in clause-final position in embedded clauses. The current standard analysis relates this difference to the distribution of complementizers: these are usually excluded in main clauses, but obligatory in embedded clauses. Paardekooper (1961) has shown that finite verbs in main clauses and complementizers in embedded clauses have the same position with respect to pronominal subjects. Leaving aside subject-initial main clauses, the examples in (26) show that such subject pronouns are always right-adjacent to the finite verb in main clauses and to the complementizer in embedded clauses.

Paardekooper concludes from this that finite verbs in main clauses have the same position as complementizers in embedded clauses. He suggests that this similarity of placement is related to the fact that complementizers and finite verbs have a similar relationship to the subject of the clause, as can be seen from the fact that in certain Dutch dialects (but not in standard Dutch) complementizers and finite verbs can agree in number and person with the subject of the clause. This is illustrated by the two examples in (27) from Van Haeringen (1939). Note that in these examples the complementizer aswhen introduces temporal adverbial clauses, but that we find similar agreement in complement clauses introduced by the declarative complementizer datthat or the interrogative complementizer ofwhether; cf. Haegeman (1992: §2.2), Hoekstra & Smit (1997), Zwart (1997), Van Koppen (2005/2017) and the references given there for examples and more information.

Paardekooper did not discuss the relation between the two positions of the finite verb in main and embedded clauses. The nature of this relationship, however, became a pressing issue in early transformational grammar, in which the surface representations of sentences were assumed to be transformationally derived from more abstract underlying forms. The main question was: which word order is more basic, that in main clauses or that in embedded clauses? Koster (1975) argued convincingly that the order found in embedded clauses is more basic, on the basis of the following economy argument. If we assume that all verbs are base-generated in clause-final position, we only need a single verb-second rule that operates in main clauses and places the finite verb in second position: the rule in (28) simply expresses that finite verbs can be placed in second position in main clauses (X, Y and Z simply stand for an unspecified string of elements).

However, if we assume that all verbs are base-generated in second position, we need at least two rules: (i) one rule that places all non-finite verbs in clause-final position, and (ii) another rule that places the finite verb in clause-final position in embedded clauses. In fact, we need many more word-order rules under this assumption; cf. Koster (1975) and Zwart (2011: Part III) for more detailed discussions.

Building on Paardekooper’s insight, Den Besten (1983) added that Koster’s verb-second rule can be formulated in such a way that we can appeal to independently needed positions, by assuming that the finite verb moves into the position normally occupied by the complementizer in embedded clauses; this was an important step, as it made the verb-second rule consistent with Emonds’ (1976) structure-preservation constraint, according to which movement cannot create a new position. The difference between main and embedded clauses is depicted in (29) on the basis of the more recently developed clause structure in (10). Note in passing that it is usually assumed that head movement cannot skip intervening heads like T or X (but moves through them in a successive cyclic fashion); we will not discuss this here, but will return to it in Section 9.3.

If we take the examples in (26) to show that subject pronouns obligatorily occupy the specifier of TP, i.e. the position left-adjacent to the T-head, this combination of the earlier findings by Paardekooper and Koster provides a simple formal account of the basic standard Dutch facts discussed so far.

This section has briefly discussed the placement of verbs in main clauses and embedded clauses. We have seen that verbs normally occur in clause-final position, with the exception of finite verbs in main clauses, which occur in second position. We argued that this second position is the same as that occupied by complementizers of embedded clauses. Verb-second position C and clause-final position V divide the clause into three parts, as shown in Figure (30). Sections 9.3 to 9.5 will discuss these parts in more detail.