The Re-Autonomization of the Modal Auxiliaries in Dutch

Home , Frisian languages, Jan Kops, Languages of Germany

The re-autonomization of the modal auxiliaries in Dutch

De herautonomisering van de Nederlandse modale hulpwerkwoorden

Faculty of Arts Department of Linguistics

The re-autonomization of the modal auxiliaries in Dutch

Thesis submitted for the degree of doctor in Linguistics at the University of Antwerp to be defended by

Wim Caers

PROMOTOR prof. dr. Jan Nuyts Antwerp 2020

Preface

This thesis could not have come about without the help of many people. I would like to express a heartfelt “thank you” to everyone. First of all, I would like to thank the Research Foundation Flanders (FWO) and the Univer- sity of Antwerp for their funding and support without which I was not able to do this research. Most in particular, I am very grateful to my supervisor, professor doctor Jan Nuyts, who has put a lot of time and effort in guiding me through this project. He was always willing to answer my many questions, to actively participate in the thinking process, and to encourage me in developing a critical mind regarding the issues encountered in this thesis. His constructive and pertinent feedback has strongly contributed to the end result of my work. Thanks are also due to Henri-Joseph Goelen (former predoctoral researcher in the linguistics department of the University of Antwerp) for his preliminary work in examining the grammatical characteristics of the less central modals in Dutch. Along with the earlier investigations of my supervisor, his pilot study has led to the outcome of my research. Since this investigation was intended to be part of a larger project into a re-autonomization process also in the (Old and (Early) Middle) English modals, conducted in Amsterdam, I am also very grateful to my colleague, PhD student Sune Gregersen, and his supervisor, emeritus professor doctor Olga Fischer, who are both affiliated at the University of Amsterdam. I would like to thank them for their input at the start of the project, for the most agreeable work meetings, for their useful and inspiring comments on my work, and last but not least for their warm welcome during my research stays in Amsterdam. I also want to thank Sune, in particular, for hosting me during these stays, for the nice collaboration, in writing and discussing the most relevant literature, as well as for the co-publication of a wonderful squib (see Caers and Gregersen 2019), which hopefully may inspire many other linguists to investigate the intriguing phenomenon of re-autonomization. In addition, I am grateful to numerous people I was fortunate to get to know on linguistic conferences. I would like to thank them all for the interesting discussions. Moreover, I want to IV THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH give credit to the anonymous reviewers for their useful suggestions on the articles that have come from this study. Many thanks are due to the members of my Individual Doctoral Committee, professors doctor Peter Petré, Jan Nuyts and Olga Fischer, as well as to the other members of the jury, professors doctor Reinhild Vandekerckhove and Freek Van de Velde, for taking the time to thoroughly assess my work. I also want to thank my colleagues at the University of Antwerp, with whom I could always share thoughts and experiences. Last but not least, I am very grateful to my parents, my brother, and especially my fiancée, who have always been very supportive. They have seen me on my best, but also on my worst, during this project. Thanks to their encouragements, their comforting words, and their help in general, I have managed to keep going and to give my all to obtain the best result possible.

Table of contents

Preface …………………………………………………………………………………………………………………………… iii Introduction ……………………………………………………………………………………………………………….…. ix

CHAPTER 1 Literature ………………………………………………………………………………………………………………………. 1 1.1 Important notions ………………………………………………….………………………………….……………..… 1 1.1.1 Grammaticalization and unidirectionality ………………………………………………………… 1 1.1.1.1 The term grammaticalization ……………………………………………………………….. 2 1.1.1.2 Parameters of grammaticalization …………………………………………..…….…….. 4 1.1.1.3 Mechanisms of change: reanalysis and analogy ………………………………….… 8 1.1.1.4 The “ontology” question ………………………………………..……………………………. 11 1.1.1.5 The “unidirectionality” question ………………………………………………………….. 14 1.1.2 Modality and (inter)subjectification ………………………………………………………………… 21 1.1.2.1 Notions of modality ……………………………………………………………………………… 21 1.1.2.2 The diachrony of modality ……………………………………………………………………. 28 1.1.2.3 (Inter)subjectification and unidirectionality …………………………………………. 29 1.1.2.4 Mechanisms of change: metaphor and metonymy ……………………………….. 36 1.2 Language-specific overview ………………………………………………….……………………………………. 37 1.2.1 (De)grammaticalization in the modal verbs ……………………………………………………… 37 1.2.1.1 The modals in Germanic ………………………………………………………………………. 37 1.2.1.2 The modals in Dutch …………………………………………………………………………….. 44 1.2.1.2.1 The central modals ……………………………………………………………………. 44 1.2.1.2.2 The less central modals ……………………………………..……………………… 46 1.2.1.2.3 The combination with a directional …………………..………………………. 49

VI THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

1.2.2 Meaning evolutions in the modal verbs …………………………………………………………… 52 1.2.2.1 The modals in Germanic ………………………………………………………………………. 52 1.2.2.2 The modals in Dutch …………………………………………………………………………….. 58 1.2.2.2.1 The central modals ……………………………………………………………………. 58 1.2.2.2.2 The less central modals ……………………………..……………………………… 60 1.2.3 Summary of the findings relevant for the present corpus study ………………………. 62

CHAPTER 2 Methodology ………………………………………………………………………………………………………………… 63

CHAPTER 3 Analytical categories ……………………………………………………………………………………………………. 71 3.1 The grammatical dimension ………………………………..……………………………………………………… 71 3.1.1 The notion of ‘transitivity’ and the concept of ‘an argument’ …………………………… 71 3.1.2 Types of arguments ……………………………………………………………………..…………………. 74 3.1.3 Types and subtypes of uses of the modal verb …………………………………………………. 76 3.1.3.1 Definitions …………………………………………………………………………………………… 76 3.1.3.2 Complications ……………………………………………………………………………………… 84 3.1.4 Directionals …………………………………………………………………………………………………….. 88 3.2 The semantic dimension ……………………….……………………………………………………………………. 89 3.2.1 Meaning categories ………………………………………………………………………………………… 89 3.2.2 Negative polarity …………………………………………………………………………………………….. 98

CHAPTER 4 The grammatical evolution ………………………………………………………………………………………… 101 4.1 Relevant frequency data from Nuyts (et al.) …………………….………………………………………… 101 4.2 Results of the present study ……………….……………………………………………………………………… 103 4.2.1 Timing of the re-autonomization process ……………………………………………………….. 103 4.2.1.1 Different types of uses ………………………………………………………………………… 103 4.2.1.2 Two subtypes of ‘new main V’ uses ……………………………………………………... 107

TABLE OF CONTENTS VII

4.2.2 Directionals in the re-autonomization process ………………………..……………………… 108 4.2.2.1 The role of directionals ………………………………………………………………………… 108 4.2.2.2 Limits to the role of directionals ………………………………………..………………… 112 4.3 Reflection and discussion ………………………….………………………………………………………………. 117

CHAPTER 5 The semantic evolution ………………………………………………………………………………………………. 123 5.1 General meaning profile according to Nuyts et al. ………………………………………………………. 123 5.2 Results of the present study …………………………….………………………………………………………… 125 5.2.1 Meaning profile of the new autonomous uses ………………………………………………… 125 5.2.1.1 A brief overview of all meanings and uses ……………………………………………. 125 5.2.1.2 Meanings of the ‘aux V implicit’ uses …………………………………..………………. 131 5.2.1.3 Meanings of the ‘new main V’ uses ………………………………..……………………. 136 5.2.2 The role of negative polarity in the re-autonomization process ………………..…….. 142 5.3 Conclusion ………………………………………………………………………………………………………………… 150

CHAPTER 6 Re-autonomization beyond (standard) Dutch ……………………………………………………….… 151 6.1 Material and methods ………………….……………………………………………………………………………. 151 6.2 Findings ………………………….…………………………………………………………………………………………. 154 6.2.1 Dutch dialects ………………………………………………………………………………………………… 154 6.2.2 Frisian ……………………………………………………………………………………………………………. 156 6.2.3 Low German …………………………………………………………………………………………………… 158 6.2.4 Afrikaans ………………………………………………………………………………………………………… 161 6.3 Concluding remarks ………………………………………………………………………………………………….. 162

CHAPTER 7 Conclusions and perspectives ……………………………………………………………………………………. 165

References ……………………………………………………………………………………………………………………. 169 Appendix 1 …………………………………………………………………………………………………………………… 183 Abstract ………………………………………………………………………………………………………………………… 223

Introduction

Although the modal verbs in Dutch (as well as in other Germanic languages) are most often used in combination with another main verb in the clause, they nevertheless may also occur independently in Present Day Dutch (in a way which is not possible in standard German or English), as illustrated in (1a–c).

(1) a. dat kan. [lit:] ‘that can [i.e. that is possible].’ b. dat mag niet. [lit:] ‘that may not [i.e. that is not permitted].’ c. het moet. [lit:] ‘it must [i.e. it is necessary].’

This phenomenon, referred to as the ‘autonomous’ use of the modals, has been investigated in a preliminary way in Nuyts (2013; Nuyts et al. 2018, 2019). These studies have shown that such uses only emerged sometime in the course of the New Dutch period, and were considerably increasing in frequency in Present Day Dutch. The present study aims to offer a more detailed analysis of this re-autonomization process in the Dutch modal verbs, by means of diachronic corpus data. Specifically, we want to investigate when the process started precisely, and how it evolves. What are the grammatical and semantic properties of the autonomous uses, and what factors play a role in their emergence and development? In a wider perspective, we will consider the implications of the empirical analyses for current views on the processes of (de)grammaticalization and (inter)subjectification. In line with the analysis in Nuyts (2013), we will argue that the re-autonomization process of the modal verbs in Dutch is a case of collective degrammaticalization, though one which does not also involve de-(inter)subjectification. This study is organized as follows. Chapter 1 reviews the relevant literature. On the one hand, it surveys current views on the notions figuring centrally in this study, including (de)- grammaticalization, modality and (inter)subjectification. On the other hand, it offers an overview of earlier work on the autonomous use of the modals, in Germanic in general, and in Dutch X THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH in particular. The latter includes a more detailed presentation of the investigations by Nuyts and colleagues, as the starting point for the present study. Chapter 2 presents and motivates the methodology of the present corpus study. In Chapter 3 we define the grammatical and semantic categories that will be applied in the data analyses. In Chapters 4 and 5 we present the results. Chapter 4 focusses on the grammatical analysis, Chapter 5 on the semantic analysis. Both chapters are structured similarly. They first offer a summary overview of the for the present study relevant frequency data obtained in the earlier studies by Nuyts and colleagues. Then we present and discuss the frequency data resulting from our present corpus study. Finally we reflect on what the analyses imply for our understanding of the mechanisms at work in the re-autonomization process. Chapter 6 considers the process in a cross-linguistic perspective. It briefly presents the results of a few small pilot studies into the occurrence of autonomous uses of the modals beyond (standard) Dutch, with special focus on Frisian, Low-German and Afrikaans. Chapter 7, finally, reflects on the theoretical implications of the investigation for our understanding of the processes of (de)grammaticalization and (inter)subjectification.

CHAPTER 1

Literature

In this chapter we present an overview of our current state of knowledge about a number of notions and dimensions that are important for our analysis of the autonomous uses of the modals. Section 1.1 surveys the literature on the notions of grammaticalization, and of modality and (inter)subjectification. Section 1.2 reports on the literature regarding the status of the diachrony of the modal verbs in the Germanic languages.

1.1 Important notions1

In this section we deal with two major concepts relevant for the present investigation: section 1.1.1 discusses the process of grammaticalization and its (often presumed) unidirectionality, section 1.1.2 focuses on the notions of modality and of subjectification and intersubjectification.

1.1.1 Grammaticalization and unidirectionality

Grammaticalization is a popular phenomenon. The term returns more than 40,000 hits in the Google Scholar database, and there are numerous monographs and collected volumes, as well as at least two textbooks (Diewald 1997; Hopper & Traugott 2003 [1993]), devoted to the subject. This apparent interest in it does not, however, reflect unanimous agreement about what it is – or even whether it is a legitimate field of research at all – and since its popularization in the 1980s, few issues in historical linguistics have been as hotly debated as the directionality, regularity, and ontology of grammaticalization. In this section, then, we will provide a brief overview of the history of the term grammaticalization since its introduction by Meillet

1 This section was written in collaboration with Sune Gregersen (University of Amsterdam). 2 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(1.1.1.1) and we will discuss a number of issues related to it: the distinguishing characteristics (1.1.1.2) and mechanisms (1.1.1.3) of grammaticalization, the ‘ontological’ question (1.1.1.4), i.e. whether grammaticalization is a process in its own right or a cover term for a number of distinct processes; and the ‘unidirectionality’ question (1.1.1.5), i.e. whether grammaticalization is reversible or always moves in the same direction, from less to more grammatical.

