Jump to ContentJump to Main Navigation
Show Summary Details
Weitere Optionen …


An Interdisciplinary Journal of the Language Sciences

Editor-in-Chief: Gast, Volker

IMPACT FACTOR 2018: 1.066

CiteScore 2018: 0.97

SCImago Journal Rank (SJR) 2018: 0.384
Source Normalized Impact per Paper (SNIP) 2018: 1.409

Alle Formate und Preise
Weitere Optionen …
Band 57, Heft 5


(Inter)subjective uses of the Dutch progressive constructions

Lynn Anthonissen / Astrid De Wit / Tanja Mortelmans
Online erschienen: 08.06.2019 | DOI: https://doi.org/10.1515/ling-2019-0019


This paper addresses the (inter)subjective functions of progressive aspect in Dutch. While the aspectual profile of the various Dutch progressive constructions has received considerable attention in the last few years, much less attention has been paid to their non-aspectual uses. As we will demonstrate in this paper on the basis of a corpus study of spoken Dutch, complemented with native-speaker elicitations, the Dutch progressive constructions can be specifically recruited to express (inter)subjective meanings such as surprise, irritation and intensity, and they differ in this respect from their simplex counterparts. Our analysis of progressive aspect in terms of backgrounded boundaries provides an explanation for (i) the general association of progressive aspect with (inter)subjectivity and (ii) our observation that some Dutch progressive constructions are more prone to such (inter)subjective exploitation than others. This semantic account also underlies the last part of this contribution, in which we discuss cases of what we call “(inter)subjective reinforcement” in complex progressive constructions, that is, the embedding of progressive constructions in other constructions that are also semantically affiliated to (inter)subjectivity (e.g. the perfect, gaan/komen ‘go/come’, modals and the bare infinitive construction), which has been largely neglected in the literature.

Keywords: progressive aspect; (inter)subjectivity; Dutch; posture/motion verbs

1 Introduction

Semantic analyses of progressive aspect typically concentrate on aspecto-temporal notions such as ongoingness, duration or incompletion. Yet various scholars have pointed out that speakers of different languages can also recruit the progressive to encode (inter)subjective meanings. In those cases, the progressive expresses “a particular attitude of the speaker vis-à-vis the epistemic status of the situation (subjective interpretations) and vis-à-vis the hearer (intersubjective interpretations)” (De Wit and Patard 2013: 119). In French, for instance, the progressive is associated with events that somehow “stand out” (1a), while it is less felicitous with events that the speaker considers expected (1b) (Franckel 1989: 78):

Ilesten train delirede travers.
hebe.prs.3sgprogread.infthe wrong way

‘He is reading the wrong way.’

? Ilesten train debienlire.

‘He is reading well.’

Similarly, speakers of German can resort to the progressive when they want to highlight an event, even if the German am-progressive is infrequently used in comparison with its Germanic equivalents (Anthonissen et al. 2016). In (2), for instance, the progressive is specifically exploited to convey a sense of intensification:

explain.prs.3sgtax consultantGüntherPöschl

‘“At the moment, we are calculating fiercely,” tax consultant Günther Pöschl says, …’

(Anthonissen et al. 2016: 19)

In English, in which the progressive is obligatorily deployed to report events that are ongoing at the time of reference (barring a few exceptional contexts – see De Wit 2017a: Ch. 4), it is harder to find uses in which the progressive is exclusively used for (inter)subjective purposes and not (also) for aspecto-temporal reasons. Yet the so-called “interpretative” use (Ljung 1980), illustrated in (3), constitutes a nice illustration of the fact that, in English too, progressive aspect can be used to put emphasis on a situation:


[In a discussion between a professor and his students about the discourse of civil rights activist Jesse Jackson]

Well he says minorities. He’s smart. He talks about minorities. But he’s really talking about African Americans.

(De Wit and Brisard 2014: 52)

In this example, the speaker first uses the simple present to report what one might think is ordinarily the case (what Jesse Jackson literally says) and then switches to the present progressive to emphasize what is actually the case.

These observations about the use of the progressive for the expression of (inter)subjective meanings extend beyond western Europe. In her study of the Western Armenian progressive and evidential constructions and their relation to salience, Donabédian-Demopoulos (2012) observes that the progressive can be used for non-aspectual reasons in contexts very similar to the ones described by Franckel (1989) for French. In the Niger-Congo language Igbo, furthermore, there is what is called a “Progressive-Unexpected” construction, which is used to express the speaker’s irony or surprise with regard to the state of affairs she describes (Emenanjo 1987: 175). In a larger-scale study on Bantu languages, Güldemann (2003) furthermore shows that progressive constructions are often used in focal contexts, i.e. those parts of linguistic expressions that contain information that is relatively the most salient within a given communicative setting. Again, the progressive is being used in such cases to present situations as “standing out”.

In this paper, we will demonstrate that, in Dutch too, the progressive can be deployed to convey (inter)subjective meanings. In those cases, the difference between the progressive-form utterance and its simplex counterpart is not aspecto-temporal, but resides instead in the sense of non-canonicity that is more naturally associated with the progressive variant. What makes the case of Dutch particularly interesting is the fact that, unlike the other European languages mentioned above, Dutch has several grammaticalized progressive constructions: 1 the aan het-progressive (4), the posture verb construction (zitten/liggen/staan te + V) (5)–(7), and the motion progressive (lopen te + V) (8). 2 The following examples illustrate the use of these different progressive constructions for the verb lezen ‘read’ (each of them translates as ‘Maria is reading’):


The original lexical meaning of the posture and motion verbs is typically bleached. While for (7) an interpretation in which Maria is standing while she is reading is still readily available, this is clearly not the case for the attested example in (9), since it is hard to believe that someone is constantly standing up while cycling:

TerwijlTom BoonenmetzijnploegmaatsvanEtixx-QuickStep
whileTom BoonenwithhisteammatesofEtixx-QuickStep

‘While Tom Boonen and his Etixx-QuickStep teammates were constantly second best the past few weeks, Vanmarcke was cycling on the first row…’

(Metro April 7, 2016)

As far as overall frequency of use and degree of entrenchment in the aspectual system of the language are concerned, it seems that the Dutch progressive is sandwiched in between its English and German analogues. While it is not obligatory to use the progressive in Dutch to report ongoing events – unlike in English, the simple tenses can still take on this role as well –, there are clearly fewer syntactic restrictions on the Dutch progressive than on its German counterpart (see e.g. Anthonissen et al. 2016: 4–6).

To some extent, the analysis of the (inter)subjective uses of the Dutch progressive constructions can be considered a follow-up study of previous analyses of the modal/(inter)subjective import of the progressive in English, French and German (see De Wit and Patard 2013; De Wit et al. 2013; De Wit and Brisard 2014; Anthonissen et al. 2016). It thus fills a gap in the extant literature on the Dutch progressive constructions, which has overwhelmingly been concerned with the temporal and aspectual meanings of these constructions (see, most notably, Beekhuizen 2010; Boogaart 1999; Booij 2008; Lemmens 2005, Lemmens 2012a, Lemmens 2012b, Lemmens 2015). Yet our study also adds important new insights to the study of progressive aspect. For one, this analysis is the first to corroborate corpus findings with native-speaker elicitations, thus providing additional independent evidence for our claim that it is the progressive, and not some other contextual factor, that generates the attested (inter)subjective reading. Secondly, the fact that speakers of Dutch have various grammaticalized progressive constructions to their avail raises an interesting and novel issue: if Dutch progressive aspect can convey (inter)subjective meanings – which is what we expect given the observations on other (Germanic) languages, and in view of our general hypothesis regarding the connection between (inter)subjectivity and progressive aspect (see Section 3) – then we might wonder whether one or more construction is particularly prone to (inter)subjective exploitation, and if so, why.

In sum, two main research questions lie at the basis of this paper, each of them having a descriptive and an analytic component: (i) is it indeed the case that the Dutch progressive constructions can express (inter)subjective senses, such as surprise, irritation or intensification, and why should we attest this association, and (ii) are there any differences in this respect between the various progressive constructions and how can we account for these differences? In order to answer these questions we have undertaken a large-scale corpus investigation and developed a native-speaker survey, the design and the results of which we will discuss in detail in Section 2. In Section 3 we analyze our findings: building to some extent on previous observations by De Wit and Brisard (2014), Lemmens (2005, 2015) and Beekhuizen (2010), we offer theoretical and cognitively plausible explanations for the connection between progressivity and (inter)subjectivity and the extent to which this connection is manifested by the different progressive constructions. Section 4, finally, concentrates on four interesting instances of, what we call, “(inter)subjective reinforcement”. These are cases in which a posture/motion verb progressive is embedded in another construction (i.e. is combined with the semi-auxiliary gaan ‘go’ or komen ‘come’, the perfect, a modal verb or the bare infinitive construction), with the effect of reinforcing the sense of incongruity that is already canonically associated with the progressive. Our conclusions will be formulated in Section 5.

2 Data, methodology and results

2.1 Corpus study

2.1.1 Data collection and analysis

The data for the corpus study were extracted from the Corpus Gesproken Nederlands [CGN] v2.0 ‘Corpus Spoken Dutch’ (Nederlandse Taalunie 2006), an annotated corpus of present-day Dutch as spoken by adults in the Netherlands and Flanders. More specifically, we selected CGN’s a component (henceforth CGNa), which comprises spontaneous face-to-face conversations and amounts to approximately 2.6 million words. All instances of the various progressive constructions listed in (4)–(8) were retrieved from CGNa with the corpus exploration software COREX.

The query for the aan het-progressive included a form of the lemma zijn 3 ‘to be’ within a seven-word window before or a five-word window after the combination of the preposition aan and the definite article het ‘the’ or its contraction ‘t. The basic query for the remaining progressive constructions consisted of the lemma of the posture or motion verb at hand (zitten/liggen/staan/lopen ‘sit/lie/stand/run’) followed by te ‘to’ and an infinitive. Separate queries were conducted to retrieve posture and motion progressives in non-finite contexts, which were not obtained by the basic queries due to their deviating morpho-syntactic behavior in those contexts: when used in the perfect and pluperfect (see Example (10)), the posture and motion progressives do not have the expected past participle form (*gezeten), but appear as infinitives (zitten) – a phenomenon known as infinitivus pro participio (IPP). The particle te ‘to’ is usually dropped, not only in IPP constructions, but also in other non-finite contexts, that is, when the posture or motion progressive is preceded by a modal (such as kunnen ‘can’ in [11]) or another (semi)-auxiliary verb (such as gaan ‘go’ in [12]). In cases like (13), the posture verb zitten appears as an infinitive in a main clause construction without there being a finite auxiliary; we will refer to this construction type as the ‘bare infinitive progressive’ (see Section 4.4).


‘I’ve already been thinking about starting selling again.’

(CGNa fv400519) 4


‘We might as well throw everything away, I’m not doing that.’

(CGNa fn00969)


‘Surely you don’t mash that.’

(CGNa fv400623)


‘I was like explaining the whole thing.’

(CGNa fn000590)

Since they involve the presence of an extra auxiliary (or, in case of the bare infinitive progressive, the absence of an auxiliary where it would have been expected), we call progressives of the types presented in (10)–(13) “complex progressives” and set them apart from “simplex progressives” (with finite forms of the progressive auxiliary verbs zijn ‘be’ or their posture/motion verb counterparts). After manual inspection for doubles and false positives, the remaining sample contains 2,465 targets in total.

