The aim of this section is to provide a bird’s-eye view of the organization of the clause in standard Dutch and to discuss some of the movements involved in the derivation of the surface forms in actual utterances. Roughly speaking, the clause consists of two main parts, called the lexical and the functional domain. The lexical domain consists of the main verb and its arguments, as well as certain types of modifiers (such as manner adverbs), which together form a proposition. For example, the verb kopento buy in (5a) takes a direct object as its complement and is then modified by the manner adverb snelquickly, and the resulting complex predicate is finally predicated of the noun phrase Jan. The complex phrase thus formed expresses the proposition that can be represented by the logical formula in (5b). For a discussion of the modifiers that can be used in the lexical domain of the clause, see Section 8.2 on VP adverbials.

Infinitival clauses such as (5a) are not normally accepted as independent sentences of Dutch, although they do occur in the special context illustrated in (6b), where participant B expresses surprise at something said by participant A.

The fact that structures such as (5a) do not normally represent acceptable sentences does not mean that the string as such is not syntactically well-formed. This will be clear from the fact that (5a) can be used as the complement of the permissive verb latento let in (7a). The structure as a whole has the propositional content in (7b), in which the proposition in (5b) is embedded in a larger proposition.

The acceptability of (7a) shows that the unacceptability of (5a) as an independent utterance cannot be attributed to the string Jan snel het boek kopen as such, but must be attributed to other factors. More specifically, the contrast between (5a) and (7a) shows that although propositions as such are well-formed expressions of artificial languages like predicate calculus, they must be supplemented with additional information in order to be usable as sentences in natural languages. One such piece of information is tense: to be usable as a sentence, a proposition must be situated in time, as in (8).

Given that the infinitival clause Jan snel het boek kopen can be used in (7a), in which the temporal information is expressed by the past tense on the verb form lietlet, we can conclude that this information is external to the lexical domain. For this reason, it has been proposed that the lexical domain of the clause is embedded in a larger functional domain. This latter domain contains not only temporal information but also information about the illocutionary force of the expression; for example, it provides an answer to the question whether we are dealing with an assertion or a question. In finite embedded clauses this information is usually provided by complementizers: the complementizer datthat is used for embedded declarative clauses, while ofwhether is used for embedded questions.

Since complementizers are usually words, it has been claimed that they occupy head positions in the functional domain of the clause. A similar line of reasoning claims that the temporal information of the clause is introduced as a temporal head in the functional domain of the clause. If correct, this would lead us to the schematic representation of the clause in (10), where C stands for the head position of the complementizer, T for the head position containing the tense features of the finite verb, and X for other functional heads in the clausal domain (if any). Like lexical heads such as V, functional heads are taken to project and thus form a CP, a TP, and an XP. The projections of V (as well as the other lexical categories N, A, and P) and functional heads will be referred to as lexical and functional projections, respectively. When referring to both the lexical and the functional domain, we will use the term extended projection of the lexical head; cf. Grimshaw (1991) for the origin of this term.

The dots in structure (10) are positions allocated to specific clausal elements (subject, object, wh-phrase, etc.), which appear as so-called specifiers of the lexical and functional heads. These specifiers can be base positions, where certain phrases are lexically inserted, or derived positions, to which certain phrases are moved from other positions in the course of the derivation.

Although the hierarchical structure in (10) is not accepted in all quarters of linguistics, it is generally accepted among generative linguists as universally valid for natural language: specific languages are derived by language-specific and sometimes construction-specific restrictions on the position of the verb in the output of the grammar (C, T, X or V), and something similar holds for the position of the arguments and modifiers of the clause. Of course, this does not change the fact that the postulation of a structure like the one in (10) and the concomitant movements are highly theory-internal. However, readers who object to the movement metaphor from generative grammar might think of structure (10) as the template in (11), in which the positions C, T, X, and V indicate potential positions for the expression of the verb, and in which the dots are designated positions for the expression of certain phrasal constituents (XPs) of the clause. The movements proposed in generative grammar can then be thought of as language/construction-specific expression rules that determine in which positions of the universal template the verbs and phrasal constituents of the clause appear; cf. Broekhuis (2011) for an illustration. Templates such as (11) are also known from theoretical frameworks that do not assume movement; cf. the abstract term patroon (pattern) in Paardekooper (1960) or the term functional pattern in Dik (1978).

Note again that we are not claiming that (10) and (11) exhaust the structural description of the clause; it may well be that the lexical and functional domains contain more heads than indicated here. Nor is it a priori clear that lexical and functional information are as neatly separated as suggested by (10) and (11); it may well be that these types of information are intermingled in more intricate ways. We will merely use structure (10) to provide a global description of the dataset that has been prominent in the discussion of Dutch clause structure in the generative literature over the past four decades (and which, in our view, should be taken into account in any theory), in order to provide the reader with some basic information that may be helpful in reading the present chapter. The reader will find in the following discussions that, despite 50 years of intensive generative research, many questions about clause structure are still unresolved and part of an ongoing debate.