1.1.1.1 The term grammaticalization

Although the phenomenon was noticed already in the 19th Century (cf. von Humboldt [1825] and von der Gabelentz [1891]), the term grammaticalization was used first in 1912 by the French linguist Antoine Meillet. In his article “L’évolution des formes grammaticales”, Meillet (1912) considered ‘grammaticalization’ as one of the two processes (the other being ‘analogy’) causing the emergence of new grammatical forms, and exemplified what he called “le passage d’un mot autonome au rôle d’élément grammatical” [‘the transition of an autonomous word to the role of grammatical element’, our translation] by means of the French verb être (Meillet 1912: 387). Within the phrase je suis celui qui suis ‘I am the one who is’, Meillet explained, the conjugated verb has an autonomous meaning (‘exist’), which it still preserves more or less in the locative sentence je suis chez moi ‘I am at home’. The phrase je suis malade ‘I am ill’, however, no longer shows such kind of autonomy, according to Meillet, since the verb’s meaning is almost superfluous. And in je suis parti ‘I left’, finally, the meaning of être has been lost completely – the conjugated verb is nothing but a purely grammatical element functioning as a time-aspect auxiliary. For several decades after Meillet’s work, the term grammaticalization was used only very rarely. This does not mean, however, that linguists were not aware of developments like those described by Meillet. For instance, in his English grammar “on historical principles”, Otto Jes- persen (1931) describes how the motion expression is going with an infinitive has developed into “an expression for future time”, and he is aware that similar developments have happened in other languages: “going loses its meaning as a verb of movement and becomes an empty, grammatical word; cf. French je vais faire and similar expressions in other Romanic languages” (Jespersen 1931: 217). Although Jespersen does not use the term grammaticalization, his description of the emergence of future be going to is very similar to the way it is described in LITERATURE 3 the grammaticalization literature, and be going to is one of the textbook examples of the phenomenon (cf. Hopper & Traugott 2003: 1). In a much-cited article on the emergence of grammatical categories in Indo-European languages, Kuryłowicz (1965) gives another example which has become a textbook case, viz. the emergence of tense markers out of a verb meaning ‘have’. In the same paragraph, almost as an afterthought, he provides what is perhaps the most cited definition of grammaticalization:

The way from Latin habeo litteras scriptas to the French past j’ai ecrit la lettre has been a rather long one. The French form represents an advanced stage of grammaticalization of a lexical phrase […] Grammaticalization consists in the increase of the range of a morpheme advancing from a lexical to a grammatical or from a less grammatical to a more grammatical status, e.g. from a derivative formant to an inflectional one. (Kuryłowicz 1965: 69, italics in original)

Another influential work is Talmy Givón’s (1971) paper on the genesis of inflectional and derivational morphology. The paper is perhaps less known for the case studies in it as for its final motto, “Today’s morphology is yesterday’s syntax” (Givón 1971: 413), which Givón attributes to Lao Tse in a somewhat different version, and which has been repeated many times in the grammaticalization literature. The main thesis of Givón’s paper is that the synchronic morphology of a language, e.g. whether it prefers prefixes or suffixes, can often be explained by the word order preferences of earlier stages of the language, since such affixes tend to derive from independent words. Givón does not use the term grammaticalization, but describes the developments in other terms, e.g. as “the condensation of main ‘modal’ verbs into modality prefixes in Bantu” (Givón 1971: 394, our italics). A 1982 working paper by Christian Lehmann, entitled Thoughts on grammaticalization, had a substantial impact on historical linguistics, despite remaining formally unpublished until 1995. It has since been chosen as the very first volume to appear in the book series Classics in linguistics by Language Science Press (Lehmann 2015). Lehmann provides both an overview of the research history of grammaticalization and a large collection of examples – including the modal auxiliaries in English – but more importantly, he introduces the notion syntagmatic and paradigmatic parameters of grammaticalization, which we will return to below. 4 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

The 1990s and early 2000s, then, saw the publication of a significant number of monographs, edited volumes, and papers on grammaticalization, both in individual languages and cross-linguistically. Heine, Claudi & Hünnemeyer (1991) primarily deal with grammaticalization phenomena in African languages, while Traugott & Heine (1991) collect a number of case studies on different languages. Another much-cited work, Bybee, Perkins & Pagliuca (1994), is a cross-linguistic study of the grammaticalization of markers of tense, aspect, and modality. We also find a number of works devoted to Germanic languages specifically, such as Diewald (1999) on grammaticalization in the German modals, Krug (2000) on ‘emerging’ modals in English such as have to and want to, and the studies of English collected in Fischer, Rosenbach & Stein (2000). Finally, Diewald (1997) and Hopper & Traugott (2003 [1993]) are two textbooks exclusively devoted to grammaticalization.

1.1.1.2 Parameters of grammaticalization

A number of different changes have been characterized as examples of grammaticalization in the literature, such as the development of auxiliaries or modal affixes out of main verbs, or the development of case affixes out of adpositions. According to Lehmann (1995 [1982]), what such changes have in common is that they involve a decrease in autonomy or “freedom” of a sign, which Lehmann takes to be the defining characteristic of grammaticalization (we quote from the 1995 edition in the following):

… the more freedom with which a sign is used, the more autonomous it is. Therefore the autonomy of a sign is converse to its grammaticality, and grammaticalization detracts from its autonomy. Consequently, if we want to measure the degree to which a sign is grammaticalized, we will determine its degree of autonomy. (Lehmann 1995: 122)

Lehmann goes on to suggest six parameters which can be used to measure the autonomy (and, conversely, the degree of grammaticalization) of a linguistic item. Since these parameters have been very influential, but also contested, in the literature, we will provide a somewhat detailed explanation of them here. Lehmann proposes that linguistic signs can vary in terms of their weight, their cohesion with other signs, and their variability. These three aspects, in turn, can vary along two axes, viz. LITERATURE 5 paradigmatic selection (the paradigmatic axis) and syntagmatic combination (the syntagmatic axis), as illustrated in Figure 1.

Paradigmatic axis

Syntagmatic axis

. . .

Figure 1. The two linguistic axes

Under this view, linguistic signs (morphemes) combine with other signs on the syntagmatic axis (the x-axis in Figure 1). However, the language user always has to make choices, i.e. make a selection, between different signs, and each sign can thus be said to stand in a paradigmatic relationship to other signs which can be used in the same ‘slot’ (visualized as broken-line rectangles on the y-axis in Figure 1). By combining the three different aspects with the paradigmatic and syntagmatic axes, then, we get six parameters, which Lehmann terms integrity, paradigmaticity, paradigmatic variability, structural scope, bondedness, and syntagmatic variability (cf. Table 1).

Table 1. Parameters of grammaticalization (after Lehmann 1995: 123)

paradigmatic axis syntagmatic axis

weight integrity structural scope

cohesion paradigmaticity bondedness

variability paradigmatic variability syntagmatic variability

6 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Beginning on the paradigmatic axis, Lehmann refers to the weight of a linguistic item as its integrity, i.e. its substantial size, both semantically and phonologically. Lehmann takes grammaticalization to be linked to a loss in integrity. A loss of semantic substance is called desemanticization (or semantic depletion or bleaching); a decrease in phonological integrity is termed phonological attrition (or erosion). To repeat the textbook example mentioned above, in developing from a motion expression to a future auxiliary, English be going to has lost the motion meaning (desemanticization) and developed the reduced variant gonna (phonological attrition). The cohesion of a linguistic item with other elements in a paradigm is termed its paradigmaticity, which Lehmann defines as “the degree to which an item enters a paradigm, is integrated into it and dependent on it” (Lehmann 1995: 123). The higher the degree of paradigmaticity, the more constrained the item is within a paradigm, and the more grammaticalized it is. Lehmann mentions the English primary prepositions “in, on, at, from, to and perhaps some others” as an example of a small, closed paradigm, as opposed to the more open-ended group of secondary prepositions formed from nouns (in front of, at the bottom of, etc.) (Lehmann 1995: 133). A closely related parameter is paradigmatic variability, which refers to the substitu- tability and omissibility of an item, i.e. the degree to which it can be substituted either by other members of the paradigm or by zero. The less paradigmatically variable an item is, the higher its obligatoriness and degree of grammaticalization. Lehmann refers to a well-known idea popularized by Roman Jakobson (1971) (who in turn attributes it to Franz Boas) that the grammar of a language consists in all those choices which the speaker is forced to make, i.e. the concepts which one has to express in the language. The same concepts can be “grammaticalized and consequently obligatory in some languages but lexicalized and merely optional in others” (Jakobson 1971: 492). In Lehmann’s terminology it is the linguistic element itself which is said to be grammaticalized, whereas in Jakobson’s it is the concept it expresses, but otherwise their ideas of obligatoriness seem to conform. (Note, however, that Lehmann repeatedly stresses the graduality of grammaticalization, i.e. that an item can be more or less grammaticalized, whereas this appears to be a binary distinction for Jakobson: Either a concept is grammaticalized or it is not.) LITERATURE 7

On the syntagmatic axis, the weight of an item is called its structural scope by Lehmann, which refers to its scope over other elements. Among the examples given by Lehmann are the development of case affixes, which have scope over a noun, out of prepositions, which have scope over the whole noun phrase, and the development of serial verbs and auxiliaries out of main verbs. For instance, the main verb have in a sentence like I have a cat functions at the clausal level, whereas the auxiliary have in I have bought a cat only functions at the verb phrase level and hence has narrower scope (Lehmann 1995: 144; but cf. below on the problems connected with this specific parameter). Syntagmatic cohesion, i.e. bondedness, is “the degree to which an item depends on, or attaches to, other signs” (Lehmann 1995: 123). The more bonded an element is, the more closely it depends on another element in the syntagmatic chain. To stay with our examples from the previous parameter, case affixes tend to be more tightly bonded to their host noun than prepositions, and auxiliaries are likely to develop into affixes or clitics, as in English I’ve bought a cat. Finally, the syntagmatic variability of an item refers to its mutability or flexibility within the syntagm, for instance when an adverb develops into an adposition, which is typically structurally fixed in relation to the nominal it forms a syntagm with. Lehmann gives the position of prepositions in Latin as an example: Unlike adverbs and other less grammaticalized parts of speech, prepositions are subject to stricter word order rules and normally have to occur before their complement (Lehmann 1995: 158). It is worth stressing that the individual parameters do not correlate with grammaticalization in the same way: Whereas increased grammaticalization is correlated with increased cohesion, i.e. paradigmaticity and bondedness, the weight and variability of a grammaticalizing item is expected to decrease. Table 2 (on top of the next page) shows how the six parameters are expected to change with increasing grammaticalization. Lehmann’s parameters have been very influential in the grammaticalization literature and are cited in most, if not all, relevant studies and textbooks (cf., inter alia, Diewald 1997: 21–29; Krug 2000: 13–15; Hopper & Traugott 2003: 30–32; Norde 2009: 123–132). However, a number of problems have been pointed out, perhaps most prominently with regard to the parameter of structural scope. Some scholars have observed that the structural scope of an item often 8 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 2. Parameter changes connected with increasing grammaticalization

paradigmatic axis syntagmatic axis

weight integrity (decrease) structural scope (decrease)

cohesion paradigmaticity (increase) bondedness (increase)

variability paradigmatic variability (decrease) syntagmatic variability (decrease)

seems to increase rather than decrease with increased grammaticalization. Tabor & Traugott (1998) propose that the structural scope of an item always increases with increased grammaticalization; Fischer (2010), on the other hand, argues that grammaticalization may be characterized by either scope increase or decrease (or no change in scope at all), and that the parameter of scope is better seen as a heuristic device than as a diagnostic of grammaticalization. Some of the other parameters have also been called into question as criteria for measuring the grammaticalization of linguistic items. For instance, Bisang (2008) argues that a number of grammatical items in Thai and other East and Southeast Asian languages, such as nominal classifiers and tense-aspect-modality markers, do not form paradigms in the sense of Lehmann (1995) and do not show signs of increased bondedness or phonetic reduction. Bisang suggests that the relevant characteristics of grammaticalization may vary between typologically different languages, and that Lehmann’s parameters do not work well for the East and South- east Asian linguistic area. We will not go further into this discussion here, but instead we turn to the mechanisms of change involved in grammaticalization.