Our data analysis proceeded in a step-wise fashion. First, instances were grouped according to construction type. From these various groups ca. 10% of the instances (255 in total) were sampled to test inter-rater agreement for the remaining variables to be annotated (most importantly, the presence of (inter)subjective qualifications, but also a number of other variables such as aspecto-temporal usage types, tense, subject animacy, presence of negation, and telicity of the main verb). 5 The sample was coded by the three authors independently, and subsequently compared and discussed to avoid inter-rater bias. 6 The remaining instances were divided among the three annotators (again with a roughly even representation of each construction type to ensure that every annotator had a similar distribution of examples). Next, every instance was checked by the two other annotators, independently from each other, though not blind for its first classification (i.e. the annotators indicated whether they agreed with the assessment of the first annotator). Any remaining cases of interrater disagreement were resolved by discussion.

Since our contribution analyzes the (inter)subjective import of the progressive in particular, some further methodological clarifications are in place. We verified for each instance whether the progressive expressed an (inter)subjective qualification in addition to its basic aspecto-temporal characterization in terms of ongoingness or related notions. Such (inter)subjective construals may result from various emotional reactions (e.g. irritation, surprise) or discourse-pragmatic motivations (e.g. to confer more emphasis on the situation at hand or to introduce an opinion or statement in a tentative manner), examples of which are discussed in Section 2.1.2. In order to categorize our corpus examples as (inter)subjective or not, we first relied on two parameters: (i) contextual indications and (ii) interchangeability with the simple tenses; in a later stage, we complemented our interpretation of the corpus data with those of (other) native speakers – see Section 2.2. Assuming that contextual cues, such as adverbials, elaborate schematic meaning elements already present in the semantic configuration of the progressive (Langacker 1987: 304–306), we consider contextual signs of irritation, surprise or other (inter)subjective notions to be strong indications of an (inter)subjective meaning component of the progressive. The fact that contextual elements need to be semantically compatible with the construction with which they collocate can, for instance, also be inferred from the observation that a time adverbial like yesterday is congruous with a past, but not with a future tense. Consider also the following examples:


Present progressive


‘He is nagging again.’


Simple present


‘?He is nagging again.’

While the sense of irritation in the constructed example in (14) may be argued to result solely from the repetitive adverb weer ‘again’ and the negative connotation already present in the verb zagen ‘to nag’, a simple-present construal of the same sentence, as in (15), would sound distinctly odd to the native speaker’s ear. Put differently, we assume that various meaning elements in the sentence contribute to conveying the speaker’s attitude towards the proposition (much like both the adverbial yesterday and past tense morphemes may redundantly express “past-ness”). Yet even if a context contains clear indications of (inter)subjective qualifications, we cannot entirely exclude that the main reason for using the progressive is aspecto-temporal (e.g. to indicate that a situation is ongoing rather than complete), and that, consequently, the attested (inter)subjective connotations are solely contextually induced. Therefore, we also took into account the effects of substitution tests: if the replacement of the progressive by a simplex form does not result in an aspecto-temporal difference, but rather in a more factual, less outstanding presentation of the situation at hand, we have strong grounds to establish that the progressive can truly be said to reflect an (inter)subjective conception of reality. The native speaker questionnaire that we developed – and whose design and outcomes are presented in Section 2.2 – serves to corroborate our intuitions on the appropriateness of progressive versus non-progressive aspect in the exact same (inter)subjective context.

2.1.2 Results

Our corpus study reveals that there are a variety of (inter)subjective qualifications associated with the progressive. For instance, Dutch progressives may carry overtones of irritation, an observation that is also made in current grammars of Dutch (e.g. Haeseryn et al. 2002: Section; Oosterhoff 2009: 91). The utterance in (16) constitutes a typical example.

[was d ook één jongen in ‘t team en z en dat vind ik nou wel een beetje irritant worden]
‘There was also one guy in the team and I think it’s getting a little annoying …’

‘… cause he’s always shouting if you do something wrong.’

(CGNa fn000481)

Speakers of Dutch also resort to progressive construals to depict the situation at hand as manifesting a quality to an unusually large extent. This intensifying use is illustrated in (17).

follow.prs.3sgmeno one

‘I was too scared to walk around most neighborhoods […]. At every street corner I was really looking around like isn’t anyone following me.’

(CGNa fn000571)

A simple present construal, though grammatically possible, would render the situation in (17) as less outstanding than the selected progressive construal. Progressives are furthermore found in contexts featuring a sense of surprise, in that the reported event is in a way incongruous in the speaker’s conception of reality. This is, for instance, the case in (18), in which the use of the simple past would be quite marked.


‘Oh Niels’s last hallway party, Niels himself was … was shaking like that on the window sill.’

(CGNa fn000536)

In essence, the (inter)subjective notions illustrated in (16)–(18) highlight the speaker’s experience of the situation. While these expressions partly reflect the speaker’s current state of mind, they also serve a more interpersonal purpose related to the expressive function of language: in accordance with the so-called maxim of extravagance, the speaker wants to “[t]alk in such a way that [she is] noticed” or in a way that is considered “amusing, funny, etc.” (Keller 1994: 101). A related maxim, “[t]alk in an especially polite, flattering, charming, etc. way” (Keller 1994: 101), is even more explicitly motivated by the speaker’s wish to achieve a particular effect on the hearer. Such interpersonal dynamics are for instance at stake when the speaker wants to introduce her ideas in a more cautious manner. We know from previous studies (e.g. Anthonissen et al. 2016 on the German am-progressive) that progressive aspect can be used to obtain such an effect, and this use is attested in the Corpus of Spoken Dutch as well, cf. (19).

momtotheshopping malltogo.inf

‘I was actually thinking of going shopping with mom tomorrow.’

(CGNa fv400384)

In (19) the speaker is politely declining an offer by tentatively announcing that she has already made other plans. She explicitly wants to construe the situation as non-consolidated, suggesting she is not yet fully committed to going shopping so as not to offend the hearer. Again, the use of the simple past would not generate the same hedging effect.

Our corpus study amply shows that the examples in (16)–(19) are not isolated cases and that there is, in fact, a recurrent association of (inter)subjective notions with progressives. Frequency statistics of our analysis show that the Dutch progressives are used for purely aspecto-temporal reasons in roughly half of all cases (n = 1,222, 49.57%). The other half (n = 1,243, 50.43%) also has an aspecto-temporal import (in that we can, for example, attest a sense of ongoingness or limited duration), yet we contend that their use is mainly motivated by (inter)subjective purposes, basing ourselves on contextual indications signaling non-canonicity and on the observation that a simple-present or -past report of the same situation would be less felicitous in combination with these contextual indications, since they yield a less subjectively colored interpretation (rather than an aspecto-temporal difference) – see also Section 2.2.

The corpus study moreover sheds light on inter-constructional differences pertaining to (inter)subjectivity. Before zooming in on these differences, let us give a brief overview of the overall frequency distribution of the various progressive constructions in Table 1, which shows that the Dutch progressives exhibit varying levels of entrenchment.

Table 1:

Frequency distribution of construction types in CGNa.

The bulk of examples in CGNa is taken up by the prepositional aan het construction and the posture verb construction with zitten, which are almost equally frequent and together account for nearly 85% of all attestations. Staan is attested considerably less often and lopen and liggen are infrequent in comparison to the overall sample. The frequency distribution thus centers around two well-entrenched constructions, which may well be considered progressive prototypes.

The extent of constructional specialization in terms of (inter)subjectivity is illustrated in Figure 1. Dark grey bars represent the proportion of (inter)subjective uses per construction type, light grey bars the proportion of instances that are exclusively aspecto-temporally motivated. The width of the bars reflects the relative prevalence of the construction types in CGNa (cf. Table 1).

Mosaic plot of (inter)subjective and non-subjective uses per construction type (n = 2,465).
Figure 1:

Mosaic plot of (inter)subjective and non-subjective uses per construction type (n = 2,465).

The aan het-progressive exhibits the lowest proportion of (inter)subjective uses among the various construction types. Moving to the posture verb progressives (liggen/staan/zitten), we witness a 25% increase in the relative number of (inter)subjective readings. Instances of motion progressives (lopen) almost invariably involve a subjective evaluation of the state of affairs. Strikingly, the posture verb progressives emerge quite clearly as a class in their own right that is positioned in between the prepositional progressive and the motion progressive on a cline of increasing (inter)subjective import. 7

It is important to remember that previous analyses have generally focused on attestations of the construction’s basic form, i.e. what we refer to as “simplex progressives”. Although these simplex progressives are attested in far greater numbers (n = 2,083, 84.50%), the group of complex progressives is not negligible (n = 382, 15.50%). These complex progressives constitute an interesting topic for further exploration because, as Table 2 shows, they are highly susceptible to (inter)subjective readings, significantly more so than the simplex progressives. 8 Therefore, a separate section (Section 4) will be devoted to complex progressives as an excursion into (inter)subjective reinforcement.

Table 2:

(Inter)subjective uses of simplex and complex progressives in CGNa.

In this section we have discussed our primary dataset, taken from spontaneous face-to-face conversations. From this analysis we can already infer that, in general, Dutch progressives are prone to be recruited for the expression of (inter)subjective qualifications and that the different construction classes differ greatly with respect to the relative frequency of (inter)subjective uses. Our native speaker survey, which we present in the following section, provides further empirical evidence for these observations.

2.2 Native speaker survey

2.2.1 Survey design

The corpus study was complemented by a native speaker survey in order to address some of the issues that could not readily be controlled for in the corpus study. In order to verify that it is indeed aspectual choice (and not context) that triggers (inter)subjective meaning, we should verify whether speakers prefer a progressive to a non-progressive in contexts featuring a sense of irritation, surprise and the like. This can be achieved by constructing minimal pairs to be evaluated by native speakers with no prior exposure to the topic of study, which provides us not only with a way to control for contextual elements, but also with a more reliable comparative basis with the non-progressive than sampling non-progressives from a corpus could achieve.

The survey was created with the Qualtrics survey system and distributed by email and on social media in Flanders and in The Netherlands. The accompanying instructions explicitly call on native speakers of Dutch without a linguistic background (e.g. in the form of a degree in Literature and Linguistics) to participate in an online linguistic study conducted at the University of Antwerp that takes about 20 minutes to complete. Potential participants were referred to a homepage, where they could complete the survey. The introductory text included a confidentiality statement and provided more information about the survey setup, which would first ask the participants to answer some general questions (to obtain meta-linguistic information such as age, sex and region), and then to judge pairs of sentences. Participants were also informed that there would be a test round first (with two questions) so that they could get accustomed with the survey design before being presented with the actual test sentences.

The stimuli presented in the questionnaire are included in Appendix A. In each question, participants were provided with a minimal pair featuring a progressive and non-progressive form 9 in context and were asked to answer which of the two options sounds best. Four possible answers were provided:

  • A.

    Example sentence A (e.g. Dat meen je niet! Zat hij te flirten met meisjes uit jouw team?) 10

  • B.

    Example sentence B (e.g. Dat meen je niet! Flirtte hij met meisjes uit jouw team?)

  • C.

    Both A and B sound equally fine/bad and there is no difference in meaning

  • D.

