Section 1.5.1 discussed the binary tense theory proposed in Te Winkel (1866) and Verkuyl (2008), according to which the three binary distinctions in (376) are used in mental representations of tense. Languages may differ in the grammatical means used to express the oppositions in (376): this may be done in the verbal system by inflection and/or auxiliaries, but it may also involve the use of adverbial phrases, aspectual markers, pragmatic information, etc. Verkuyl claimed that Dutch expresses all the oppositions in (376) in the verbal system: [+past] is expressed by inflection, [+posterior] by the verb zullenwill, and [+perfect] by the auxiliaries hebbento have and zijnto be.

Section 1.5.2 has argued at length that the claim that zullen is a future auxiliary is incorrect: it is an epistemic modal, and it is only because of pragmatic considerations that examples with zullen are sometimes interpreted with a future time reference; cf. also Broekhuis & Verkuyl (2014). If this is true, then the Dutch verbal system only expresses the binary features [±past] and [±perfect]; it does not make an eightfold, but only a fourfold tense distinction based on inflection and auxiliaries. This means that the traditional view on the Dutch verbal tense system in Table 9 from Section 1.5.1, sub I, must be replaced by the one in Table 11; the verb zullen is no longer seen as a marker of a separate set of future tenses. Of course, this does not mean that the future tense is not relevant for Dutch, but only that posteriority must be expressed by other means (e.g. the use of time adverbials such as morgentomorrow).

	present	past
imperfect	simple present (o.t.t.) Ik wandel/Ik zal wandelen. I walk/I will walk	simple past (o.v.t.) Ik wandelde/Ik zou wandelen. I walked/I would walk
perfect	present perfect (v.t.t.) Ik heb gewandeld/ Ik zal hebben gewandeld. I have walked/I will have walked	past perfect (v.v.t.) Ik had gewandeld/ Ik zou hebben gewandeld. I had walked/I would have walked

This revised view on the Dutch verbal tense system implies that utterances in the simple present/past can normally refer to any event time interval in the present/past-tense interval i; eventuality k can precede, overlap or follow n/n', as shown in Figure 24. Recall that the number of possible worlds is in principle infinite and that we simply select a number of them that suit our purpose.

The representation of the perfect tenses is virtually identical to that in Figure 24, the only difference being that the eventualities are construed as completed autonomous units within the present/past-tense interval. As before, we indicate this in Figure 25 by a vertical line at the end of the event time interval k.

In the figures above we have assumed that the default value of the present j of eventuality k (i.e. the time interval within which the eventuality denoted by the lexical projection of the main verb must take place) is equal to that of the complete present/past-tense interval i. The following sections will show that contextual information (linguistic and non-linguistic) can override this default interpretation and lead to more restricted interpretations.

Before proceeding with a discussion of the Dutch verbal tense system, we would like to note that although Verkuyl (2008) was probably wrong in assuming that the binary tense theory is perfectly reflected in this system, it seems that Dutch is very suitable for studying the interaction of tense, modality and pragmatic information, because it can be characterized as a strongly “tense-oriented” language. First, Dutch does not normally mark mood on the verb (the exception being imperative marking), so it differs from German in that it does not have a productive subjunctive marking on the verb; cf. Section 1.4.3. Second, Dutch does not normally mark syntactic aspect on the verb, so it differs from English in that progressive aspect can be expressed by using the simple present/past. Third, Dutch does not require epistemic modality to be marked, so it differs from English in that the expression of non-actualized (“future”) events need not be marked by the modal will; Dutch zullen is optional in such cases. Finally, it may be useful to mention that adverbial phrases such as gisterenyesterday that refer to temporal intervals preceding speech time can be used in Dutch present-perfect constructions, i.e. Dutch does not have the property found in English that such adverbials can only be used in past-tense constructions: cf. Broekhuis (2021) and Verkuyl (2022: §3.3.3) for two possible accounts of this contrast from a binary tense perspective. All in all, this means that Dutch allows us to directly investigate the interaction of past tense, epistemic modality, and pragmatics in the derivation of special meaning effects, without the intervention of any of the more idiosyncratic properties concerning mood/modality, aspect and adverbial modification of the type mentioned above.