1.1.1.3 Mechanisms of change: reanalysis and analogy

In the previous sections, we have presented a number of views on what characterizes grammaticalized items. Another question which has been discussed in the literature is how grammaticalization comes about. Most linguists seem to agree that grammaticalization relies on the working of two general mechanisms: reanalysis and analogy. The first one, reanalysis, is said to take place when a speaker attributes a new semantic and/or syntactic interpretation to a sequence of linguistic elements, and may be defined as “a change in the structure of an LITERATURE 9 expression or class of expressions that does not involve any immediate or intrinsic modification of its surface manifestation” (Langacker 1977: 58). A similar formulation has been found in Hopper & Traugott (1993), who consider reanalysis a mechanism that “modifies underlying representations, whether semantic, syntactic, or morphological, and brings about rule change” (Hopper & Traugott 1993: 32). In other words, reanalysis is essentially conceived of as a ‘covert’ process, since the construction looks formally the same before and after the change. This can be represented, schematically, as follows:

(2) X Y [Z] X [Y Z]

As illustrated in (2), reanalysis may thus be seen as a matter of rebracketing the linguistic structure of an expression. Note, however, that the schema in (2) is not entirely complete, in that it does not take into account the possible coexistence of the old and the new construction, a phenomenon which Hopper (1991) has termed layering. So, instead of (2), the schema in (3) is more appropriate in many cases:

(3) X Y [Z] X Y [Z] X [Y Z]

To repeat one of the examples from section 1.1.1.1, the development of future auxiliaries from motion verbs in some West Germanic languages (English BE going to; Dutch gaan) did not lead to a loss of the original motion verbs, but rather to the coexistence of two different structures. Compare, for instance, I’m going [to visit Bill] with I’m going to [visit Bill]. In the former structure (i.e. the first stage), BE going is a progressive with a verb of motion and a purposive clause. In the latter structure (i.e. the second stage), however, the form has been reanalysed as a future auxiliary with a verb of activity (‘visit’). Whereas reanalysis could be considered a ‘covert’ process (both structures look formally the same), analogy rather can be seen as an ‘open’ mechanism of grammaticalization: “[it] modifies surface manifestations, but in itself does not effect rule change” (Hopper & Traugott 1993: 32). Nevertheless, as Hopper & Traugott (1993) point out, what it does effect, is rule spread. So, the difference between the two mechanisms of grammaticalization corresponds to 10 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH another distinction as well: the distinction between the innovation of a new form (corresponding to reanalysis), and the gradual spread of that new form (comparable to analogy) (Hopper & Traugott 2003: 46). An example of the latter may be found, for instance, in the further extension of BE going to as a future auxiliary with, specifically, a verb of activity (cf. I’m going to [visit Bill], mentioned above) to an auxiliary of the future which may combine with any possible verb (see, for instance, I’m going to [like Bill]). A few critical points can be raised, however, regarding the mechanisms of grammaticalization. Here we will discuss the two most important ones. Firstly, although reanalysis and grammaticalization are sometimes considered equivalents of each other in the literature (as, for instance, suggested by Meillet’s (1912) definition, and even explicitely formulated in Heine & Reh (1984: 95)), they may not be equated. It is a fact, indeed, that many cases of reanalysis are clear examples of grammaticalization (cf. the BE going to case, for instance), but not all of them are. An often cited counterexample is the emergence of the English verb and noun up (as in to up the price and with ups and downs), out of the homonymous preposition. As pointed out in Hopper & Traugott (2003), this change clearly involves “a shift from a grammatical to a lexical structure, rather than from a lexical to a grammatical structure” (which is the norm for grammaticalization) (Hopper & Traugott 2003: 58).2 And secondly, while most scholars (such as, e.g., Harris & Campbell 1995, or Hopper & Traugott 2003) acknowledge the primary role of “reanalysis” as mechanism of grammaticalization, others (like Fischer 2007, for instance) rather perceive “analogy” as being the primary mechanism of the process. According to the former group of scholars, reanalysis is “a pre- requisite for the implementation of the change through analogy” and, thus, should be considered the primary mechanism behind grammaticalization (Hopper & Traugott 2003: 39). Moreover, it is the only mechanism that is able to “create new grammatical structures” (Hopper & Traugott 2003: 64). This does not mean, however, that these enthusiasts of reanalysis neglect the role of analogy entirely – on the contrary, analogy is an important mechanism too (though not the primary one), since it “makes the unobservable changes of reanalysis observable”

2 This provisionally leaves open the question/discussion of how to refer to the shift, then: either as an instance of “degrammaticalization” or an example of “lexicalization” – Hopper & Traugott (2003) will opt for the latter analysis, as will be explained in section 1.1.1.5. LITERATURE 11

(Hopper & Traugott 1993: 61). Fischer (2007), on the other hand, emphasizes the role of analogy more strongly, by considering it at least as equal to the mechanism of reanalysis. After all, reanalysis takes place “within the contours of the communicative situation and the grammatical system in which a structure operates”, and therefore is also “confined and shaped by the formal structures that already exist” (Fischer 2007: 123). As a consequence, Fischer hypo- thesizes that the “reanalysis of a structure will not as a rule result in a totally new structure, but in one that is already in use elsewhere”, and she goes on by saying that “it is the superficial similarity [i.e. the analogy] that a language user perceives between two structures and between two communicative uses of them that causes a reanalysis in one of them, so as to bring it in line with the other” (Fischer 2007: 123–124). Interpreted thusly, the “perception of similarity [i.e. the analogy] must be logically primary to the reanalysis”, according to Fischer (2007: 124). Although one could tell much more about these mechanisms of change, and especially about the points of critique that can be raised regarding them, a further discussion would lead us too far. Hence, we will not go into a further discussion here, but instead we turn to some (other) objections regarding grammaticalization “theory” in the next two sections.

1.1.1.4 The “ontology” question

One of the most frequently voiced objections to the notion of grammaticalization is that it has no independent status as a process or mechanism of change, but rather consists in a number of different – and separate – mechanisms. This point has been argued by Campbell (2001a) and Joseph (2001), among others, in a special issue of Language Sciences devoted to (critiques of) the grammaticalization literature (Campbell 2001b). As Campbell points out and illustrates with several examples, most contributors to this literature are well aware that the term is used to refer to a number of different changes:

The events that occur during this process may be discussed under rubrics of semantic, functional, grammatical, and phonological changes, though we will argue that these processes are intimately connected with one another. (Bybee et al. 1994: 5–6)3

3 Cf. also the following quote from Lehmann: “It is obvious that phonological attrition is omnipresent in linguistic change. It plays its role not only in grammaticalization, but affects, in the long run, practically every sign. … Consequently, it would be wrong to infer from phonological attrition to gramamticalization [sic]. We will 12 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Campbell argues that while such changes may be connected with each other, they need not be, and none of them is necessary or sufficient to define grammaticalization. Since the developments which have been discussed under the heading of grammaticalization in the literature are always reducible to other mechanisms (such as reanalysis and analogical extension, cf. the previous section), Campbell denies that the term has any value except as a heuristic:

In short, grammaticalization is derivative, epiphenomenal, and has no independent status of its own. “Grammaticalization theory” has no explanatory value because what it claims to explain is explained already by other well-understood mechanisms which lie behind it and, as is generally agreed, it cannot “explain” without appeal to these other mechanisms and kinds of change. (Campbell 2001a: 151)

Joseph argues the same point in his contribution as well as in later work (2004, 2011), describing grammaticalization in similar terms as “an epiphenomenon resulting from other processes of change” (Joseph 2001: 185; cf. also Fischer 2010). Moreover, he argues that the development of grammaticalization theory has had the unfortunate consequence that historical linguists have privileged one cluster of changes over others:

One has to wonder … why one particular grouping of changes (semantic shifts of a certain type + phonetic reductions + extension of usage into novel realms, etc.) should be treated as special, deserving its own label, conferences, textbooks, and other compendia, and not just an accidental confluence of factors. (Joseph 2001: 197)

These objections have met various responses in the grammaticalization literature. Hopper & Traugott (2003), for instance, ignore them in the second edition of their textbook, instead focusing only on the proposed counterexamples to unidirectionality (for a discussion of the “unidirectionality” question, cf. below):

meet the same situation with some of the other parameters. None of them is by itself sufficient to define grammaticalization; it is only by the interplay of all of them that grammaticalization comes about.” (Lehmann 1995: 127) LITERATURE 13

The bulk of the arguments [in Campbell 2001b] was devoted to the discussion of a small number of cases in which a reversal of unidirectionality can be argued, some of these, however, such as the English possessive ‘s, being themselves quite controversial. (Hopper & Traugott 2003: 34)

Hopper & Traugott devote an entire chapter to the unidirectionality question, but do not engage with Campbell’s and Joseph’s objection that grammaticalization is not a process in its own right. This question is addressed in a paper by Lehmann (2004), who argues that there are cases of ‘pure’ grammaticalization which cannot be explained by or reduced to any other mechanism. He gives as examples the development of preverbal cross-reference markers out of personal pronouns in Romance varieties, the Ancient Greek passive marker, and the development of indefinite and definite articles in Germanic and Romance languages.4 Other scholars have attempted to provide more restrictive definitions of grammaticalization in functional terms. Diewald (2010) argues that grammaticalization has not been sufficient- ly clearly defined in the literature because the notion of grammar has not been explicitly defined. While there are no mechanisms of change that are unique to grammaticalization, according to Diewald there is a cluster of changes which lead to the development of grammatical items, which she defines in terms of deixis. According to this conception, grammatical items are inherently deictic in nature because they are always anchored to an origo in the context or in the linguistic system; for instance, tense and mood marking anchors an event to the speech situation, and anaphoric pronouns provide anchor points within the linguistic message. The deictic relation can also be of a more abstract nature, such as within a case paradigm where the individual case forms are said to be anchored to the unmarked case (Diewald 2010: 45). An alternative definition is given by Boye & Harder (2012), who criticize the notion that grammaticalization is a complex process consisting of a bundle or cluster of different mechanisms, specifically referring to Heine & Reh (1984) and Lehmann (1995, 2002). Instead, they propose a functional-structural definition of the notion of grammar according to which grammatical expressions are “by convention ancillary and as such discursively secondary” –

4 But cf. Sommerer (2015) for an analogy-based account of the emergence of the definite article in English. 14 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH lexical items, in turn, are “by convention capable of being discursively primary”, i.e. being the main point of an utterance (Boye & Harder 2012: 7). Grammaticalization, then, is the change whereby a discursively secondary use of a lexical item becomes conventionalized and loses the ability to be discursively primary. While phonetic reduction, semantic bleaching, etc. may follow from grammaticalization and often do, these processes are not necessary parts of grammaticalization itself, but possible consequences of it.

1.1.1.5 The “unidirectionality” question

Another often contested issue concerns the question of ‘unidirectionality’, and more specifically, the question whether grammaticalization is reversible or not. In the 1980s, with the renewed interest in grammaticalization studies, linguists believed that the process of grammaticalization was exclusively unidirectional: lexical items turned into grammatical ones, but not vice versa. This hypothesis, also termed the “strong” hypothesis of unidirectionality, has been subject to considerable debate since the 1990s, however. Christian Lehmann (1995 [1982]) used the term degrammaticalization to refer to a phenomenon which he considered non-existent. He encounters grammaticalization as an irreversible process and, consequently, the hypothesis of unidirectionality as an absolute principle. As we will see below, Lehmann’s hypothesis prompted a number of linguists to look for examples of the reverse direction, but many of the proposed examples of degrammaticalization have themselves been called into question. For instance, Heine et al. (1991) agree with Lehmann that unidirectionality is a central characteristic of the grammaticalization process, and consider the few counterexamples that have been suggested in the literature “to be the result of an inadequate analysis” (Heine et al. 1991: 5). Other scholars have made less strong claims about unidirectionality, however. Hopper & Traugott (2003 [1993]), for instance, do not go as far as Lehmann and Heine (et al.) by considering unidirectionality an absolute principle, since counterexamples – though relatively infrequent – do exist (Hopper & Traugott 2003: 132). This does not mean, however, that they accept all instances presented in the literature – on the contrary, most examples are interpreted no real cases of degrammaticalization by Hopper & Traugott. Instead, the exceptions are said to rather represent some other phenomena such as “lexicalization”, a process which Hopper & Traugott characterized as “a shift from grammatical LITERATURE 15 to lexical structure” (Hopper & Traugott 1993: 49). To repeat one of the examples discussed in section 1.1.1.3, the English preposition up has ‘lexicalized’ into a verb, as exemplified, for instance, by the short sentence “He upped the price”. Other examples involve the change of a derivational morpheme into a noun (see, e.g., “I dislike her use of isms”) or the ‘lexicalizing’ of the French second-person singular tu into the verb tutoyer. According to Hopper & Traugott, what these changes have in common is their instantaneousness: “[O]ne can take any element of language, […] and use it lexically” (Hopper & Traugott 2003: 134). So interpreted, these “lexicalizations” may not be regarded as ‘real’ counterexamples to the unidirectionality hypothesis (which is a strong tendency); the only changes that deserve that ‘label’ in the opinion of Hopper & Traugott are those in which “more grammatical items become less so” (Hopper & Traugott 1993: 129) – examples are given below. Haspelmath (2004) follows Hopper & Traugott by arguing that unidirectionality cannot be considered an absolute principle, and acknowledges a few good exceptions as well, for which he introduces the term antigrammaticalization (Haspelmath 2004: 27). These exceptions, however, as Hopper & Traugott also state, are extremely rare and restricted, so they do not challenge “[t]he basic generalization of unidirectionality” as a strong hypothesis (Haspelmath 2004: 23). To be concrete, Haspelmath describes eight cases of “degrammaticalization”, which he, by definition, interprets as legitimate examples of counterdirectionality:

By this [i.e. antigrammaticalization], I mean a change that leads from the endpoint to the starting point of a potential grammaticalization and also shows the same intermediate stages. For instance, a change from a case suffix to a free postposition with the intermediate stage of a postpositional clitic would be an antigrammaticalization. (Haspelmath 2004: 27–28)

In other words, like Hopper & Traugott (2003), Haspelmath clearly emphasizes the gradualness of a counterdirectional change rather than its instantaneousness. One of the examples he refers to, is the development of clitic =s out of the genitive suffix -s in English and Mainland Scan- dinavian (Haspelmath 2004: 29). Another example involves the change from the English infinitive prefix to- into the proclitic to=, for instance. However, as Norde (2009) correctly remarks, what is striking about these changes, and by extension all counterdirectional changes 16 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Haspelmath lists up (except for one5), is that “none of these cases meets Haspelmath’s own definition of antigrammaticalization” (Norde 2009: 109). After all, these changes “involve only a single counterdirectional shift, and not a series of gradual transitions with the same intermediate stages as observed in grammaticalization chains” (Norde 2009: 109). Norde (2009), in turn, discusses some clear counterexamples as well, but more importantly, she pays way more attention to the subject, and even suggests a detailed typology in her monograph, entirely devoted to the phenomenon of degrammaticalization. This phenomenon has been conceived of as an overarching term encountering three different types of degrammaticalization, all of them falling under Norde’s general definition of the process:

Degrammaticalization is a composite change whereby a gram in a specific context gains in autonomy or substance on more than one linguistic level (semantics, morphology, syntax, or phonology). (Norde 2009: 120)

In the first type, i.e. in degrammation, a function word has been reanalysed as “a member of a major word class” (Norde 2009: 135). It has acquired the morphosyntactic properties typical of that word class, and it has gained in semantic substance. Nevertheless, an important pre- condition has to be fulfilled: the degrammation has to occur “typically in ambiguous contexts” (Norde 2009: 135). If, by contrast, an item would be ‘context-extracted’ (as e.g. in to up the volume), then we are dealing with a case of “lexicalization”, according to Norde, and not with an instance of degrammation. An often cited example of the latter concerns the development of a modal verb into a lexical verb (cf. section 1.2.1, about the (de)grammaticalization of the modal verbs specifically). However, as Norde argues, only a few of these cases may be considered genuine examples of degrammation. The shift from the preterite subjunctive of the modal verb welle ‘to want to’ into the full verb wotte ‘to wish’ in Pennsylvania German is one of them (cf. later on in section 1.2.1.1). Others, however, (such as the development of English dare and need, for instance – often suggested as counterexamples in the literature, cf. again section 1.2.1.1) are no ‘real’ cases of degrammation, according to Norde, since they lack in

5 This concerns the development of the Saame abessive suffix *-ptaken into the free postposition taga through the intermediate stage of clitic =taga. (Haspelmath 2004: 29) LITERATURE 17 semantic change (Norde 2009: 136). Instead, they overwhelmingly express a ‘retraction’ [see Haspelmath 2004: 33] to an earlier, less grammaticalized use, rather than a degrammaticalization from a more to a less grammaticalized form. One should be careful with this view, though, according to Nuyts (2013: 131–132), since the core modals in Dutch (kunnen ‘can’, mogen ‘may’ and moeten ‘must’) have shown some clear ‘lexical uses’ in Present Day Dutch too, without any gain in semantic substance – on the contrary. Moreover, as will be discussed in much more detail later on in this thesis (cf. section 1.2.1.2), these changes can hardly be seen as the result of any “retraction to an earlier, less grammaticalized use”. After all, the lexical (or main verbal) uses in Present Day Dutch are quite different from those in the earlier periods (but, again, see section 1.2.1.2 for a more elaborate discussion). As an alternative, then, Nuyts proposes “not to make definitions of grammaticalization or degrammaticalization dependent on a semantic criterion”, but instead “to judge them purely in terms of what they are in the first place: a change in the grammatical [or] structural status of an element” (Nuyts 2013: 132). Therefore, based on the observation made before in the literature (cf. Traugott 2010) that “grammaticalization and (inter)subjectification do not necessarily coincide”, Nuyts argues it would be better to consider any change from function word to full lexical verb which cannot be shown a case of ‘retraction’ as an instance of degrammaticalization, “no matter what happens semantically” (Nuyts 2013: 132) (cf. section 1.2.1.2). Norde’s second type of degrammaticalization, then, i.e. deinflectionalization, involves a shift from a ‘more grammatical’ to a ‘less grammatical’ item, which nevertheless remains bound. To be concrete, in terms of a definition, “deinflectionalization is a composite change whereby an inflectional affix in a specific linguistic context gains a new function, while shifting to a less bound morpheme type” (Norde 2009: 152). One of the examples Norde refers to, has been discussed by Haspelmath (2004) before, and concerns the development of genitive -s “from inflectional affix to phrase-final (enclitic) determiner” in English and Mainland Scan- dinavian (Norde 2009: 160ff). Other examples involve the shift of Swedish -er from inflectional (case) suffix to derivational (noun) suffix, and the development of Swedish -on as ‘berry-name suffix’ out of a ‘number suffix’. We will not go into a further discussion of the different examples here (since they are less relevant to our study of the modal verbs in Dutch), but we move on immediately to Norde’s third and last type of degrammaticalization. 18 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

The third type, i.e. debonding, is the most frequently attested one of all three types, and may be defined as the change “whereby a bound morpheme in a specific linguistic context becomes a free morpheme” (Norde 2009: 186). Again, as in degrammation and deinflectionalization, in debonding too, Norde argues the importance of the reanalysis of a grammatical item “within the context of [its] own construction”, and the continuation of its previous function, at least until it becomes a free morpheme. Once it has become a free morpheme, however, it “may […] start to appear in other constructions as well” (Norde 2009: 186). To illustrate this, Norde refers to the ‘degrammaticalization’ of “the Northern Saami abessive suffix haga into a postposition meaning ‘behind’”, which subsequently “started to appear independently as an adverb, and as a preposition” (Norde 2009: 186). Other examples of debonding, discussed by Norde, are, e.g., the splitting of the infinitives in English (as in to boldly go where no man has gone before…, cf. Norde 2009: 190ff), the evolution of the Old Estonian ep “from clitic to free particle” (see Norde 2009: 201ff) and the development of the quantifiers tig, tich and zig in respectively Dutch, Frisian and German, out of the homonymous suffixes (cf. Norde 2009: 213ff) – again, we will not go into a further discussion of the examples here; instead we turn to the different characteristics of ‘degrammaticalization’. According to Norde (2009: 120ff), each of the three types discussed above can be characterized by four different features: (i) counterdirectionality, (ii) novelty, (iii) infrequency, and (iv) discontinuity. The first criterion to discover a potential case of degrammaticalization concerns “the direction of the change”. Whereas grammaticalization could be defined as a shift from the left to the right on Hopper & Traugott’s (2003: 7) ‘cline of grammaticality’ (rendered in (4) below), degrammaticalization can be regarded as a shift in the opposite direction on this cline (i.e. from the right to the left).

(4) content item > grammatical word > clitic > inflectional affix

A second criterion in order to qualify a case as degrammaticalization has to do with an item’s novelty. As argued by Norde, one should only be allowed to consider an instance as ‘degrammaticalized’ when a new gram has emerged. If, by contrast, a less grammatical item, “however marginalized”, exists alongside a more grammatical item and at a certain moment in time becomes more frequent again, then, the change should not be considered a case of degram- LITERATURE 19 maticalization, according to Norde. Instead, it should be qualified as a matter of ‘retraction’ (cf. again the case of English dare, mentioned above). In other words, in degrammaticalization, “a less grammatical item must be shown to derive from a more grammatical item” (Norde 2009: 122). A third criterion to identify a case of degrammaticalization concerns its infrequency. Especially when compared to the process of grammaticalization, degrammaticalization is a rather infrequent phenomenon. This may not, however, be a reason, according to Norde, to dismiss the phenomenon as non-existent. The fourth and last criterion relevant to distinguish a case of degrammaticalization, then, is termed discontinuity. It joins and at the same time it expands the first criterion, in that the opposite or counterdirectional change (i.e. the movement from right to left on the ‘cline of grammaticality’) does not go all the way back. Whereas grammaticalization may (but need not) go all the way along the cline, i.e. from content item to inflectional affix, most cases of degrammaticalization entail “a single shift from right to left on the cline” (Norde 2009: 123). This does not mean, however, that subsequent changes are impossible – on the contrary (cf. the case of the abessive suffix haga, for instance), but in general “there is no domino effect” (Norde 2009: 123). In other words, degrammaticalization may not be considered the complete reverse or “mirror image” of grammaticalization (Norde 2009: 112). Although Norde clearly recognizes the process of degrammaticalization, she is also aware of the extreme rarity of examples of it (Norde 2009: 49). Nevertheless, they should not be ignored. So, instead of questioning and discussing how many examples would be sufficient as counterevidence to the unidirectionality claim, it would be better, according to Norde, to accept that changes in both directions exist, although “movement in the direction of greater grammatical status” is the more frequent one (Norde 2009: 52). Unidirectionality, in that way, is better conceived of as general tendency, instead of an absolute principle. Other linguists have voiced much stronger criticisms to the claim of unidirectionality, however. Campbell (2001a), Janda (2001) and Newmeyer (2001), for instance, criticize the fact that the notion of unidirectionality “is essentially built into the [very] definition of grammaticalization”, and as such cannot be considered “an empirical hypothesis that can be tested” (Campbell 2001a: 124). After all, by defining grammaticalization as a change from ‘lexical to 20 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH grammatical’ or from ‘less grammatical to more grammatical’, “any change not going in[to] this direction” clearly falls outside the definition of grammaticalization. Nevertheless, as shown above, counterexamples do exist, and according to the authors mentioned (i.e. Campbell, Janda and Newmeyer), they are even “quite numerous”. This, of course, depends on one’s conception of degrammaticalization, i.e. whether or not examples of what Norde (2009), for instance, has called cases of “lexicalization” are included into the counterdirectional phenomenon or not6, but any how the phenomenon may be conceived of, one cannot deny the existence of some clear counterexamples (cf. above). Their existence compels Newmeyer (2001) to the conclusion that the claim of unidirectionality is false – after all, he takes “any example of upgrading as sufficient to refute [the claim]” (Newmeyer 2001: 205). Nevertheless, as Newmeyer goes on, “it is not all that false”, since downgradings have occurred (estimated somewhat roughly) “at least ten times as often as upgradings” (Newmeyer 2001: 213). Campbell (2001a) and Janda (2001), finally, do not want to go as far as Newmeyer by refuting (the claim of) unidirectionality completely, but they agree that acceptance of the claim “in its strongest form” (i.e. by ignoring the exceptions entirely) would clearly be wrong too (Campbell 2001a: 133). From this section, then, we may conclude that movement towards increasing grammaticality has generally been accepted to be the preferred one, and that unidirectionality has been viewed of as general tendency by most scholars. Points of discussion, however, are whether unidirectionality should be considered an absolute principle or an empirical hypothesis, whether counterexamples exist or not, and, most importantly, if they exist, whether they stand on their own, are part of a countermovement (i.e. degrammaticalization) or refer to other processes such as “lexicalization”. In section 1.2.1, we will consider these notions (of unidirectionality and (de)grammaticalization) in light of the (evolution of the) Germanic modals in general (1.2.1.1) and the Dutch modals in specific (1.2.1.2), but first we will turn to an overview of the semantic developments discussed in the literature.