    Both A and B sound equally fine/bad, but there is a difference in meaning

Subjects were asked to provide a brief explanation why their preferred option sounds better, or if A and B sound equally natural, but have a difference in meaning (option D), what the nature of this difference is. The survey progressed in a piece-meal fashion: participants were presented with only one question pair at a time, and could not go back to the previous questions (to make sure they rely on their first thoughts). The survey design also forced them not to leave the explanation field empty.

In addition to two test questions, participants were presented with 20 example pairs, 15 of which represented (inter)subjective contexts (of irritation, intensification, surprise and tentativeness). Importantly, these examples were not constructed, but are real-life examples attested in CGNa in progressive form, which were slightly adapted to make them more readable, and supplemented with their non-progressive counterpart to obtain a minimal pair. In all cases, the non-progressive was a grammatically and temporally (i.e. aspectually) acceptable alternative. The sample was internally balanced in the sense that more entrenched construction types (the aan het and zitten progressive), as evidenced by their frequency distribution in CGNa, were more often featured in the questionnaire; likewise, the total of complex progressives was restricted as simplex progressives are more commonly attested. An overview of the progressive types featuring in these 15 (inter)subjective contexts is presented in Table 3.

Table 3:

Types included in the survey, based on attested progressives in (inter)subjective contexts.

Five other question pairs were included which served as controls: (a) two constructed control sentence pairs, where one utterance each was clearly grammatically incorrect so as to ensure participants were attentive and had paid attention to these minimal differences (see Appendix Q03 and Q15), and (b), three pairs in non-subjective, generic or conditional contexts (cf. Appendix Q05, Q17 and Q20), which are not typically compatible with the progressive (in these cases, it is the non-progressive utterance that was attested in CGNa).

2.2.2 Results

In total, we collected 43 fully completed surveys, thereby reaching all age groups from adolescence onwards, both male and female speakers, and speakers from Flanders as well as the Netherlands (though the former are better represented). 11

In order to rank our results on a cline of perceived acceptability, the data were interpreted as a three-point Likert scale, with the various options coded as 0 (non-progressive form sounds most natural) – 0.5 (neutral) – 1 (progressive form sounds most natural). In case the D option was chosen, we assigned a 0, 0.5 or 1 score depending on the explanation given by the speaker. Only in instances where the speaker in her explanation explicitly mentions (inter)subjective notions such as surprise, disbelief, annoyance/disapproval, intensity etc. as being more highlighted in the utterance with the progressive, did the D option receive a score of 1. In cases of doubt or when no explanation was given, the neutral score (0.5) was assigned.

Computing the mean values of answers given will reflect the speakers’ preference for either progressive or non-progressive aspect. The closer the value is to 0, the stronger the preference for the non-progressive form; the closer to 1, the stronger the preference for a progressive form. Figure 2 plots the mean values per question as a proxy for the appropriateness of progressive vs. non-progressive aspect in subjective contexts (dark grey dots) and non-subjective generic control contexts (light grey dots). 12

Proportion of PROG choices per question (774 observations).
Figure 2:

Proportion of PROG choices per question (774 observations).

As expected, the non-subjective generic control contexts trigger non-progressive forms. Regarding subjective contexts, the chart shows an overall preference for the progressive form. In some questions, most notably the top five rows (Q08, Q18, Q10, Q02, Q01), this preference is strongly pronounced and statistically significant 13 if you take 0.5 (the neutral option) as the baseline of comparison, i.e. as the expected result. The remaining dark grey dots do not significantly differ from the 0.5 neutral option, 14 that is, their preference for the progressive (with the exception of Q06, Q11 – see infra) could be attributed to coincidence. It should, however, be noted that taking 0.5 (“both utterances sound equally natural”) as the expected mean is a very conservative measure if you consider that there is a strong frequency effect in favor of non-progressive aspect, even in contexts that encode ongoingness. Mortier (2008), for instance, concludes that in Dutch (as well as in French) simple tenses are the unmarked expression of progression. Fewer than 1% of all sentences in Dutch contain a progressive marker, 15 and even in the description of events that are “ongoing” or “in progression” the non-periphrastic simple tense is estimated to be 4 to 5 times as likely to be used as the progressive form (Mortier 2008: 5–6). Progressive constructions can therefore be said to occupy a rather marginal position in the grammar of Dutch. Given the fact that non-progressive aspect is the default in progressive contexts, we expect that the baseline for comparison of what is the “most appropriate” form in a particular context of ongoingness is in reality lower than the 0.5 mean we set as a baseline. This means that aspecto-temporal motivations for the use of progressive marking are low, 16 and that the observation that the progressive is favored in all but 2 of the 15 question pairs must be attributed to its affiliation with expressions of (inter)subjectivity.

There is one sentence in particular that seems to deviate from this pattern. Question 6, whose A and B answer options are repeated in (20a) and (20b), exhibits a statistically significant preference for the non-progressive form (20a):

ginderinhetMidden Oostenvechten?
over.thereintheMiddle Eastfight.prs.3pl
ginderinhetMidden Oostenliggentevechten?
over.thereintheMiddle Eastlie.prs.3pltofight.inf

‘We can’t help that they’re fighting over there in the Middle East.’

(CGNa fv400466)

From the explanations given by our informants, we can infer that this is largely due to the low grammaticalization status of the liggen progressive, especially for speakers of Netherlandic Dutch, who seem to require that the posture verb liggen ‘lie’ is semantically congruent with the event designated by the main verb. In the context of (20), one would not immediately conceive of the parties involved as lying down, which is why speakers of Netherlandic Dutch in particular and some speakers of Belgian Dutch as well prefer the non-progressive form (cf. also Q09 with liggen). For speakers who have adopted a more grammaticalized liggen progressive as a part of their idiolect, however, the progressive highlights the speaker’s frustration with the state of affairs. For instance, Belgian informants note that (20b), featuring the progressive form, “sounds more disapproving” than the non-progressive form (Sp14) and that it “implies the needlessness of the conflict” (Sp09).

Overall then, judgements by native speakers with no background in language and linguistics complement and corroborate our corpus findings. While it was established by means of the corpus study that progressive utterances frequently yield (inter)subjective readings, the survey results show that, on the whole, progressives are more compatible with (inter)subjective meaning elements than non-progressive simple tenses.

3 Discussion

3.1 General association of progressive aspect with (inter)subjective senses

The findings discussed in Section 2 confirm our hypothesis that the Dutch progressive constructions, like progressives in other languages, are highly liable to (inter)subjective interpretations, and that they differ in this respect from their simple counterparts, which are typically used for the expression of situations that are more readily assimilated in the speaker’s conception of reality. At first sight, an attractive way of analyzing this association of progressive aspect with (inter)subjective interpretations draws on the diachrony of the simple and progressive constructions, and the relative markedness of the latter. Studies focusing on the role of linguistic extravagance as an impetus for grammaticalization, such as Keller (1994), Haspelmath (1999) and Petré (2017), have demonstrated that individuals can select unconventional linguistic forms – such as a new construction, or a construction that is not normally used in a given context – in order to attract attention to what they are saying and thus to themselves. Linguistic change begins when this unexpected usage is adopted by peers wishing to identify with these individuals (cf. the rise in use of I kid you not in the past decades after it was introduced in a popular novel and adopted by Tonight television host Jack Paar), until the construction’s expressivity fades away because it is too frequently used in a given context to stand out. Petré (2017) convincingly demonstrates that extravagance played a crucial role in the extension of the Early Modern English progressive construction to present-tense main clauses. Early-stage present progressive attestations typically collocated with expressions of atypicality or irritation, and they differed in this respect from their simple present counterparts, which suggests that they were recruited by speakers as a means to stand out. Other developing constructions that appear to have been used for (inter)subjective purposes are GO-constructions: as noted in De Wit and Brisard (2017), the first auxiliary uses of English and French GO were attested in contexts featuring extravagant indices. Emerging perfect constructions, too, appear to lend themselves to extravagant exploitation (cf. e.g. Labov (1998) on the AAVE perfect). 17 Along the same line of analysis, we might hypothesize that the frequent association of the Dutch progressive with (inter)subjective readings is a result of the relatively recent emergence of the category (compared to the simple tense), and that the propensity of an individual Dutch progressive construction to taking on (inter)subjective interpretations is correlated with the relative degree of grammaticalization of that construction. This would certainly be worth investigating (see De Wit et al. ms), yet such an analysis in terms of extravagance, which essentially focuses on transient properties such as novelty and degree of grammaticalization, overlooks the synchronic stability and crosslinguistic dispersion of (inter)subjectively used progressives. How, for instance, can we account for the fact that the English present progressive still features (inter)subjective uses (see De Wit and Brisard 2014), in spite of its remarkable degree of grammaticalization? Similarly, why do we attest (inter)subjective uses for progressive constructions in genetically and geographically unrelated languages, irrespective of their degree of entrenchment in the verbal paradigm of those languages (see Section 1)? These observations suggest that an explanation for the association of the progressive with (inter)subjectivity needs to be sought in the semantics of the category. Building on De Wit et al. (2013), De Wit and Brisard (2014), Anthonissen et al. (2016), we will argue that this crucial semantic property is the fact that the progressive is an event selector. 18

The essential distinction between progressive and general imperfective constructions, such as the French imparfait, is that only the latter can combine with stative verbs; progressives combine with event verbs, by definition. Unlike events, states are unbounded and homogeneous (i.e. they do not change over time), and therefore contractible: any random sample of a state is representative of the state in its entirety. Events, on the other hand, are not contractible – they are construed as bounded in time and also, typically, heterogeneous. The function of the progressive is to imperfectivize events by imposing an internal perspective – an aspectual immediate scope (ISA) (Langacker 1987, Langacker 1991) – on these originally bounded and heterogeneous situations, as depicted in Figure 3.

Progressive aspect (based on Langacker 2001: 259).
Figure 3:

Progressive aspect (based on Langacker 2001: 259).

As a result of the internal viewpoint imposed by the progressive, the situation that is designated (i.e. that part of the situation that lies within ISA) is made unbounded and homogeneous, just like a state. Yet an important difference between genuine states and imperfectivized events is that, as can be inferred from Figure 3, the latter still feature backgrounded boundaries (which are located within the expression’s maximal scope or MS). This intrinsic association of progressive aspect with backgrounded boundaries has crucial epistemic implications: given the fact that events cannot be fully identified on the basis of one random sample (since they are not contractible), the partial view imposed by the progressive implies less than full knowledge about the further development and endpoint of the denoted event. Such bounded situations that cannot be fully identified by the speaker are necessarily contingent or incidental, i.e. they represent states of affairs that are not readily assimilable to known schemas. As De Wit and Brisard (2014: 68) put it, it seems that “the English progressive has inherited the epistemic and temporal traits associated with individual occurrences of events, the majority of which do not represent anything structural” (emphasis in the original). In our view, this analysis in terms of backgrounded boundaries and the epistemic implications thereof holds for progressive aspect in any language, including Dutch. The association of progressives with situations that are not (yet) assimilated within the speaker’s knowledge structure obviously makes these constructions excellent candidates for (inter)subjective extensions: while not every situation reported by means of a progressive necessarily entails (inter)subjective readings – it is not because a situation is not yet assimilated that it is also difficult to assimilate – every situation that is in some way perceived as non-canonical has not attained a structural status yet, and is grammatically marked accordingly by means of the progressive.

In what follows, we will offer an explanation for the differences between the various progressive constructions in terms of their relative susceptibility to (inter)subjective exploitation, thereby providing further evidence for the essential role of backgrounded boundaries in the epistemic status conferred upon situations reported by means of a progressive. In doing so, we will build on a number of previous proposals regarding the differences between the various Dutch progressives, which we therefore first introduce in the following section.

3.2 Differences between progressive constructions in Dutch: Previous observations

The various Dutch progressive constructions and the differences between them have been studied by, most notably, Beekhuizen (2010), Boogaart (1999), Booij (2008) and Lemmens (2005, 2012a, 2012b, 2015). Highly relevant for our analysis is Lemmens’s (2015) proposal to analyze the distinction between the aan het-progressive and the posture verb progressive in terms of different semantic profiles: the aan het-progressive is said to evoke a processual profile and as such “zoom[s] in on the ongoing process itself” (Lemmens 2015: 5), whereas the posture verb progressive construction has a situational profile, “typically locating the ongoing process in a well-described spatio-temporal frame” (Lemmens 2015: 5). The situational focus of the posture verb progressive can be regarded as a remnant of the original lexical (i.e. postural) use of these verbs. 19

These different semantic profiles for the two construction types are indeed reflected in many of the differences described in the literature so far. For one thing, although both construction types in principle select activity verbs and exclude state verbs (Boogaart 1999: 176; Lemmens 2015: 14), Lemmens (2015) notes – on the basis of a distinctive collostructional analysis of the relevant construction types – that the aan het-progressive clearly prefers telic (typically change of state or motion) verbs, whereas the posture verb progressive primarily selects non-telic verbs (see also Beekhuizen 2010: 111). The actual overlap of complement verbs appearing in both construction types is therefore rather small (about 10% of the verb types in Lemmens’s (2015) corpus occur in both construction types). Unsurprisingly, then, most distinctively attracted to the aan het-progressive are the change of state (hence processual) verbs worden ‘to become’ and veranderen ‘to change’, whereas the non-telic verb wachten ‘to wait’ is the verb most strongly attracted to the posture verb progressive (Lemmens 2015: 23). With regard to the types of subject both construction types prefer, Beekhuizen (2010: 107) notes that the likelihood of a posture verb progressive increases “if a subject is non-volitionally participating in the situation denoted by the main verb”, i.e. if the subject does not have prototypical agent properties. Again, this fits in with the situational profile of the posture progressive, which is more descriptive (and hence somewhat less agent-oriented) than the processual aan het-progressive. Another difference between the aan het-progressive and the posture verb progressive concerns the increased presence of locative modifiers with the latter (Lemmens 2005, Lemmens 2015; Beekhuizen 2010: 112), which is again consonant with the posture verb progressive’s situational focus. 20 As we will demonstrate in Section 3.3, this difference between the more “processual” prepositional progressive and the more “situational” posture verb construction can be held responsible for the different frequency of intersubjective readings attested for the respective constructions.

As to the (relatively infrequent) motion verb progressive with lopen, Lemmens (2005) notes that it typically combines with verbs denoting physical motion like voetballen ‘playing soccer’ and as such has a more dynamic meaning that the posture verb progressives. When lopen selects non-motion verbs, it “typically expresses repetition, often combined with a verb that displays a negative evaluation on the speaker’s part” (Beekhuizen 2010: 106). Previous observations thus already point into the direction of a connection between the repetitive profile of lopen and a sense of irritation – a connection that we will further elaborate in the subsequent section.

3.3 Why some progressive constructions are more prone to (inter)subjective readings than others

In Section 3.1, we have explained that every progressive construction constitutes, in principle, an excellent device to report situations that are in some way non-canonical, given the heterogeneity and backgrounded boundaries (reflecting a non-consolidated status) by definition present in the semantic configuration of the progressive. Still, prepositional constructions are clearly less prone to (inter)subjective interpretations than posture verb constructions, which are in turn less frequently associated with (inter)subjectivity than the motion verb construction. In order to account for these differences, we have to connect the previous observations on these constructions’ semantic profiles with our analysis presented in Section 3.1: the central idea is that the more the progressive accentuates the boundaries or repetition of the denoted situation, the more likely it is to receive (inter)subjective interpretations.

Let us first zoom in on the differences between prepositional and posture verb constructions. While Figure 3 is applicable to both types of progressive, that part of the event that is backgrounded in the expression’s maximal scope is more highlighted in the case of the posture verb construction, since the posture verb construction is generally more concerned with where and how something happened rather than with the ongoingness of a process. This is depicted in Figure 4, which can be considered an extension (i.e. slight modification) of Figure 3:

Posture verb progressives.
Figure 4:

Posture verb progressives.

While the designated part of the event is still its imperfectivized part within ISA, the higher degree of boldness of the backgrounded part of the event (compared to Figure 3) reflects that it has a more important contribution to the semantics of the expression. Since the presence of a sense of (inter)subjectivity can be connected to the association of progressive aspect with non-consolidated (i.e. bounded and (typically) heterogeneous) situations, and since posture verb progressives give emphasis to these (normally more) backgrounded elements, it should come as no surprise that these constructions are more frequently associated with (inter)subjective readings than their prepositional counterparts, which are not specifically concerned with the non-structural properties (the “here-and-now”) of the events they report.

The semantic configuration of the motion verb progressive can equally be analyzed as an extension of Figure 3, yet this time it is not so much the original boundaries (and possible heterogeneity) that are highlighted, but rather the duration and iteration of the ongoing event. The difference between (21a) and (21b), for instance, is that in the latter case the situation is perceived as occurring repeatedly within the same temporal interval:


‘John is whining.’


‘John is constantly/continually whining.’

Thus, the configuration for the motion verb progressive can be rendered as in Figure 5, the difference with Figure 3 being the fact that the progressive zooms in on a kind of virtual generalization that the speaker makes on the basis of actual repeated occurrences. That is, the iterated event (the whining in [21b]) is represented as continually present at reference time.

Motion verb progressive.
Figure 5:

Motion verb progressive.

The fact that lopen te is used to generate this iterative/durative effect can be analyzed a result of its semantics. A motion verb does not only entail a relatively extended process, but also a cyclical one: running/walking requires the continued repetition of the same type of movement. As a semi-serial verb (i.e. in collocation with another verb V), lopen te has the meaning of running/walking while doing something else (expressed by V). By extension, V takes over the sense of continued repetition conveyed by lopen te. Clearly, events that are constantly repeated within a single temporal interval (to the point that they are perceived as continually ongoing) are likely to cause irritation – hence the frequent exploitation of lopen te to express some kind of annoyance on the part of the speaker or subject. In view of this plausible semantic connection between motion, temporal extension and irritation, it should come as no surprise that motion verb progressives in other (European) languages are typically associated with a durative meaning as well (Bertinetto et al. 2000: 523), and that they often give rise to negative readings (Ebert 2000: 616; Tommola 2000: 664).

4 (Inter)subjective reinforcement in complex progressives

Table 2 shows that so-called complex progressives are particularly prone to (inter)subjective interpretations. To repeat, we have coined the term complex progressives to refer to those cases in which the progressive construction is embedded in another construction, whereby the progressive auxiliary (be, the posture verb or the motion verb) is non-finite; this holds for progressive constructions in the (plu)perfect, for cases featuring another (semi-)auxiliary verb, such as gaan ‘go’, a modal or a combination of the two, and for those cases in which no finite auxiliary is present (the so-called root or bare infinitive progressive). We have also observed that these constructions are largely ignored in the extant literature, in spite of their noteworthy semantic and pragmatic features, i.e. their propensity to (inter)subjective extensions. Table 4 gives an overview of the absolute and relative frequency of (inter)subjective interpretations per type of complex progressive.

Table 4:

Complex progressives.

Table 4 shows that of the different types of complex progressives, the perfect of the prepositional progressive is least prone to (inter)subjective interpretations. Since it appears to behave like non-perfect progressives in terms of its inclination towards (inter)subjective readings, we will not devote any separate discussion to this construction. The remaining complex progressives, on the other hand, are very likely to feature (inter)subjective overtones – an observation that we will try to account for in this section, thereby distinguishing between the perfect posture/motion progressive (Section 4.1), the “gaan/komen + posture/motion progressive” (Section 4.2), “modals + posture/motion progressive” (Section 4.3), and posture/motion verb progressives without auxiliary (Section 4.4), since – as we will see – each involves a different theoretical motivation for their association with (inter)subjectivity.

4.1 The perfect posture/motion verb progressive and (inter)subjectivity

Examples (22)–(23) illustrate some (inter)subjective uses of perfect progressives with posture/motion verbs in Dutch:

itbe.sorry.prs.3sgmevery muchyesthatIhave.prs.1sg

‘I’m really sorry that I’ve been raging on.’

(CGNa fv400421)

‘Then my father had still been making remarks about my clothes, about how stupid he thought they were.’
(CGNa fn007972)

‘She had of course been messing about and she hadn’t had the time anymore to put it back in the right spot where I had put it down.’

(CGNa fn008413)

While (22) involves a sense of intensity, the speakers in (23) and (24) are obviously irritated. We have seen that each of these readings can also be conveyed by simplex posture/motion verb progressives, yet as the frequencies in Table 4 show, these senses of intensity and irritation are reinforced when the posture/motion progressive is embedded in a perfect construction.

Note that the English translations of these examples also feature (present and past) perfect progressives, which preserve the (inter)subjective meanings conveyed by their Dutch counterparts. In fact, many Dutch examples are reminiscent of the so-called “explanatory-resultative” uses of the (present) perfect progressive in English (Declerck et al. 2006: 253–254), as illustrated in (25):


You’re behaving so awkwardly – have you been drinking again?

In cases such as this, the present perfect progressive is used to highlight the unintended side-effects of a past event, which typically comes with a sense of surprise or irritation on the part of the speaker. Thus, not only in Dutch, but also in English, do we find that the combination of perfect and progressive – or at least, one type of progressive in the case of Dutch – can generate (inter)subjective meaning effects, which suggests that there is something about these two types of aspect that makes their combination inherently susceptible to such interpretations. Yet, to our knowledge, no plausible explanation for this observation has been offered for either language.

We argue that the perfect can reinforce the (inter)subjective quality that is already intrinsic to progressive constructions because it is itself semantically associated with (inter)subjectivity. In other words, perfect progressives are complex combinations of two constructions that often receive (inter)subjective readings across languages. For the progressive, we have offered evidence and a theoretical explanation for this connection with (inter)subjectivity, and we have shown that the posture/motion verb progressives are even more likely to receive such readings than their prepositional counterparts. The perfect, in turn, has been singled out as potentially signaling incongruity in a variety of languages by, among others, DeLancey (2001) and De Wit (2017b). For instance, as pointed out by Slobin and Aksu (1982: 195) in their analysis of the Turkish -miş perfect, “the essence of all uses of –miş is to encode situations for which the speaker is not somehow prepared – situations on the fringe of consciousness, learned of indirectly, or not immediately assimilable to the mental sets of the moment”. In his crosslinguistic analysis of mirativity (i.e. the grammatical expression of surprise), DeLancey (2001: 378) proposes a conceivable semantic explanation as to why perfect constructions lend themselves to the expression of incidental situations for which the speaker lacks premonitory awareness: since (present) perfect aspect involves past situations of which the speaker witnesses some current consequences (in the form of, e.g. a result), the information about the past event which the speaker has at the time of speaking is necessarily indirect. As DeLancey (2001: 378) notes, “a fact which one knows only when ones sees secondary evidence for it is necessarily unexpected to some degree.” If the speaker did witness a past event directly, it is no longer particularly novel at the time she reports this event. It thus seems that both the perfect and the progressive have inherent properties that make them suitable for the expression of situations that are in a way incongruous or non-canonical. It should not be surprising, then, that the combination of the two yields (inter)subjective readings even more frequently. We can once more propose a diagram that reflects the semantic origin of the (inter)subjective present perfect progressive. Figure 6, which is based on Brisard (2013: 232), contains the representation of a regular present perfect in English.