6 According to Campbell (2001a: 131), e.g., all instances showing a conversion of ‘grammatical into lexical’ are clearly “against the unidirectionality claim, regardless of whether such examples are given a new name (‘lexicalization’) or not”. LITERATURE 21

1.1.2 Modality and (inter)subjectification

As mentioned at the beginning of this chapter, two major concepts are important for our exploration of the re-autonomization process of the modals in Dutch. Next to (de)grammaticalization, as a shift in the structural status of an element, there is also subjectification and intersubjectification, as a change in the semantic status of an item and often said to be connected (though not necessarily coinciding) with the process of “grammaticalization”. Before turning to the latter, we first introduce a number of views on the definition and classification of modality in general (1.1.2.1). More specifically, we discuss (i) the traditional distinction between deontic and epistemic modality as found e.g. in Lyons (1977), (ii) the alternative modal proposed by Bybee et al. (1994), (iii) the influential approach in van der Auwera & Plungian (1998), and (iv) – most important to us – the classification suggested by Byloo & Nuyts (2014), Nuyts & Byloo (2015) and Nuyts (2016). Next we will discuss the diachronic development of modality, i.e. the predictable pathways along which modality develops (1.1.2.2), as well as the interpretation of this in terms of the semantic processes of “subjectification” and “intersubjectification” and their presumed unidirectionality (1.1.2.3). Finally, we will end this section by giving an overview (rather briefly) of the discussion held in the literature about the possible mechanisms involved into the processes of “subjectification” and “intersubjectification”, i.e. metaphor and/or metonymy (1.1.2.4).

1.1.2.1 Notions of modality

Modality is a notoriously difficult notion, and one which has been defined in numerous ways. Traditionally, it is conceived as the semantic domain of possibility and necessity as it is understood in modal logic, and although some scholars have pointed out problems with this approach when applied to language, it is still often adhered to for lack of a better definition. A common view is that modality in language comprises at least two distinct types of meanings, one concerned with truth values and the other with allowing or obliging people to carry out certain actions. The two types typically go under the names epistemic and deontic modality. Whereas epistemic modality concerns the necessary or possible truth of pro- positions, as in (5a–b) (cf. Lyons 1977: 797), deontic modality “is concerned with the necessity or possibility of acts performed by morally responsible agents” (Lyons 1977: 823). Expressions 22 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH of obligation and permission, as in (6a–b) (see Lyons 1977: 839), hence belong to the field of deontic modality under this view:

(5) a. Alfred must be unmarried [epistemic necessity] b. Alfred may be unmarried [epistemic possibility]

(6) a. You must open the door [deontic necessity] b. You may open the door [deontic possibility]

A third type, dynamic modality, is usually distinguished in the literature as well, but there is much disagreement about the precise delineation of this category. Under the most restrictive view, it comprises only expressions of the ability or capacity of an agent, as in (7). This is the view adhered to by Traugott & Dasher (2002), who rename the category ability/capacity but explicitly equate it with what other scholars have called dynamic modality (cf. Traugott & Dasher 2002: 107, fn. 2).

(7) Jane can swim [dynamic possibility]

Some linguists, however, would also consider certain expressions of need or necessity to fall under the notion of dynamic modality, as for instance in (8) (taken from Nuyts 2016: 34).

(8) I must find something to eat or I’ll starve [dynamic necessity]

In addition, expressions of will (‘volition’) and intention are sometimes included under dynamic modality as well (cf. e.g. Palmer 1990, 2001; Warner 1993; Jacobsson 1994). Traugott (1989: 38), on the other hand, considers ‘volition’ a subtype of deontic modality, but this appears to be a minority view (see Nuyts 2016: 37). However the field of dynamic modality is defined, many agree that there is a close connection between deontic and dynamic modality as opposed to epistemic modality. As such, Palmer (2001) groups deontic and dynamic modality together under the heading of event modality, because both types concern the realization of the event (i.e. state of affairs) of the LITERATURE 23 clause. Epistemic modality, on the other hand, which concerns the truth of a proposition, is categorized together with evidentiality as a type of propositional modality. A similar view is expressed by van der Auwera & Plungian (1998: 81ff). Coates (1983) goes even further by not distinguishing between dynamic and deontic modality at all, but by treating them both as part of a larger category of root modality, which she regards as a scalar category with some uses being more subjective than others. Thus, what other scholars term ‘dynamic’ and ‘deontic’ should rather be seen as the less and more subjective ends of a continuum according to Coates (1983: 20ff). A rather different classification of modality is proposed by Bybee, Perkins & Pagliuca (1994), who distinguish between four different types: agent-oriented, speaker-oriented, epistemic, and subordinating modality. It is central to their approach that they view modality not as a coherent notional domain, but as “a set of diachronically related functions” (Bybee et al. 1994: 176). In other words, modality is seen as a number of meanings which are related historically and hence are often expressed by the same forms cross-linguistically. Specifically, the authors take agent-oriented modality to be historically ‘prior’ to the other three types, which develop from it along largely predictable lines (cf. section 1.1.2.2 below). The domain of agent-oriented modality comprises meanings which report “the existence of internal and external conditions on an agent with respect to the completion of the action expressed in the main predicate” (Bybee et al. 1994: 177). This includes not only ability and possibility, but also necessity, obligation, desire, intention, and willingness. It is thus a broader category than dynamic modality as introduced above. Whereas dynamic modality under a more traditional view (exclusively) features a first argument participant as its ‘source’, agent-oriented modality also includes meanings where the first argument participant is the ‘target’ of the modality, as it were, such as the ‘strong’ and ‘weak’ obligation meanings in (9–10) (cf. Bybee et al. 1994: 177; quoted from Coates 1983: 35, 59). Under the traditional view, these would be treated as examples of deontic modality:

(9) All students must obtain the consent of the Dean … before entering for examination

(10) I just insisted … on calling her Miss Tillman, but one should really call her President

24 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

In this way, the notion of agent-oriented modality is comparable to the category of ‘root’ or ‘non-epistemic’ modality in Coates (1983). The notion of speaker-oriented modality, on the other hand, consists of expressions which are directive rather than descriptive. Rather than re- porting the presence/absence of certain conditions on the first argument participant, speaker- oriented modality expresses the speakers’ imposition of such conditions on the addressee (Bybee et al. 1994: 179). Imperatives are thus considered part of the domain of speaker- oriented modality, as are prohibitive, hortative, admonitive, and permissive meanings. It is unclear to us to what extent agent-oriented and speaker-oriented modality may overlap in this framework, however. On the one hand, they are discerned as two separate semantic categories, but on the other, Bybee et al. (1994: 178ff) also write that agent-oriented modality can be ‘used’ in directives such as (11) (from Coates 1983: 88), and they classify directive meanings as speaker-oriented. Therefore, the classification made up here, and more specifically the distinction between both types of modality discussed so far, does not work well in our opinion.

(11) You can start the revels now

It is also worth mentioning that the terminology used by Bybee et al. is potentially confusing. Whereas the term ‘agent-oriented’ (in ‘obligation’ meanings) refers to the ‘target’ of the modal expression – i.e. the participant who is permitted, obliged, etc. – ‘speaker-oriented’ refers to its ‘source’ – i.e. the speaker who allows, demands, etc. The term ‘agent-oriented’ is also somewhat unfortunate because the participant in question is not necessarily an agent. This is precisely the reason why van der Auwera & Plungian (1998) prefer the terms participant- internal and participant-external instead, cf. below. The notion of epistemic modality in Bybee et al. is more in line with the traditional concept. They describe it as an indication of “something less than a total commitment by the speaker to the truth of the proposition” (Bybee et al. 1994: 179), and distinguish between possibility, probability, and inferred certainty, as in (12–14) (cf. Bybee et al. 1994: 180). They briefly mention that epistemicity is closely related to indirect evidentiality, but do not otherwise go into the discussion about the relation between epistemic and evidential meanings.

(12) I may have put them down on the table: they’re not in the door. LITERATURE 25

(13) The storm should clear by tomorrow.

(14) There must be some way to get from New York to San Francisco for less than $600.

The fourth and last domain distinguished by Bybee et al. (1994) is subordinating modality or mood. According to Nuyts (2016), however, this is actually a formal rather than a semantic category, since its distinguishing characteristic is that it only occurs in a particular clause type, namely subordinate clauses. The category includes both inflectional subjunctives, as in Latin and its daughter languages, and uses of modals such as should in (15) (taken from Bybee et al. 1994: 215).

(15) The police are expecting that the Libyans should make the first move

Coates (1983) terms such uses ‘quasi-subjunctive’ and characterizes the meaning of the modal as more or less ‘eroded’ depending on how harmonic it is with the meaning of the matrix clause. While uses like the one in (15) are often not classified as part of the modal domain in the literature, it is clear that there must be a relation between modal and ‘quasi-subjunctive’ meanings, since they are expressed by the same forms in a number of languages. ‘Quasi-subjunctive’ meanings are thus also distinguished by van der Auwera & Plungian (1998) and Nuyts & Byloo (2015), although neither of these studies considers such meanings part of the modal domain (cf. below). The study by van der Auwera & Plungian (1998) builds on Bybee et al. (1994) but departs from their work in important ways. Firstly, they define the field of modality in terms of a paradigmatic contrast between possibility and necessity. This means that expressions of volition, intention, desire, willingness, etc. are not regarded as modal, and that ‘subordinating modality’, as mentioned in the previous paragraph, is not seen as part of the modal domain either. The link between modal and subjunctive-like meanings is acknowledged, however, by treating subjunctive-like uses as ‘postmodal’. Secondly, they take the category of ‘speaker-oriented modality’ in Bybee et al. to be a conflation of deontic modal meanings and meanings which pertain to the illocutionary type of the utterance rather than modality (van der Auwera & Plungian 1998: 83). Hence, imperatives, prohibitives, etc. are not regarded as modal categories under this view. 26 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

To be concrete, van der Auwera & Plungian (1998) distinguish between four different types of modality: participant-internal, non-deontic participant external, deontic participant- external, and epistemic modality. ‘Participant-external’ modality is taken to encompass both a non-deontic and a deontic subtype; hence all deontic meanings are also participant-external, but participant-external meanings need not be deontic, cf. Table 3 (as well as the examples below). The authors do not suggest a shorter term for ‘non-deontic participant-external’ modality.

Table 3. Classification of modality in van der Auwera & Plungian (1998)

1. Participant-internal modality Dynamic possibility/ability/capacity – need

2. Non-deontic: possibility – necessity Participant-external modality 3. Deontic: permission – obligation

4. Epistemic modality Uncertainty – probability

Participant-internal modality refers to possibilities and needs which are internal to the first argument participant (cf. (7–8) before, for instance), whereas participant-external modality refers to external circumstances, as in (16a–b) (see van der Auwera & Plungian 1998: 80).

(16) a. To get to the station, you can take bus 66. [participant-external non-deontic possibility]

b. To get to the station, you have to take bus 66. [participant-external non-deontic necessity]

Deontic modality is also external to the participant, but differs from non-deontic participant- external modality by having ‘some person(s)’ or ‘some social or ethical norm(s)’ as its source, as in (17a–b) (in which “there is some person with authority and/or some norm with respect to whom John’s leaving is possible/necessary”) (cf. van der Auwera & Plungian 1998: 81). This definition is thus in line with the traditional use of the term. LITERATURE 27

(17) a. John may leave now. [participant-external deontic possibility]

b. John must leave now. [participant-external deontic necessity]

Van der Auwera & Plungian’s conception of epistemic modality is similar to the one found in Bybee et al. (1994): they define epistemic possibility as expressions of uncertainty about the truth of the proposition (cf. the example in (12) above) and epistemic necessity as certainty or “a relatively high degree of probability” about its truth (see the examples in (13–14) mentioned before) (van der Auwera & Plungian 1998: 81). Hence, the subtypes ‘probability’ and ‘inferred certainty’ from Bybee et al. (1994) qualify as epistemic necessity in van der Auwera & Plungian’s terms. Evidentiality is not regarded as a subtype of modality by van der Auwera & Plungian, with the exception of inferential evidentiality, illustrated in (18), which is treated as an overlap category belonging to both epistemic modality and evidentiality (see van der Auwera & Plun- gian 1998: 85).