English present perfect.
Figure 6:

English present perfect.

In contrast to the previous diagrams, Figure 6 does not only contain an aspectual immediate scope (ISA), but also a temporal immediate scope (IST), which is imposed by the tense marker. Since the present perfect is a present-tense construction (in English at least – cf. below for Dutch), this immediate scope coincides with the time of speaking, indicated by the boxed squiggly line (cf. De Wit 2017a: Ch. 2 for more details). Aspectually, the perfect zooms in on the (resultant) end-state of a past situation (since it deals with the current consequences of that situation), and thus IST and ISA coincide. This past situation (an event in the case of Figure 6) is therefore not in focus, but it does belong to the overall conceptualization of the present perfect, represented by its maximal scope (MS). The difference between a regular present perfect and a present perfect progressive resides in the nature of this past situation in MS: while in the former case we are dealing with a simple state or event, the latter involves an event that has undergone imperfectivization by means of the posture verb progressive before being taken up in the configuration of the perfect. The present perfect progressive can thus be rendered as suggested in Figure 7, which connects parts of Figure 4 with parts of Figure 6 by means of interrupted correspondence lines.

Present perfect progressive.
Figure 7:

Present perfect progressive.

Figure 7 shows that, in the case of the present perfect progressive, the past situation of which the consequences continue into the present is actually an internal sample of an event that is conceived of as phenomenal (rather than structural), by virtue of the backgrounded boundaries in MS1.

One might object that Figure 6, and therefore also Figure 7, are relevant for English, but not so much for Dutch, in which the original present perfect has evolved into a construction that comprises both past perfective and present perfect readings. Thus, situations that took place at a definite moment in the past and that are not connected to the present can be reported by means of a present perfect (Brisard 2013: 233–234), as is illustrated in (26):


‘[lit.] The cats have caught two mice last week.’

Since the connecting line between the past situation and the time of speaking in Figure 6 need not be present in the case of the Dutch present perfect, and since there is therefore not necessarily a relation of secondary evidence, one might wonder to what extent the analysis of the Dutch present perfect as a construction prone to (inter)subjective exploitation still holds. Yet the current range of uses of the Dutch present perfect (progressive) does not, we believe, determine the relative bias of the construction toward (inter)subjectivity. Most likely this bias is inherited from older varieties of the construction, which did only feature true present-tense uses, establishing a sense of current relevance. We have no reason to believe that the fact that the Dutch perfect can now take on past perfective meanings on top of its present perfect meaning has fundamentally altered the kind of (inter)subjective extensions it can give rise to.

4.2 Gaan ‘go’ and komen ‘come’ + posture/motion verb progressives

Complex constructions in which posture/motion verb progressives combine with the auxiliaries gaan ‘go’ or komen ‘come’ also turn out to be particularly likely to receive an (inter)subjective interpretation, as can be inferred from Table 4. The combination of gaan with posture/motion verb progressives is notably more frequent in our corpus than combinations with komen. Yet as the following examples show, both can (and often do) receive (inter)subjective readings:


‘Surely you don’t mash that.’

(CGNa fv400623)

‘It was the middle of winter and I was scraping off my car. And then he came like helping me just like that. I thought that was really quite kind.’
(CGNa fv400197)

While (27) involves a sense of atypicality, the speaker in (28) is obviously surprised. In both cases, these (inter)subjective readings would be less straightforward if the simple present or past were used. In line with our analysis of the present perfect progressive in Section 4.1, we believe that this strong sense of (inter)subjectivity is the product of the combination of two constructions that are each semantically affiliated to non-canonicity. We have amply shown that this is the case for posture/motion verb progressives as a result of the relative profiling of the events they imperfectivize (in the case of posture verb progressives) or the iteration and stretching of an event (in the case of the motion verb progressive). A sense of emphasized boundaries is, as we will demonstrate now, also the reason why gaan and komen are so likely to reinforce the mirative reading of posture/motion verb progressives.

Analyses of a wide range of languages have pointed out that the motion verbs COME and GO are likely to grammaticalize into tense and aspect auxiliaries (cf. Bybee et al. 1994, among many others). 21 Thus, as an auxiliary, COME often conveys a recent past or resultative meaning, whereas the auxiliary GO typically appears in future or inchoative contexts. Yet both verbs can also be recruited for other grammatical functions, as has been extensively described in, most notably, Devos and van der Wal (2014). The uses that are of interest here have been most extensively (though not exclusively) analyzed in studies on GO and COME in French and English, in which they were described as “evaluative” (Bourdin 2003), “extraordinary” (Bres and Labeau 2013) or “mirative” (Celle and Lansari 2015) – terms that are very similar to the notion of (inter)subjectivity as defined in this paper. As Examples (27) and (28) show, Dutch komen and gaan can equally be used to confer an (inter)subjective evaluation on the part of the speaker. 22 The crosslinguistic attestation of (inter)subjective readings (see especially Bourdin (2003) on languages other than French and English) suggests that there is something intrinsic about COME and GO constructions that makes them prone to (inter)subjective exploitation. Celle and Lansari (2015) point out that French aller, which acts as an inchoative/future marker, but can also indicate surprise on the part of the speaker, puts a focus on the initial boundary of a process. This property, which we believe is typical of grammaticalized GO constructions across languages, makes these constructions likely to take on (inter)subjective interpretations as a consequence of the relationship we posit between the presence and relative prominence of boundaries and (inter)subjectivity. When the Dutch gaan construction combines with posture verb progressives, the initial backgrounded boundary of the imperfectivized event, which is already quite highlighted by means of the posture verb construction (which gives relative emphasis to both boundaries), is even more highlighted; we thus find double (inter)subjective strengthening, as is depicted in Figure 8. Similarly, with the motion verb progressive, gaan reinforces the designated event’s contingent status on top of the sense of duration/iteration that the “lopen te + V” construction conveys – see Figure 9.

gaan + posture verb progressive.
Figure 8:

gaan + posture verb progressive.

gaan + motion verb progressive.
Figure 9:

gaan + motion verb progressive.

For komen we can propose a similar analysis, which is again inspired by Celle and Lansari (2015), who tentatively suggest that COME constructions are used to highlight the transition that takes place within the conceptualizer’s deictic center as a result of a process that intrudes into this deictic center (Celle and Lansari 2015: 298). In this case, it is not the initial transition, but the final one that is highlighted: COME constructions indicate that the occurrence and completion of an event causes the start of something new – a situation that is possibly discrepant with what is expected on the basis of what used to be the case. This property of COME constructions does not only make them likely to evolve into resultative markers, it also turns them into likely markers of non-canonicity. Again, in combination with posture and motion verb progressives, this (inter)subjective reading that is already inherent in komen is reinforced. The diagrammatic representations of “komen + posture/motion verb progressives” are equivalent to Figures 8 and 9, the only difference being that the emphasis now lies on the final boundary.

4.3 Modals + posture/motion verb progressives

The combination of a modal verb with a posture or motion progressive also enhances the (inter)subjective potential of an utterance, as the following examples show.


‘But then she also shouldn’t be complaining about rheumatism and such things.’

(CGNa fn007972)


‘But I can’t chatter with you all the time, you know.’

(CGNa fn000621)

In (29) the speaker voices his or her irritation about somebody complaining about her health, whereas example (30) involves a sense of intensity (also marked by the adverbial de hele tijd ‘the whole time’ and the inherent iterativity in the verb kleppen ‘chatter’). Note that both examples also feature the verb gaan ‘go’ as an extra (inter)subjective re-enforcer (see Section 4.2); in fact, in almost half of the cases in which a modal verb is present, it is combined with gaan in our data sample (see Table 4).

The modal verbs appearing here are mainly moeten ‘must’ and kunnen ‘can’, typically with a deontic meaning, i.e. referring to societal norms or more speaker-oriented assessments in terms of (moral) acceptability, as in the examples above. Dynamic meanings, involving particular subject- or situation-internal capacities and necessities, show up as well, as in the following examples. In (31), the speaker expresses her irritation at the behavior of the subject who cannot but weep, whereas in (32), the speaker is clearly irritated about the fact that one has to, i.e. is forced by particular circumstances, to elbow one’s way through.


‘Who only sits there and weeps.’

(CGNa fn000664)


‘Then you’ve got to elbow your way through, you have to hold on to those dirty balusters.’

(CGNa fv400615)

In (33), which features the modal verb mogen ‘to be allowed to’, the speaker expresses irritation at the fact that everybody is allowed to talk bullshit on the phone in the library (whereas smoking is forbidden everywhere).

youmay.prs.2sgGod damn itnowhereacigarettelight.inf
‘You’re not allowed – damn it – to light a cigarette anywhere but everybody can sit in the library with his phone and talk bullshit but there’s absolutely nothing being done about it.’
(CGNa fn000665)

Just like progressives, dynamic and deontic modal verbs are often used to construe a current situation as incongruous with the way the speaker wants or expects it to be. In a force-dynamic analysis of modals (see Talmy 1988), subjects of modal verbs are conceptualized as being under the influence of particular forces clearly delimiting the subject’s potential range of action, which might lead to irritation. This is obviously the case with modals denoting obligation and necessity, such as the modal moeten ‘must, have to’ in (34): the necessity of having to go to work does not align with what the speaker thinks should be the case (i.e. staying at home). 23

‘Help! My child is ill and I have to go working.’

The same applies to modals expressing ability and permission. A good example of the latter case is (35), in which dynamic kunnen ‘can’ also expresses the speaker’s irritation at her having to arrange everything.


‘Your granddad is giving a party tonight. And I can make all the preparations.’

It should in this connection also be emphasized that the modal verbs in complex progressive constructions with intersubjective meanings do not seem to feature epistemic meanings, but only dynamic and deontic ones, which typically select eventive (and thus bounded) predicates allowing a transition from a stage in which the state of affairs is not realized to one in which it is (see Heine 1995; Abraham 1998 on the eventive nature of verbs combined with non-epistemic modals). Again, it seems that both components of the complex construction – deontic/dynamic modals and posture/motion verbs – share a propensity for (inter)subjective meanings.

4.4 Posture/motion verb progressives without auxiliary

The use of bare infinitives where one would expect a finite verb form has mostly attracted attention in the domain of language acquisition (see, e.g. Blom 2007 on Dutch and English, among many others). It is thus primarily considered a reflection of a stage in an incomplete learning process. Yet adult speakers too may resort to bare infinitives in main-clause contexts (see Nikolaeva 2014 on so-called root infinitives in French and Latin; Avrutin 1999: Ch. 8 on Russian and English; Lasser 2002 on root infinitives in German; IJbema 2002: 118–122 on root infinitives in Dutch, among others). The examples in (36)–(38) illustrate the use of posture/motion verb progressives in Dutch without a finite auxiliary. Note that, in terms of semantics, these sentences cannot be interpreted as non-finite: they report past perfective events and would therefore normally feature the present-perfect auxiliary hebben ‘have’ (recall that the present perfect has evolved into a past perfective in Dutch – cf. Section 4.1).


‘The teacher going outside to watch. Like, yes you have been drinking. The whole place upside down. All those classes brought in. All those book bags searched through.’

(CGNa fv400495)


‘Everything was carried into that cave. Crawling all the way under a hole where you could barely go through, crawling on your knees.’

(CGNa fn007879)


‘But dad and Bert, they have also called. I find that strange as well, I hadn’t expected that at all. Both calling with their mobile phones.’

(CGNa fn000583)

Each of these examples illustrates the association of progressive infinitives without an auxiliary with a sense of incongruity – an association that is relatively strong, as Table 4 shows. Both (36) and (37) feature a sense of intensity. In (38), the speaker obviously depicts the reported situation as atypical. Observe that, in this case, the progressive infinitive combines with gaan ‘go’ (cf. Section 4.2), which thus reinforces the (inter)subjective reading of (38). In fact, this example constitutes, we believe, a case of double reinforcement of the posture/motion verb progressive: one by gaan and one by the bare infinitive progressive. In other words, just like with the perfect, gaan/komen and the dynamic/deontic modals, we argue that bare infinitive progressives instantiate a complex construction that combines two constructions that are inherently liable to (inter)subjective manipulation, in this case the posture/motion verb progressive and the root infinitive construction. Thus, instead of assuming that uses of bare infinitives in adult language constitute illustrations of some random auxiliary ellipsis in spoken Dutch, we argue that utterances of sentences feature bare infinitives for specific semantic reasons. Consider the following bare infinitive constructions: 24


[In response to someone who has been complaining about elaborate renovations in her house:]


‘And me worrying about a donkey that’s a little bit upset.’


[In a group therapy session, one of the participants complains about three other participants that do not cooperate:]


‘And we’re doing nothing but working and we’re doing nothing but coming out of our shell. and they, they just keep sitting where they’re sitting and you think that’s fine too.’


[About someone who wants to distance herself from her father:]


‘And I kept saying: “No, I don’t have anything to do with him.” But it kept following me.’


‘Him clean the house? No way!’

Examples (39)–(41) instantiate what we suggest to call the “en SUBJ maar INF Construction”. Example (39) illustrates the use of bare infinitives to report repeated useless actions on the part of the subject: actions that should not have taken place. Both (40) and (41) involve intensified events, which are in addition irritating to the speaker. Clearly, each of these contexts involve events that in a way diverge from what is considered canonical, and this sense of non-canonicity would not be obviously present if the root infinitives were to be replaced by means of finite constructions. In (42), the speaker is using another construction, the so-called “Incredulity Response Construction” (Fillmore et al. 1988: 511) involving “Mad Magazine sentences” (Akmajian 1984), which has been analyzed at length for English by, among others, Lambrecht (1990). The “Incredulity Response Construction” is typically used for the expression of “surprise, disbelief, skepticism, scorn, and so on” (Akmajian 1984, cited in Lambrecht 1990: 215) in English. As Example (42) illustrates, such uses are also attested in Dutch: in the example, the speaker expresses her surprise at the suggestion that the subject would be cleaning (or would have cleaned) the house.

We contend that the “en SUBJ maar INF Construction” and the “Incredulity Response Construction” do not accidentally feature bare infinitives in Dutch, but that this form is specifically chosen for (inter)subjective purposes. As Lambrecht (1990: 224) points out in his analysis of “Incredulity Response Constructions”, the speaker’s “incredulous response does not concern the time at which the situation or event described in the context proposition takes place”. Rather, it refers to the general awkwardness of the hypothetical occurrence of a situation within a certain context, involving a certain subject. Building on Langacker (1987), Szcześniak and Pachoł (2015: 10) similarly argue that non-finite verbs profile “concepts considered in the abstract rather than realized events”. Since the “Incredulity Response Construction” involves connections between a subject and a situation that are not normally/actually connected, it makes sense to use non-finite verbs which can represent that situation as abstract and hypothetical. This analysis may be extended to account for the relation between incongruity and bare infinitives in general: situations that do not normally take place (or do not take place to that extent) in the world of the conceptualizer (i.e. situations that are typically presented by means of constructions that have taken on (inter)subjective meanings) are more likely to be represented in an a-temporal fashion, e.g. by means of infinitives.

Even if this hypothesis requires further verification on the basis of empirical data, we hope to have shown in this section that, in those contexts in which bare infinitives are used to express situations that might have been expressed by means of finite verbs as well, we appear to end up with (inter)subjective readings. Thus, root infinitives seem to be intrinsically connected to a sense of incongruity, and they can therefore reinforce the (inter)subjective meanings that are conveyed by means of posture/motion verb progressives in Dutch.

5 Conclusion

This paper set out to demonstrate that Dutch progressive constructions can be exploited to present situations as in a way incongruous with the speaker’s conception of how reality is or should be. While the aspectual meaning of the various progressive constructions in Dutch has received considerable attention in the last few years, much less attention has been paid to their non-aspectual uses, in particular their potential to express (inter)subjective meanings such as surprise, irritation and intensity, which, in the attested examples, often seem more prominent than the purely aspectual meaning of ongoingness or duration that is usually implied. Such (inter)subjective extensions of progressive constructions have been observed in languages other than Dutch, both European and non-European. What makes the situation in Dutch particularly intriguing, however, is the fact that there are at least five more or less grammaticalized progressive constructions, which exhibit a different (inter)subjectivity potential. On the basis of our corpus study, we have shown that all the Dutch progressive constructions under investigation can be used to convey a sense of non-canonicity, and we have accounted for this relation between progressivity and non-canonicity in terms of the backgrounded boundaries that are by definition part and parcel of the semantic configuration of the progressive, which thus presents situations as non-structural or phenomenal within the speaker’s conception of reality. Yet, of these different progressive constructions in Dutch, the motion verb progressive turns out to trigger (inter)subjective readings more frequently than the posture verb constructions, which are in turn more prone to (inter)subjectivity than the prepositional construction. We have connected these results to the different semantic configurations of the progressive construction classes, arguing that the more the progressive accentuates the boundaries or repetitive character of the denoted event, the more likely it is to receive an (inter)subjective interpretation. By way of concluding our analysis, we have discussed four complex progressive constructions: the present perfect posture/motion verb progressive, “gaan/komen + posture/motion verb progressive”, “modals + posture/motion verb progressive” and bare infinitival uses of posture/motion verb progressives. In each of these complex constructions, the progressive combines with another construction that is also inherently capable of undergoing (inter)subjective meaning extensions. We are thus dealing with cases of (inter)subjective reinforcement, and consequently, a very strong connection between these complex progressives and interpretations of non-canonicity.

More generally, our analysis demonstrates that speakers can adapt the prototypical affordances of grammatical constructions for affective purposes, i.e. to qualify a situation and/or to generate a certain effect on the part of the addressee. There is no reason to assume that this important evaluative function of grammar is restricted to the domain of aspect. Future work will have to unveil whether systematic (inter)subjective exploitation can also be attested beyond the verb phrase, and if so, how such notable uses can be integrated in a semantic analysis of the constructions at hand.


We wish to thank two anonymous reviewers for their useful comments on an earlier version of this paper.

Appendix A: Native speaker survey

The following stimuli were presented via online survey tool Qualtrics. For reasons of clarity, the minimal pairs are printed in bold and accompanied with an English translation (which was naturally not the case in the online survey version). More information about the survey design can be found in Section 2.2.1.


A. Toen gingen al die taxi’s weer rijden want die hadden daar super lang gewacht.

‘Then all the taxis went driving again because they had waited [no progressive] there for a very long time.’

B. Toen gingen al die taxi’s weer rijden want die hadden daar super lang staan wachten.

‘Then all the taxis went driving again because they had been waiting [lit. had stand wait; complex posture progressive with staan ‘stand’] there for a very long time.’


A. Ik denk… Je werkt voor ’t volk, je verdient een salaris van hier tot Tokyo. Ga je nog zo lopen kafferen.

‘I think… You work for the people, you earn an amazing lot of money. And still you are railing [lit. go run rail; complex motion progressive with gaan ‘go’ and lopen ‘run’].’

B. Ik denk… Je werkt voor ’t volk, je verdient een salaris van hier tot Tokyo. Ga je nog zo kafferen.

‘I think… You work for the people, you earn an amazing lot of money. And still you rail [no progressive].’


A. Ik ben vergeten de afwas te doen.

‘I have forgotten to do [no progressive] the dishes.’

B. Ik ben vergeten de afwas aan het doen.

‘I have forgotten to be doing [ungrammatical] the dishes.’


A. Dat meen je niet! Zat hij te flirten met meisjes uit jouw team?

‘You must be kidding. Was he flirting [lit. sat he to flirt; simplex posture progressive with zitten ‘sit’] with girls on your team?’

B. Dat meen je niet! Flirtte hij met meisjes uit jouw team?

‘You must be kidding. Did he flirt [no progressive] with girls on your team?’


A. Zit u over het algemeen veel TV te kijken?

‘Are you generally watching [lit. sit you to watch; simplex posture progressive with zitten ‘sit’] a lot of TV?’

B. Kijkt u over het algemeen veel TV?

‘Do you generally watch [no progressive] a lot of TV?’


A. Wat kunnen wij er hier aan doen dat ze ginder in het Midden Oosten vechten?

‘What can we here do about the fact they fight [no progressive] over there in the Middle East?’

B. Wat kunnen wij er hier aan doen dat ze ginder in het Midden Oosten liggen te vechten?

‘What can we here do about the fact that they are fighting [lit. lie to fight; simplex posture progressive with liggen ‘lie’] over there in the Middle East?’


A. Ik vind het wel vervelend dat die daar zo echt op kan kijken.

‘I find it a bit annoying that he can really look [no progressive] at it like that.’

B. Ik vind het wel vervelend dat die daar zo echt op kan zitten kijken.

‘I find it a bit annoying that he can be looking [lit. ‘can sit look’, complex posture progressive with modal verb kan ‘can’ and posture verb zitten ‘sit’] at it like that.’


A. Maar ik denk even of ik ’s middags niet ook naar huis ga omdat ik tot half één les heb.

‘But I just think [no progressive] whether I won’t go home at noon as well since I have classes until 12:30.’

B. Maar ik zit even te denken of ik ’s middags niet ook naar huis ga omdat ik tot half één les heb.

‘But I am just thinking [lit. sit to think; simplex posture progressive with zitten ‘sit’] whether I won’t go home at noon as well since I have classes until 12:30.’


A. Het spijt mij dat ik heb liggen doorrazen.

‘I’m sorry that I have been raging on [lit. have lie rage; complex posture progressive with liggen ‘lie’].’

B. Het spijt mij dat ik heb doorgeraasd.

‘I’m sorry that I have raged on [no progressive].’