(18) Ahmet must have come. [inferential evidentiality = epistemic necessity]

Finally, an even more fine-grained classification of modality is employed by Byloo & Nuyts (2014) and Nuyts & Byloo (2015) in two studies of the Dutch modal verbs, as well as in an overview article by Nuyts (2016). This classification is similar in many ways to the one proposed by van der Auwera & Plungian (1998), but differs in at least two important respects. Firstly, the term dynamic is used to cover the participant-internal and participant-external non-deontic subtypes in van der Auwera & Plungian (cf. types 1 and 2 in Table 3), but importantly, three rather than two dynamic subtypes are distinguished. Secondly, the term deontic is redefined to encompass only notions of moral or ethical acceptability, and not expressions of permission and/or obligation. Such expressions are termed directive and are taken to be illocutionary rather than modal. Nuyts & Byloo would thus agree with Bybee et al. (1994) that You can start the revels now (cf. example (11) above) is used directively, but they would consider it to fall outside the field of modality proper. 28 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Since our investigation of the semantic development of the Dutch modal verbs relies on the classification of Byloo & Nuyts (2014) and Nuyts & Byloo (2015), we will not go into a further discussion of the different semantic categories here; they will be treated extensively in section 3.2.1. Instead, we now turn to the diachrony of modality, and the notions of subjectification and intersubjectification.

1.1.2.2 The diachrony of modality

As the preceding section should have made clear, there are several competing definitions and classifications of modality. Nevertheless, there seems to be agreement in the literature that the diachronic development of modal meanings is predictable to some extent. Specifically, epistemic meanings are assumed to be secondary to deontic, root, or agent-oriented modality, the terminology varying according to the framework in question. This notion of predictability (i.e. of unidirectional modal development) has been around in the historical linguistic literature at least since the early 1980s. In a short study based on data from Antiguan Creole and English child language acquisition, Shepherd (1982) formulates the hypothesis strongly: “If a particular modal form changes in meaning (with respect to deonticity or epistemicity) the direction of the change will be from deontic to epistemic, not from epistemic to deontic” (Shepherd 1982: 316). In the same volume, Goossens (1982) investigates the development of epistemic modals in the history of English. While he finds ‘traces’ of epistemic usage of a few modals in Old English (specifically, the ancestors of may, shall, and will), the conclusion is that this semantic field only developed fully after the Old English period. Unlike Shepherd’s and Goossens’s studies, Bybee et al. (1994) include data from a larger number of languages from around the world. While they find that modal expressions may develop from many different lexical sources, the semantic pathways within the modal domain are largely predictable. Agent-oriented meanings may develop into any of the other three types directly, and speaker-oriented and epistemic meanings may develop into subordinating modality (cf. Figure 2). The reverse developments are not attested in the data. The ‘semantic maps’ in Bybee et al. are elaborated on by van der Auwera & Plungian (1998), who also find that epistemic meanings are secondary to agent-oriented ones (specifically, participant-external meanings in their terminology). LITERATURE 29

Figure 2. Pathways of modality (Bybee et al. 1994: 241)

1.1.2.3 (Inter)subjectification and unidirectionality

The supposedly unidirectional development to epistemic meanings (discussed in the previous section) is linked to a general linguistic tendency to increased subjectivity by Traugott (1986, 1989) and Hanson (1987). In addition, these authors claim that the development from less to more subjective is also found within the epistemic domain itself. While Hanson (1987) primarily deals with the history of epistemic adverbs in English, Traugott (1989) also includes speech act verbs and modals in her analysis; we will only concern ourselves with the latter here. Traugott (1989) sees the two developments – from non-epistemic to epistemic and from ‘weakly’ to ‘strongly’ subjective epistemic – as instantiations of the same diachronic tendency, and uses the term subjectification from Langacker (1985) for it. Traugott defines it thus: “Meanings tend to become increasingly based in the speaker’s subjective belief state/attitude toward the proposition” (Traugott 1989: 35). An example of the two degrees of subjectivity within the epistemic domain is taken from Lyons (1982: 109):7

(19) You must be very careful a. It is obvious from evidence that you are very careful [weakly subjective reading] b. I conclude that you are very careful [strongly subjective reading]

According to Traugott, the history of the modals shall, will, and must provides evidence that weakly subjective epistemic meanings always come before strongly subjective ones. For

7 Traugott also provides examples of more and less subjective deontic modality (also from Lyons 1982). But since Traugott’s paper is primarily about developments within the epistemic domain, we restrict ourselves to the epistemic examples here. 30 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH instance, to the extent that must is found epistemically in Old English – and Traugott is not sure that it is – it is always only weakly subjective; but it then develops strongly subjective epistemic uses during the Middle English period. And while shall is attested in Old English as a hearsay evidential as in (20), which Traugott (1989: 41) classifies as a weakly subjective epistemic meaning, it is never found with a strongly subjective meaning.

(20) & to Þam Pentecosten […] wæs gesewen blod weallan of eorþan. swa swa mænige sæden Þe hit geseon sceoldan. ‘and at the Pentecost … blood was seen welling up from the ground, as many said who supposedly saw it’

It is worth noting, however, that Traugott does not actually show an unambiguous development from weakly to strongly subjective epistemic in any of the three individual modals. Shall is not argued to have ever developed a strongly subjective sense, and the strongly subjective uses of will which are said to occur from the 19th Century, as in (21) (see Traugott 1989: 43), are not argued to have developed from less subjective epistemic uses, but from habitual or future uses.

(21) This will be your luggage, I suppose

For must, the problem is that Traugott is not sure that it is actually attested in weakly subjective epistemic uses in OE – she uses the otherwise undefined term ‘nondeontic’ in one place (1989: 42) – and the analysis of the Middle English data is unclear. It is argued that the strongly subjective epistemic sense only developed in contexts with a ‘strongly epistemic adverb’ such as nedes in (22) (cf. Traugott 1989: 42), and that must alone was thus only weakly subjective at the time.

(22) He that dooth good & doth not goodly … must nedes be badde ‘Whoever does good, but does not do it with good intentions … must necessarily be bad’

LITERATURE 31

However, Traugott’s characterization of the adverbs as epistemic and her gloss ‘without doubt’ (Traugott 1989: 42) are problematic. The Middle English Dictionary (MED) glosses the adverb nedes as ‘necessarily, inevitably, unavoidably’, and the dictionary contains numerous examples where must and nedes co-occur without an epistemic meaning, as in (23).

(23) I wyll go do Þat nedys must be don. ‘I will go and do that which necessarily has to be done’ (our translation)

Table 4 gives an overview of the development of weak and strong subjectivity in shall, will, and must according to Traugott.

Table 4. Subjectivity in three epistemic modals according to Traugott (1989)

Old English Middle English Modern English

shall weakly subjective no epistemic uses no epistemic uses

will no epistemic uses no epistemic uses strongly subjective

must (weakly subjective?) strongly subjective only with strongly subjective nedes etc.

We also note, however, that Traugott’s categories are not particularly well-defined and that the criteria for deciding between a ‘weakly’ or ‘strongly’ subjective meaning are unclear. The weakly subjective epistemic sense is paraphrased as “It is obvious from evidence that”, and the strongly subjective sense as “I conclude that” (Traugott 1989: 36; cf. example (19) above). But either of these paraphrases seems appropriate to us in a number of cases, for instance in (24) (from Traugott 1989: 42).

(24) the fruit muste be delicious, the tree being so beautiful

Traugott (1989: 42) classifies this as a strongly subjective epistemic use, but it seems to us that the subjectivity of the statement primarily lies in the evaluative adjective beautiful rather than 32 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH the modal. In fact, the subjectivity of the modal could be argued to be rather weak here – based on the available evidence (which happens to be a subjective opinion), it is obvious that the proposition ‘the fruit is delicious’ must be true. The history of must is investigated in more detail by Traugott & Dasher (2002). Unlike in Traugott (1989), the Old English verb is not analysed as primarily deontic, but as having a ‘premodal’ meaning of ability or permission, cf. (25), which then developed a deontic necessity meaning in late Old English and early Middle English (26), and finally unambiguous epistemic uses in ‘middle’ Middle English (27) (The examples are taken from Traugott & Dasher 2002: 122, 124, 129):

(25) Ic hit Þe Þonne gehate Þæt Þu on Heorote most sorhleas swefan ‘I promise you that you will be able to sleep free from anxiety in Heorot’

(26) we moton eow secgan eowre sawle Þearfe, licige eow ne licige eow ‘we must tell you about your soul’s need, whether it please you or not’

(27) For yf that schrewednesse makith wrecches, than mot he nedes ben moost wrec- chide that lengest is a schrewe. ‘For if depravity makes men wretched, then he must necessarily be most wretched that is wicked longest’

As in Traugott (1989), the deontic and epistemic meanings are assumed to have become more subjective over time. The same tendencies, from deontic to epistemic and from less to more subjective within each of these domains, are observed for English ought to in a shorter chapter (Traugott & Dasher 2002: 137–143). The development of these modals is taken as evidence for the unidirectionality of subjectification.8 In addition to subjectification, Traugott & Dasher distinguish another diachronic tendency, intersubjectification, which is also assumed to be unidirectional. Traugott & Dasher’s notion of intersubjectivity, which is inspired by Benveniste (1971), involves the speaker’s attention to the

8 We will not go further into the specifics of must in Old and Middle English here, but it should be noted that there is much uncertainty about how to classify it semantically in Old English (cf. Traugott & Dasher’s hesitant analysis of ex. (4) on p. 123 and their repeated use of quotation marks around “ability”). LITERATURE 33 addressee and is defined as the coded expression of this attention, e.g. by “hedges, politeness markers, and honorific titles” (Traugott & Dasher 2002: 23). Intersubjectification, then, is the development of such expressions out of subjective ones as described for the development of Japanese honorifics in chapter 6 of their textbook. As far as we can see, Traugott & Dasher do not invoke this notion of intersubjectification in their discussion of the English modals, despite a brief reference to a ‘particular kind of (inter)subjectivity’ in epistemic modals on p. 115. Following Sweetser (1990) and van der Auwera & Plungian (1998), they suggest that concessive uses of modals as in (28) (see Traugott & Dasher 2002: 115) may be analysed as intersubjective because they involve a qualification of something another speaker has said. They do not go further into this use, but remark that the textual record does not provide enough evidence to investigate exactly how it may have arisen in English.

(28) She may jog, but she sure looks unhealthy to me.

It thus appears that Traugott & Dasher’s notion of intersubjectification is only marginally relevant to the development of the (English) modals. However, there are several competing definitions of intersubjectivity and intersubjectification in the literature, some of which have more bearing on the analysis of the modals. Here, we will concern ourselves with the notions as they are understood by Byloo & Nuyts (2011, 2014) and Nuyts & Byloo (2015). Like Traugott (1989) and Traugott & Dasher (2002), Byloo & Nuyts distinguish between subjectification and intersubjectification, the definitions of which are comparable to the ones in Traugott’s work. However, the ways of characterizing the two notions display some differences between the authors. Byloo & Nuyts raise the criticism that the notions of subjectification and intersubjectification as used by Traugott (et al.) are formulated “somewhat vague[ly], relying heavily on intuition” and hence are “vulnerable to multiple interpretations” (Byloo & Nuyts 2014: 92). In order to make the notions more “objectively reconstructible and testable”, Byloo & Nuyts suggest to reinterpret them in terms of a “hierarchy of qualificational categories”, which offers an account for “the relative extension of semantic scope”, and of which the categories involved could be represented as in Figure 3 (cf. Byloo & Nuyts 2011: 11). Note that not all possible categories are shown in this scheme, only those relevant to the modals, cf. section 3.2.1. 34 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

> (inferential) evidentiality > epistemic modality > deontic modality/boulomaic modality > time > quantitative aspect/dynamic modality > phasal aspect > state of affairs

Figure 3. Hierarchy of qualificational categories (Byloo & Nuyts 2011: 11)

To get an idea of what the hierarchy precisely involves, we refer to the different illustrations discussed in Byloo & Nuyts (2014: 92ff) and Nuyts & Byloo (2015: 39ff), which aim to clarify the correlation between “climbing up the scheme” and “widening of the scope”. To state it briefly: “the higher up in the scheme, the wider the scope of the categories” (Byloo & Nuyts 2014: 93). More important to us, however, is the relevance of this hierarchy for the understanding of the notions of subjectification and intersubjectification. As pointed out by Byloo & Nuyts, the scheme not only concerns the semantic scope of the categories; it also reflects “the relative position of the different categories in our conceptualization of the world” (Byloo & Nuyts 2014: 93). So, climbing up the hierarchy also encompasses “a gradual widening of the perspective on the SoA”, which, correspondingly, implies “an increasing role for the qualifying subject, in terms of interpreting the situation” (Byloo & Nuyts 2014: 93). To put it differently: “the higher up in the scheme, the more the subject has to interpret the situation to qualify the SoA” (Nuyts & Byloo 2015: 41). In this way the qualificational hierarchy “also reflects increasing subjectivity”, and as such Traugott’s notion of subjectification can be redefined as “the diachronic process by which the meaning of a linguistic element gradually climbs up in the hierarchy” (Nuyts & Byloo 2015: 41). The notion of intersubjectification, then, can be reconsidered “in the context of the distinction between conceptual and communicative dimensions of language use” (Nuyts & Byloo 2015: 41). After all, the regulation of interaction has nothing to do with the “conceptual representation of world knowledge”, but all with the “planning of communication acts about that knowledge with others” (Byloo & Nuyts 2014: 93). In that way, intersubjectification can be reinterpreted “as the process whereby a linguistic element ‘semantically leaves’ the conceptual LITERATURE 35 hierarchy, to assume a function in the realm of interaction and discourse planning” (Nuyts & Byloo 2015: 41). This includes politeness markers as in Traugott & Dasher (2002), but also illocutionary markers like directives and imperatives, and ‘textual’ markers like clause connectors. In Table 5, a number of intersubjective categories relevant to the modals are listed in the right column. The left column shows the hierarchy of qualificational categories again, cf. Figure 3 above.