A. Dus ik was aan het werken en die staat daar zo.

‘So I was working and he just stands [no progressive] there like that.’

B. Dus ik was aan het werken en die staat daar zo te staan.

‘So I was working and he is just standing [lit. stands to stand; simplex posture progressive with staan ‘stand’] there like that.’


A. En Bart was maar aan ’t zeggen van ‘Haast u, haast u!’

‘And Bart was constantly saying [simplex prepositional progressive] like ‘Hurry up, hurry up!’

B. En Bart zei maar van ‘Haast u, haast u!’

‘And Bart constantly said [no progressive] ‘Hurry up, hurry up!’


A. Ik keek daar zo van “Oh fuck – wat is hier allemaal gebeurd.”

‘I looked there like “Oh fuck – what on earth has happened here.”’

B. Ik zat daar zo te kijken van “Oh fuck – wat is hier allemaal gebeurd.”

‘I was looking [lit. sat to look; simplex posture progressive with zitten ‘sit’] there like “Oh fuck – what on earth has happened here.”’


A. Maar jullie willen echt gaan zeilen weer? – Ja, nou, mogelijk in ieder geval. We zijn nog een beetje aan ’t brainstormen.

‘But you really want to go sailing again? – Yes, well possibly in any case. We are still brainstorming [simplex prepositional progressive] a little’

A. Maar jullie willen echt gaan zeilen weer? – Ja, nou, mogelijk in ieder geval. We brainstormen nog een beetje.

‘But do you really want to go sailing again? – Yes, well possibly in any case. We brainstorm [no progressive] a little’


A. We hadden een pot op tafel liggen, en ineens was die pot weg. Echt superlullig. We hadden wel een beetje een vermoeden wie het gedaan zou kunnen hebben, maar ja, we hadden niets gezien en… Echt superlullig, dan ben je gewoon je maten aan ‘t naaien

‘We had a money pot lying on the table, and all of a sudden it was gone. Really very annoying. We had a bit of a suspicion about who could have done it, but well, we hadn’t seen anything and… Really very annoying, then you are really deceiving [simplex prepositional progressive] your mates…’

B. We hadden een pot op tafel liggen, en ineens was die pot weg. Echt superlullig. We hadden wel een beetje een vermoeden wie het gedaan zou kunnen hebben, maar ja, we hadden niets gezien en… Echt superlullig, dan naai je gewoon je maten…

‘We had a money pot lying on the table, and all of a sudden it was gone. Really very annoying. We had a bit of a suspicion about who could have done it, but well, we hadn’t seen anything and… Really very annoying, then you really deceive [no progressive] your mates…’


A. Gisteren zal ik zitten eten met mijn vrienden.

‘Yesterday I will be eating [ungrammatical] with my friends.’

B. Gisteren zat ik te eten met mijn vrienden.

‘Yesterday I was eating [lit. I sat to eat; simplex posture progressive] with my friends.’


A. Als ik nou ineens zo ga doen dan heeft mijn moeder ook wel door dat ik slijm.

‘If I start behaving like that all of a sudden, my mother will also know that I brown-nose [no progressive] her’

B. Als ik nou ineens zo ga doen dan heeft mijn moeder ook wel door dat ik loop te slijmen.

‘If I start behaving like that all of a sudden, my mother will also know that I am brown-nosing [lit. run to brown-nose; simplex motion progressive] her.’


A. Oké, en wie belt wie? – Degene die wakker is, die belt gewoon.

‘Okay, and who calls who? – The person who is awake, simply calls [no progressive]’

B. Oké, en wie belt wie? – Degene die wakker is, die ligt gewoon te bellen.

‘Okay, and who calls who? – The person who is awake, is simply calling [lit. lies to call; simplex posture progressive with liggen ‘lie’].’


A. Daar trok ik me dan liever aan op dan dat ik ging denken van “Hé, wat storend. D’r zijn mensen aan het lullen” of zo.

‘I preferred to draw support from that rather than thinking like “Hey, this is so annoying. People are talking bullshit [simplex prepositional progressive]” or something like that.’

B. Daar trok ik me dan liever aan op dan dat ik ging denken van “Hé, wat storend. D’r lullen mensen” of zo.

‘I preferred to draw support from that rather than thinking like “Hey, this is so annoying. People talk bullshit [no progressive]” or something like that.’


A. Ik heb echt thuis al in de kookboeken gesnuffeld.

‘I really have browsed [no progressive] through the cookbooks at home.’

B. Ik ben echt thuis al in de kookboeken aan het snuffelen geweest.

‘I really have been browsing [complex prepositional progressive] through the cookbooks at home.’


A. Als hij ergens mee zit, dan is hij me aan het opbellen.

‘If he worries about something, then he will be calling (lit. is he calling me; simplex prepositional progressive) me.’

B. Als hij ergens mee zit, dan belt hij me op.

‘If he worries about something, then he’ll call [no progressive] me.’


  • Abraham, Werner. 1998. The morphological and semantic classification of ‘evidentials’ and modal verbs in German: The perfect(ive) catalyst. Groninger Arbeiten zur germanistischen Linguistik 42. 192–206. Google Scholar

  • Akmajian, Adrian. 1984. Sentence types and the form-function fit. Natural Language & Linguistic Theory 2(1). 1–23. Google Scholar

  • Anthonissen, Lynn, Astrid De Wit & Tanja Mortelmans. 2016. Aspect meets modality: A semantic analysis of the German am-progressive. Journal of Germanic Linguistics 28(1). 1–30. CrossrefGoogle Scholar

  • Avrutin, Sergey. 1999. Development of the syntax-discourse interface. Dordrecht: Springer. Google Scholar

  • Beekhuizen, Barend. 2010. On abstraction in construction grammar. Leiden: University of Leiden MA thesis. https://www.academia.edu/9690223/On_abstraction_in_construction_grammar_An_exercise_in_methodology (accessed 20 June 2016). 

  • Bertinetto, Pier Marco, Karen H. Ebert & Casper de Groot. 2000. The progressive in Europe. In Östen Dahl (ed.), Tense and aspect in the languages of Europe, 517–558. Berlin & New York: Mouton de Gruyter. Google Scholar

  • Blom, Elma. 2007. Modality, infinitives, and finite bare verbs in Dutch and English child language. Language Acquisition 14(1). 75–113. CrossrefGoogle Scholar

  • Boogaart, Ronny. 1999. Aspect and temporal ordering: A contrastive analysis of Dutch and English. Amsterdam: Free University of Amsterdam dissertation. Google Scholar

  • Booij, Geert. 2008. Constructional idioms as products of linguistic change: The aan het + INFINITIVE construction in Dutch. In Alexander Bergs & Gabriele Diewald (eds.), Constructions and language change, 79–104. Berlin & New York: Mouton de Gruyter. Google Scholar

  • Bourdin, Philippe. 2003. On two distinct uses of go as a conjoined marker of evaluative modality. In Roberta Facchinetti, Frank R. Palmer & Manfred Krug (eds.), Modality in contemporary English, 21–45. Berlin & New York: Mouton de Gruyter. Google Scholar

  • Bres, Jacques & Emmanuelle Labeau. 2013. Allez donc sortir des sentiers battus! La production de l’effet de sens extraordinaire par aller et venir. Journal of French Language Studies 23. 151–177. CrossrefGoogle Scholar

  • Brisard, Frank. 2013. An account of English tense and aspect in cognitive grammar. In Kasia M. Jasczcolt & Louis de Saussure (eds.), Time: Language, cognition, and reality, 210–235. Oxford: Oxford University Press. Google Scholar

  • Bybee, Joan, Revere Perkins & William Pagliuca. 1994. The evolution of grammar: Tense, aspect and modality in the languages of the world. Chicago: University of Chicago Press. Google Scholar

  • Celle, Agnès & Laure Lansari. 2015. On the mirative meaning of aller + infinitive compared with its equivalents in English. Cahier Chronos 27. 289–305. Google Scholar

  • De Wit, Astrid. 2017a. The present perfective paradox across languages. Oxford: Oxford University Press. Google Scholar

  • De Wit, Astrid. 2017b. The relation between mirativity and aspect. Review of Cognitive Linguistics 15(2). 385–410. Google Scholar

  • De Wit, Astrid & Frank Brisard. 2014. A cognitive grammar account of the semantics of the English present progressive. Journal of Linguistics 50(1). 49–90. CrossrefGoogle Scholar

  • De Wit, Astrid & Frank Brisard. 2017. GO auxiliaries as markers of mirativity. Paper presented at the 15th international pragmatics conference, 16–21 July, Belfast. Google Scholar

  • De Wit, Astrid & Adeline Patard. 2013. Modality, aspect and the progressive: The semantics of the present progressive in French in comparison with English. Languages in Contrast 13(1). 113–132. CrossrefGoogle Scholar

  • De Wit, Astrid, Adeline Patard & Frank Brisard. 2013. A contrastive analysis of the present progressive in French and English. Studies in Language 37(4). 846–879. CrossrefGoogle Scholar

  • De Wit, Astrid, Peter Petré & Frank Brisard. Ms. Standing out with the progressive. Google Scholar

  • Declerck, Renaat, Susan Reed & Bert Capelle. 2006. The grammar of the English verb phrase, vol. 1: The grammar of the English tense system. Berlin & New York: Mouton de Gruyter. Google Scholar

  • DeLancey, Scott. 2001. The mirative and evidentiality. Journal of Pragmatics 3. 371–384. Google Scholar

  • Devos, Maud & Jenneke van der Wal (eds.). 2014. ‘COME’ and ‘GO’ off the beaten grammaticalization path. Berlin & Boston: De Gruyter Mouton. Google Scholar

  • Donabédian-Demopoulos, Anaïd. 2012. Evidentiel et progressif: quel statut grammatical pour la saillance prédicative? Faits De Langues 39. 65–82. Google Scholar

  • Ebert, Karen H. 2000. Progressive markers in Germanic languages. In Östen Dahl (ed.), Tense and aspect in the languages of Europe, 605–654. Berlin & New York: Mouton de Gruyter. Google Scholar

  • Emenanjo, E. Nolue. 1987. Elements of modern Igbo grammar: A descriptive approach. Ibadan: Ibadan University Press. Google Scholar

  • Fillmore, Charles J., Paul Kay & Mary Catherine O’Connor. 1988. Regularity and idiomaticity in grammatical constructions: The case of let alone. Language 64(3). 501–538. CrossrefGoogle Scholar

  • Franckel, Jean-Jacques. 1989. Etude de quelques marqueurs aspectuels du français. Geneva: Droz. Google Scholar

  • Güldemann, Tom. 2003. Present progressive vis-à-vis predication focus in Bantu: A verbal category between semantics and pragmatics. Studies in Language 27(2). 323–360. CrossrefGoogle Scholar

  • Haeseryn, Walter, Kirsten Romijn, Guido Geerts, Jaap de Rooij & Maarten C. van den Toorn. 1997. Algemene Nederlandse Spraakkunst [General grammar of Dutch], 2nd edn. Groningen: Martinus Nijhoff. Google Scholar

  • Haeseryn, Walter, Kirsten Romijn, Guido Geerts, Jaap de Rooij & Maarten C. van den Toorn. 2002. Algemene Nederlandse spraakkunst [General grammar of Dutch]. http://ans.ruhosting.nl/e-ans/index.html (accessed 22 December 2016) 