Table 5. Subjective and intersubjective dimensions (Nuyts & Byloo 2015: 42)

CONCEPTUAL REPRESENTATION COMMUNICATIVE PLANNING

> (inferential) evidentiality ILLOCUTIONARY DIMENSIONS: > epistemic modality volition > deontic modality/boulomaic modality intention > time directivity

> quantitative aspect/dynamic modality TEXTUAL DIMENSIONS: > phasal aspect condition > state of affairs concessivity

As the table shows, Nuyts & Byloo (2015) consider epistemic modality more subjective than deontic modality (which, in turn, is more subjective than e.g. tense and dynamic modality). Un- like Traugott (1989) and Traugott & Dasher (2002), however, they do not distinguish between different ‘strengths’ of subjectivity within epistemic modality. Another difference between the authors is the relation between subjectification and intersubjectification, which Traugott & Dasher (2002) view as two parts of the same diachronic process, with intersubjective meanings as the endpoint, as depicted in Figure 4.

Figure 4. Subjectification and intersubjectification in Traugott & Dasher (2002) 36 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

According to Nuyts & Byloo (2015), on the other hand, intersubjective meanings do not have to develop out of highly subjective ones, which means that intersubjectification is not seen as the ‘last stage’ of subjectification, but as a different process, cf. Figure 5. Intersubjective meanings may develop out of meanings at any point on the vertical axis from least to most subjective. In fact, Byloo & Nuyts (2014) find that intersubjective meanings (cf. e.g. Je kan gaan ‘you can go’, meaning ‘[I want you to] leave’) seem to develop primarily out of dynamic modality in the Dutch core modals (and never out of epistemic modality).

Figure 5. Subjectification and intersubjectification in Nuyts & Byloo (2015)

1.1.2.4 Mechanisms of change: metaphor and metonymy

How exactly a meaning shift such as that from dynamic to epistemic modality or directivity happens, is a question that has received some attention in the literature. While some authors, such as Sweetser (1990) e.g., assume that ‘processes of metaphorization’ are foundational to diachronic meaning shifts, other linguists, like Traugott (1989) and Goossens (1999) e.g., believe that such developments should be explained by conversational implicatures, hence ‘metonymic processes’ (Byloo & Nuyts 2011: 13). Traugott & Dasher (2002) argue for what they term the ‘Invited Inferencing Theory’, which stresses the importance of polysemy. Byloo & Nuyts (2011, 2014) and Nuyts & Byloo (2015) make a case for a similar conception by pointing out that although a metaphorical transfer could be conceivable at a macro-level (i.e. “as a hindsight at source meaning and target meaning”, cf. Byloo & Nuyts 2011: 13), only a metonymic shift can explain the processes of subjectification and intersubjectification. The idea that metonymic meaning shifts are crucial can be illustrated with the Dutch example in (29) (taken from Nuyts & Byloo 2015: 48). LITERATURE 37

(29) Iets minder zon morgen, en er kan een buitje vallen ‘Somewhat less sunshine tomorrow, and there can/may be a shower’

Nuyts & Byloo (2015) classify this example as ambiguous between dynamic modality (specifically, the situational subtype, cf. section 3.2.1) and epistemic modality. Under the dynamic interpretation, kan expresses a potential inherent in the situation – it is physically or logically possible that it is going to rain tomorrow. Under the more subjective epistemic interpretation, it expresses the speaker’s evaluation of the likelihood that it is going to rain (which can be paraphrased with ‘perhaps’). The idea is that the dynamic interpretation is the original one, and that the possibility of an epistemic reading has developed through the conventionalization of a conversational implicature: by saying that it is logically possible that it is going to rain tomorrow, the speaker presumably intends to express that he or she considers it somewhat likely. By systematically studying ambiguities like this one, it is possible to get a better picture of the specific meaning developments rather than merely an overview of when certain meanings are first attested (which, of course, is valuable too). Inspired by Byloo & Nuyts (2014: 107–113), we also intend to take into account the phenomenon of ambiguity in our own semantic analyses (cf. section 3.2.1).

1.2 Language-specific overview

Before turning to the design of our present corpus study, however, let us first consider what the literature says about the role of (de)grammaticalization (1.2.1) and (inter)subjectification (1.2.2) in the diachrony of the modal verbs in Germanic (1.2.1.1 resp. 1.2.2.1), and in Dutch in particular (1.2.1.2 resp. 1.2.2.2). Section 1.2.3 summarizes the main findings relevant for the present corpus study.

1.2.1 (De)grammaticalization in the modal verbs

1.2.1.1 The modals in Germanic9

The modal auxiliaries in various Germanic languages, especially English, have figured prominently in the literature on grammaticalization. To give just a few examples: Hopper &

9 This subsection was largely written in collaboration with Sune Gregersen (University of Amsterdam). 38 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Traugott (2003) characterize the development of the English modals as a clear instance of grammaticalization; Diewald (1997) uses the German modals to illustrate the grammaticalization phenomenon itself; and in the cross-linguistic study by Bybee et al. (1994), Danish is included to represent the Germanic languages and is argued to show a number of grammaticalization paths common to modal auxiliaries, such as e.g. skulle ‘owe’ > ‘must, shall’ and kunne ‘know’ > ‘can’ (both of which have parallel developments in English, Dutch, and other Germanic languages). A synchronic overview of the modals across the Germanic language family is given by Mortelmans et al. (2009), who use Lehmann’s parameters of paradigmaticity, paradigmatic variability and structural scope to analyse their degree of grammaticalization in present-day Dutch, German, Danish, and Icelandic. They include a few references to the historical developments in these four languages, but do not carry out any systematic investigation of diachronic data. Diewald (1999) is a monograph entirely devoted to the diachronic development of the German modals, but otherwise most book-length studies present synchronic analyses (cf. e.g. Palmer 1979 and Coates 1983 on English; Öhlschläger 1989 on German; Brandt 1999 on Danish; Eide 2005 on Norwegian). The English modals, on the other hand, have received substantial attention from historical linguists. This interest was, to a large extent, sparked by Lightfoot (1979), who suggested that the development of modal auxiliaries out of full verbs was a case of ‘radical’ syntactic reanalysis. This analysis provoked responses from a number of scholars who argued that the diachronic development of the modals is better explained in terms of grammaticalization; most prominently, Plank (1984: 308) argues that they are a ‘paradigm case’ of the phenomenon (cf. Lehmann 1995: 27–28 as well). Grammaticalization accounts have also been suggested for other auxiliaries and modal expressions in English, such as the ‘emerging’ modals have to, have got to, and want to (Krug 2000), and the development of epistemic uses of parenthetical I think (Aijmer 1997). The history of the English modals has also been approached from other theoretical angles. Warner’s (1993) monograph on the modals in Old and Middle English is primarily a formal- structural account and focuses on the establishment of the modals as a separate syntactic category. However, Warner also includes a short section on grammaticalization (1993: 195– LITERATURE 39

197), which he cautiously suggests may have been formally rather than semantically motivated in the case of the modals: “one is bound to suspect that the cohesion of formal properties … takes precedence over the semantics” (Warner 1993: 127). Warner does not explicitly state whether he considers grammaticalization ‘epiphenomenal’ or a mechanism in its own right, but his short discussion suggests that he does not consider it to have much explanatory potential. Regarding the unidirectionality question, there have also been suggestions that some modals in English – as well as other Germanic languages – do not conform to the expectations about the pathways of grammaticalization. A number of counterexamples, variously termed de-auxiliarization, degrammaticalization, de-modalization, etc… have been proposed. To begin with English, the verb dare has been argued on several occasions to have moved from a more to a less grammatical status. Beths (1999) gives a number of arguments for why the development of dare is a counterexample to the unidirectionality hypothesis, and Schlüter (2010) comes to the same conclusion after a detailed investigation of the development of the verb in a corpus of 16th- to 19th-Century English. The tenor of the argument is that a number of formal changes have happened to dare which have made it less auxiliary-like, such as the development of weak (i.e. regular) instead of preterite-present (i.e. irregular) morphology, and the use of the to-infinitive, which starts to occur in the Early Modern period.10 Similarly, the equivalent of dare in Danish has also been proposed as an example of de- auxiliarization. On the Danish verb, Andersen (2008) writes the following:

In the 1900s, Da. turde ‘dare’ is reanalysed as a lexical verb. It comes to compete with the lexical vove ‘dare’, which it tends to replace. As a lexical verb, it comes to combine with at-infinitives and to receive full stress. (Andersen 2008: 22)

This analysis is similar to the English ones cited above: the change from bare infinitive to at- infinitive (corresponding to the West Germanic to/te-infinitive) and the strengthening of the

10 These formal changes to dare had in fact already been pointed out by Nagle (1989) and Warner (1993) (and, even earlier, in an unpublished PhD dissertation by Reed [1981]). However, these studies were written before the question of unidirectionality became a contested issue in the grammaticalization literature, and hence they do not attach much significance to the fact that the verb seems to have become more ‘lexical’. Taeymans (2004) does discuss the unidirectionality question in her study of dare in Present-Day English, but does not come to any definite conclusion about the matter. 40 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH stress pattern is taken as an indication of decreased grammatical status. A claim based on semantics is put forth in the most recent grammar of Danish. Hansen & Heltoft (2011) refer to turde as a “former modal verb” and write that the verb has now lost its earlier modal uses and only has the meaning ‘dare’, which the authors do not consider modal (Hansen & Heltoft 2011: 778). However, neither Andersen (2008) nor Hansen & Heltoft (2011) back up their claims about the development of turde with any corpus data, so the interpretations rely entirely on the intuitions of the authors. Burridge (1998) suggests that the verb wotte in Pennsylvania German (specifically, the variety spoken in the former Waterloo County in Ontario) is a degrammaticalized past subjunctive form of the modal welle (Standard German willen). The verb in the present-day language has the meaning ‘wish, would like’, as in (30):

(30) Pennsylvania German Ich wott, du kennscht frieher kumme ‘I wish you could come sooner’

According to Burridge, the new verb wotte has split off from the original modal verb willen in the context of “modest wish”, as in (31), but has now lost all traces of its earlier modal status. Hence, for instance, it does not take infinitival complements (*Ich wott kumme ‘I want to come’) and has acquired a new past participle form, gewott (Burridge 1998: 28–29).

(31) Modern Standard German Ich wollte, ich wäre zu Hause ‘I wish I was home’

Burridge also proposes a somewhat controversial explanation for this development, namely that it is due to socio-cultural factors: because of strong social mores against immodesty and impudence in the Mennonite community, the speakers of Pennsylvania German have adopted this more ‘modest’ way of expressing wishes rather than the original verb winsche (cognate of German wünschen). It should be noted, however, that the specific use of willen which wotte is supposed to have developed from (and which is still found in present-day standard German in LITERATURE 41

(31)) is actually not auxiliary in nature, but a transitive verb with a complement clause as direct object, and that the past participle form gewollt also exists in Standard German. In addition, as Börjars & Vincent (2011) point out, the analysis is based entirely on a reconstruction of the relevant steps as there is no historical evidence available for this variety of Pennsylvania German. For this reason, the authors profess to “remain sceptical” about the example. Börjars & Vincent (2011) also mention another example which has been cited in the literature: the apparent development in Middle Swedish of a (regular) full verb må ‘feel’ out of the (preterite-present) modal må (the cognate of English may and Dutch mogen). This example was proposed by van der Auwera & Plungian (1998: 105; cf. also van der Auwera 2002: 23–24). However, as Ziegeler (2003: 243) points out, these two verbs might also have developed out of an original lexical source, and this is indeed what Andersson (2008) finds in his study of må in Old and Middle Swedish: the full verb må meaning ‘feel’ (as well as with the now obsolete meaning ‘have power’) is attested throughout the entire history of Swedish, hence there is no evidence that the development was modal > full verb rather than the other way around. In fact, as the verb is attested with the meaning ‘feel’ in Old Icelandic, Old English, and Old High German as well, it seems reasonable to assume that this meaning goes back to Proto-Germanic (Andersson 2008: 27–28). The only change that has happened is that the full verb variant has acquired regular instead of preterite-present morphology, which Andersson (2008: 24–26) shows happened in the Early Modern period, several centuries after the first attestations of the ‘feel’ meaning. Hence, the Swedish development is better classified as a case of ‘retraction’. An example from the history of German is suggested by van der Auwera & Plungian (1998: 105), namely the development of a transitive verb ‘like’, as in (32), out of the modal mögen (another cognate of English may and Dutch mogen).