  • Haspelmath, Martin. 1999. Why is grammaticalization irreversible? Linguistics 37(6). 1043–1068. Google Scholar

  • Heine, Bernd. 1995. Agent-oriented vs. epistemic modality. In Joan L. Bybee & Suzanne Fleischman (eds.), Modality in grammar and discourse, 17–53. Amsterdam & Philadelphia: John Benjamins. Google Scholar

  • IJbema, Aniek. 2002. Grammaticalization and infinitival complements in Dutch. Leiden: University of Leiden dissertation. Google Scholar

  • Keller, Rudi. 1994. On language change: The invisible hand in language. London: Routledge. Google Scholar

  • Labov, Wiliam. 1998. Co-existent systems in African-American vernacular English. In Salikoko S. Mufwene, John R. Rickford, Guy Bailey & John Baugh (eds.), The structure of African-American English, 110–153. London: Routledge. Google Scholar

  • Lambrecht, Knud. 1990. “What, me worry?” – ‘Mad magazine sentences’ revisited. Proceedings of the sixteenth annual meeting of the Berkeley linguistics society. 215–228. Google Scholar

  • Langacker, Ronald W. 1987. Foundations of cognitive grammar, vol. 1: Theoretical prerequisites. Stanford: Stanford University Press. Google Scholar

  • Langacker, Ronald W. 1991. Foundations of cognitive grammar, vol. 2: Descriptive application. Stanford: Stanford University Press. Google Scholar

  • Langacker, Ronald W. 2001. The English present tense. English Language and Linguistics 5. 251–273. CrossrefGoogle Scholar

  • Lasser, Ingrid. 2002. The roots of root infinitives: Remarks on infinitival main clauses in adult and child language. Linguistics 40. 767–796. Google Scholar

  • Lemmens, Maarten. 2005. Aspectual posture verb constructions in Dutch. Journal of Germanic Linguistics 17. 183–217. Google Scholar

  • Lemmens, Maarten. 2012a. Variatie en verankering bij progressiefconstructies in het Nederlands [variation and entrenchment of progressive constructions in Dutch]. Nederlandse Taalkunde 17. 284–291. CrossrefGoogle Scholar

  • Lemmens, Maarten. 2012b. Een contrastieve analyse van progressiefconstructies in het Nederlands, het Engels en het Zweeds [a contrastive analysis of progressive constructions in Dutch, English and Swedish]. Handout IVN-colloquium, workshop contrastieve constructiegrammatica, University of Antwerp, 27–31 August 2012. Google Scholar

  • Lemmens, Maarten. 2015. Zit je te denken of ben je aan het piekeren? Persistentie in het synchrone gebruik van de PREP- en POS-progressiefconstructies in het Nederlands [‘sit’ you thinking or ‘are’ you worrying? persistence in the synchronic use of PREP- and POS-progressive constructions in Dutch]. Nederlandse Taalkunde 20(1). 5–36. Google Scholar

  • Levshina, Natalia. 2015. How to do linguistics with R: Data exploration and statistical analysis. Amsterdam & Philadelphia: John Benjamins. Google Scholar

  • Ljung, Magnus. 1980. Reflections on the English progressive. Gothenburg: Acta Universitatis Gothoburgensis. Google Scholar

  • Mortier, Liesbeth. 2008. An analysis of progressive aspect in French and Dutch in terms of variation and specialization. Language in Contrast 8(1). 1–20. CrossrefGoogle Scholar

  • Nikolaeva, Irina. 2014. The narrative infinitive construction in French and Latin. In Hans C. Boas & Francisco Gonzálvez-García (eds.), Romance perspectives on construction grammar, 139–180. Amsterdam & Philadelphia: John Benjamins. Google Scholar

  • Oosterhoff, Jenneke A. 2009. Intermediate Dutch: A grammar and workbook. London: Routledge. Google Scholar

  • Petré, Peter. 2017. The extravagant and the progressive: An experimental corpus study on the grammaticalization history of [BE Ving]. English Language and Linguistics 21(2). 227–250. CrossrefGoogle Scholar

  • Slobin, Dan & Ayhan Aksu. 1982. Tense, aspect, modality, and more in Turkish evidentials. In Paul Hopper (ed.), Tense-aspect: Between semantics and pragmatics, 185–200. Amsterdam & Philadelphia: John Benjamins. Google Scholar

  • Szcześniak, Konrad & Małgorzata Pachoł. 2015. What? Me, lie? The form and reading of the incredulity response construction. Constructions 1. https://www.constructions.uni-osnabrueck.de (accessed 25 November 2016). 

  • Talmy, Leonard. 1988. Force-dynamics in language and cognition. Cognitive Science 12. 49–100. CrossrefGoogle Scholar

  • Tommola, Hannu. 2000. Progressive aspect in Baltic Finnic. In Östen Dahl (ed.), Tense and aspect in the languages of Europe, 655–692. Berlin & New York: Mouton de Gruyter. Google Scholar


  • 1

    We also searched for constructions with bezig and the posture verb construction hangen te + V, but ended up excluding them from the database because they yielded too few relevant examples. 

  • 2

    In German, progressive aspect can also be expressed via beim V-inf sein or im V-inf sein, yet these constructions are marginal compared to the am V-inf sein construction. 

  • 3

    Technically, zijn ‘to be’ is not the only auxiliary that can be used with aan het and a nominalized infinitive (see Booij 2008: 85–88 for a discussion of the verbs that may appear with the pattern), though it is by far the most frequent one. 

  • 5

    The variables listed here were merely included to verify whether they interacted with the presence of (inter)subjective qualifications in significant ways. Since they do not turn out to be obviously relevant for the present investigation, we will not return to these parameters in the remainder of this paper. Note also that no principal distinction will be made between utterances by speakers of Netherlandic Dutch (NL) and utterances by speakers of Belgian Dutch (FL). While Netherlandic Dutch does have a higher proportion of (inter)subjective uses, this is related to the observation that (in CGNa) speakers of Netherlandic Dutch use posture and motion progressives relatively more often than speakers of Belgian Dutch. As will be shown in Section 2.1.2, these constructions are more readily associated with (inter)subjective uses than the aan het-progressive (the latter being relatively more common in Belgian Dutch). Moreover, the motion progressive, which is almost exclusively used in (inter)subjective contexts, seems largely restricted to Netherlandic Dutch: the lopen progressive is only attested twice in the Belgian part of CGNa. 

  • 6

    For the remainder of our analysis it proved to be most informative to discuss (the few) cases of disagreement and examine the thought processes that led to a particular classification. Since there were no consistent interpretational differences among the three annotators, no statistic interrater measurements (such as Cohen’s kappa) were applied. 

  • 7

    Running a chi-squared test for independence allows us to reject the null hypotheses that the categorical variables ‘construction type’ and ‘(inter)subjective use’ are not correlated (χ² = 208.06, p < 0.05). Applying the same test only to the posture verb constructions yields a negative result (χ² = 0.22, p = 0.90), that is, the difference between the various posture verbs is not significant (they form a class). We mention these statistics only for sake of completion. Strictly speaking, we should not perform a chi-squared test on the CGNa data because it is possible that a single speaker is responsible for more than one of the progressive utterances, which would violate the condition of independence of observations (note that this condition is often relaxed or downright ignored in corpus linguistics (cf. Levshina 2015: 212, fn. 2 )). 

  • 8

    Chi-squared test statistic: χ² = 51.35, p < 0.05. Our reservations concerning the independence of observations (cf. fn. 7) hold here as well. 

  • 9

    As can be inferred from the questionnaire (Appendix A), progressive and non-progressive forms were randomly assigned to either A or B (i.e. it is not the case that either form was always presented as A or B). 

  • 10

    See Appendix A for translations. 

  • 11

    Metalinguistic variables include age (<20 years: 20.93%, 20–39 years: 20.93%, 40–59 years: 25.58%, 60 + : 32.56%), gender (male: 37.21%, female: 62.79%) and region (The Netherlands: 23.26%, Flanders: 76.74%). 

  • 12

    Constructed control contexts (Q03, Q15), where one of the options was grammatically impossible for reasons other than aspectual ones (cf. 2.2.1), were not included in this graph. 

  • 13

    The chi-squared test statistic for goodness of fit provided the following results: Q08: χ² = 15.72; Q18: χ² = 11.26; Q10: χ² = 11.26; Q02: χ² = 6.72; Q01: χ² = 3.93. For all the questions mentioned here, the results are significant at p < 0.05. 

  • 14

    The exception is Q06 (χ² = 3.93, p = 0.047), which exhibits a preference for the non-progressive form. 

  • 15

    In spontaneous spoken Dutch (as represented by CGNa), too, the percentage of progressive verb forms does not exceed 1% (even if you exclude a list of the most commonly used stative verbs). There are, however, large differences between individual verbs. 

  • 16

    Only 10% of the progressive preferences in the survey were given a temporal motivation. Of those, in more than half of the cases (53.57%) the informants mention that the utterance with the progressive stresses the long duration of the event, which is not merely a temporal motivation, but at the same time constitutes an (inter)subjective use as highlighting the duration typically intensifies an event. 

  • 17

    We thank an anonymous reviewer for pointing this out to us. 

  • 18

    See De Wit (2017b) for a similar account of GO-auxiliaries and the perfect. 

  • 19

    It should be noted that Dutch generally makes use of posture verbs to locate animate and inanimate entities (instead of the verb zijn ‘to be’). 

  • 20

    Apart from these semantic differences, register is also found to influence the use of both construction types, since the aan het-progressive occurs more often in more formal language (Beekhuizen 2010, following Haeseryn et al. 1997: 735), whereas the posture verb progressive is more frequent in colloquial language. Note that register also influences the choice of posture verb: more formal (written) language prefers the posture verb staan in the posture verb progressive (see Lemmens 2012b), whereas in informal (spoken) language, the verb zitten is most frequent (Beekhuizen 2010: 117; Lemmens 2012b, see also the findings in our corpus study). 

  • 21

    We use capitalized COME and GO as crosslinguistic labels to refer to those verbs that are glossed as ‘come’ and ‘go’ in individual languages. 

  • 22

    It is noteworthy that both (27) and (28) involve inverted sentence structure. In this respect it is interesting to point out that Dutch exclamatives, which by definition involve a sense of discrepancy, obligatorily involve inversion, even with pronominal subjects. Unfortunately though, the precise relation between (inter)subjective expressions and word order in Dutch and other languages lies beyond the scope of this paper. 

  • 23

    Examples (34) and (35) were taken from the internet. 

  • 24

    Examples (39)–(41) have been taken from the internet (accessed November 24, 2016). Example (41) is an excerpt from José van Reijen and Ton Haans. 2008. Groepsdynamica in gedragstherapeutische en psychodynamische groepen. BohnStafleu van Loghum: Houten, available via Google Books. Example (42) is a constructed example. 


Online erschienen: 08.06.2019

Erschienen im Druck: 25.09.2019

Quellenangabe: Linguistics, Band 57, Heft 5, Seiten 1111–1159, ISSN (Online) 1613-396X, ISSN (Print) 0024-3949, DOI: https://doi.org/10.1515/ling-2019-0019.

Zitat exportieren

© 2019 Walter de Gruyter GmbH, Berlin/Boston.Get Permission

Kommentare (0)