(32) Modern Standard German Ich mag Tom nicht. ‘I don’t like Tom.’

However, this example has been questioned too. As already shown by Lühr (1987: 273) and further discussed by Diewald (1999: 315–316) and Ziegeler (2003: 244–245), the transitive use is found in the earliest available sources, as in (33). German mögen, then, just like Swedish må, 42 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH appears to continue an older lexical use alongside the modal one, and there is no evidence that the former developed out of the latter.

(33) Old High German Der stárchemo féhe gíbet sîne fuôra. Er uuêiz die starchen die das héuue múgen. ‘who gives nourishment to the strong cattle. He knows the strong ones, who like hay.’

Finally (and in the light of this thesis, most importantly), the development of the Dutch modals has been argued to show counterdirectional tendencies by Nuyts (2011, 2013) and Nuyts et al. (2018). Unlike the proposed examples from other Germanic languages discussed above, the Dutch development in question is not limited to a single verb, but has been proven to affect a whole group of modals, which appear with modal meaning – but without a main verb – in examples like (34) (taken from Nuyts 2013: 127).

(34) Present Day Dutch vind ie dat bepaalde dingen mogen of niet mogen? ‘Are certain things acceptable or not acceptable to him?’

There thus appears to be a return (no ‘retraction’) to a main verbal status in the Dutch construction, but with retention (and even expansion) of the modal meanings. Since they are the direct motivation for the present corpus study, we present the developments as observed in the studies by Nuyts (et al.) in more detail in sections 1.2.1.2 (the grammatical evolution) and 1.2.2.2 (the semantic evolution). But before doing so, we should mention one final observation regarding the Germanic modals which is potentially relevant for the analysis of the evolution in Dutch. It has repeatedly been pointed out that in many Germanic languages the modal verbs can be used independently provided there is a directional element in the clause. Although this pattern is no longer possible in Present Day English, in Old, Middle and even Early Modern English, combinations of the kind in (35a–b) were quite common (examples from Visser 1963: 165, 163, respectively; see also Mitchell 1985: 415–421). Moreover, they are normal in the LITERATURE 43 present day Scandinavian languages, as illustrated in (35c–e) (respectively from Svenska Akade- miens grammatik 1999: 470, Faarlund et al. 1997: 527, and Hansen & Heltoft 2011: 809; cf. also Hansen 1972, and Brandt 1999: 64–76), and in current German, as exemplified in (35f) (from Mortelmans et al. 2009: 29).

(35) a. Middle English Wrechyd sowle, Þou muste to helle. [lit:] ‘Wretched soul, you must [go] to hell.’

b. Early Modern English Thou shalt with me to Iona. [lit:] ‘You shall [go] to Iona with me.’

c. Present Day Swedish Han måste till stan. [lit:] ‘He must [go] to town.’

d. Present Day Norwegian Hun skal hjem. [lit:] ‘She must [go] home.’

e. Present Day Danish Du skal hjem. [lit:] ‘You must [go] home.’

f. Present Day German Er muss nach Hause. [lit:] ‘He must [go] home.’

This kind of use is often linked to ‘strong’ modals (expressing meanings such as need, necessity, inevitability, requirement, obligation, etc.), including must and shall and equivalents in the other languages, while reports on ‘weak’ modals allowing this type of use (i.e. modals expressing meanings such as ability, potential, possibility, acceptability, permission, etc.), like 44 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH can or may and equivalents, are much rarer. In the Scandinavian languages, for instance, the ‘weak’ forms seem to exclude this kind of use entirely (for unknown reasons, cf. Hansen & Heltoft 2011: 812–813). But, as the Present Day Dutch example in (36) (from Mortelmans et al. 2009: 25) indicates, this may not be considered a cross-linguistically absolute rule.

(36) Present Day Dutch Hij mag naar huis. [lit:] ‘He may [go] home.’

It will be relevant to investigate whether directionals may also play a role in the emergence of the independent uses of the Dutch modals of the kind in (34) above, even if the latter do not require a directional.

1.2.1.2 The modals in Dutch

1.2.1.2.1 The central modals

The grammatical evolution of the three most ‘central’ modals, kunnen ‘can’, mogen ‘may’ and moeten ‘must’ has been investigated in Nuyts (2011, 2013, 2014). These studies are based on diachronic corpus data, and focus on four stages in the history of Dutch: Old Dutch (OD, before 1150), Early Middle Dutch (EMD, 1200–1300), Early New Dutch (END, 1550–1650) and Present Day Dutch (PDD, after 1950), subdividing the latter into written (PDDW) and spoken (PDDS) language. In line with the typical scenario in other languages, the Dutch modals have developed from full verbs through grammaticalization. The earlier studies have shown that the timing of this evolution differs between the individual modal verbs, though. In mogen and moeten, the process of grammaticalization was already completed in the oldest available documents. There may be some traces of their original main verbal uses even today, in uses of the type ik mag soep [lit:] ‘I may soup’ [i.e., I eat soup] or ik moet hem niet [lit:] ‘I must him not’ [i.e., I cannot stand him] (cf. the discussion of (32)–(33) above), but these have undergone meaning change. Mogen originally meant something like ‘have power’, and moeten (most probably) something like ‘measure’ (see e.g. Van Wijk 1912: 437–438, and De Vries 1971: 450–451). An evolution to the meanings ‘like’ and ‘stand’ is not that hard to LITERATURE 45 imagine, though: if something is within one’s power, one can handle it, hence in a way stand it; and if one can measure something, it is within one’s reach, hence at least in some sense manageable. Instances of this type are, however, very marginal in the data of Nuyts – already in OD and EMD, both mogen and moeten are nearly exclusively auxiliary verbs. In kunnen, however, the evolution towards an auxiliary verb is still in full progress in OD and EMD. Auxiliary uses are already dominant in those periods, but the original main verbal use, meaning ‘to know’, remains clearly present too. And it still exists today, although rather marginally, especially in constructions of the type hij kan Spaans [lit:] ‘He can Spanish’ [i.e., he knows Spanish] or ik kan die liedjes uit mijn hoofd [lit:] ‘I can those songs of my head’ [i.e., I know these songs by heart]. As mentioned, the above developments fit into a classic pattern of grammaticalization. But in the transition from END to PDD, the three central modals start to develop in a (from the perspective of a grammaticalization scenario) less expected way: they all show a (sometimes very drastic) increase of autonomous uses again. Some of these may be considered to involve main verb elision, but in others the modal verb itself has to be considered the main verb of the clause. (See section 3.1.3 for a detailed discussion of the types of autonomous uses.) It is remarkable that these main verbal instances (called the ‘new’ main verbal uses) are grammatically very different from the original main verbs from which the auxiliaries have emerged (i.e. the ‘old’ main verbal uses). The ‘old’ main verbs are transitive and feature a first argument expressing an entity (very often an animate or human being), as illustrated by (37) (taken from Nuyts 2014: 363). The ‘new’ main verbal uses, illustrated in (38) (see Nuyts 2013: 127), by contrast, are nearly always intransitive and their single argument refers to a state of affairs (in (38) by means of a relative clause with an embedded antecedent, wat u doet ‘what you are doing’).

(37) Dat voel liede sijn die en geen latijn en conen noch en verstaen. ‘There are many people who do not know [lit: ‘who can no’] Latin nor do understand it.’

46 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(38) En ik zei, dat gaat me niet aan, maar wat u doet kan helemaal niet, een klooster bouwen op het grootste joodse kerkhof ter wereld! ‘And I said, that’s none of my business, but what you are doing is not tolerable [lit: ‘what you are doing cannot’] at all, to build a monastery on the largest Jewish cemetery in the world.’

In all three central modals the upsurge of autonomous uses in PDD is stronger in the spoken than in the written data. In PDDS it is significant in all three verbs, and most prominent in mogen. In PDDW the rise is only significant in kunnen. Overall, the evolution is weakest in moeten. The fact that the autonomous uses are more prominent in the spoken language probably has to do with the more progressive nature of this variant, which therefore reveals ongoing tendencies more clearly than the written language (but cf. section 4.2.1.1). On the basis of these observations, then, Nuyts (2013: 129) concludes that the three central modals in Dutch are currently witnessing a tendency towards re-autonomization in which analogy plays a crucial role. The fact that the tendency is observable even in the more ‘conservative’ written language data of kunnen suggests that this verb is “leading the way for the two other modals.”

1.2.1.2.2 The less central modals

The three ‘less central’ modals, hoeven ‘need’, zullen ‘shall/will’ and willen ‘want’, have been studied in Nuyts et al. (2018, 2019). The setup of these studies is identical to that of the investigations of the three central modals, and was aimed to see whether these less central modals show the same diachronic evolution. The data showed that the three verbs behave quite differently. The situation in hoeven is quite complicated (see Nuyts et al. 2018). This verb only emerged as a modal auxiliary in the transition from EMD to END, out of the original main verb behoeven, meaning ‘to need’. It immediately assumes a predominantly auxiliary status, although the original main verbal use is still prominent as well in END. So, in END, hoeven is probably still mainly a phonologically or morphologically shortened variant of main verbal behoeven, though with more emphasis on the auxiliary use. In the transition to PDD, then, hoeven continues on its grammaticalization path towards a purely auxiliary verb, with the LITERATURE 47 share of main verbal instances drastically decreasing, at least in PDDW. But in PDDS the share of main verbal uses is suddenly substantially higher again. Most main verbal instances in PDD are very different from those in the earlier stages, though: while in END they are still mainly of the ‘old’ transitive type, in PDD most of them are of the ‘new’ intransitive type, with their first (and only) argument referring to a state of affairs, as in (39) (taken from Nuyts et al. 2018: 24). This trend is even more outspoken in the spoken than in the written data.

(39) En rijdt er tussen het feestgedruis door toch nog soms eens een trein, dan is dat mooi meegenomen. Maar het hoeft niet meer per se. [lit:] ‘And if there is an occasional train passing during the festivities, that is an asset. But it need not anymore [i.e. it is not necessary/indispensable].’

Hence, as Nuyts et al. (2018: 32–33) argue, hoeven – as a negative polarity item – is undergoing the same kind of re-autonomization process as kunnen, mogen and moeten, more or less simultaneously with the latter verbs. Since the verb has only started to auxiliarize in the course of the END period, however, this re-autonomization trend ‘breaks’ – as it were – into the still ongoing process of grammaticalization. According to Nuyts et al. we are dealing here with a case of ‘collective degrammaticalization’: hoeven joins kunnen, mogen and moeten, probably helped by the still well-known original main verbal use of the verb and of its predecessor behoeven – its status as a main verb, even if with different grammatical properties, is not uncommon for the language user. As reported in Nuyts et al. (2019), the picture in zullen and willen is very different, however. Just like the other modals, they, too, have grammaticalized from full verbs. In the data for zullen there is no trace anymore of the original main verb, in any period. And also intuitively it does not seem to exist anymore: if you want to express that you owe something to someone – which was the meaning of the original main verbal use of zullen – you can’t do that anymore by saying: *ik zal hem 10 euro [lit:] ‘I shall him 10 euros’ [i.e., I owe him 10 euros]. So the process of auxiliarization was already completed in the oldest available documents of zullen. Willen mainly occurs as an auxiliary from the oldest documents onwards as well, although its original main verbal use, meaning ‘to wish’, clearly remains present too, until today (as e.g. in ik wil een snoepje ‘I want a candy’). 48 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

However, in their further grammatical evolution, unlike the other modals, zullen and willen do not really participate in the re-autonomization process. Zullen does feature ‘new’ autonomous uses of the type in (38) and (39) above. But in all of them, one can (still) assume the presence of an elided main verb (most often: zijn ‘to be’; cf. section 3.1.3.1). So, we are never dealing with an undeniable main verbal use, of the kind we find in kunnen, mogen, moeten and hoeven. (40) offers an illustration from the spoken material of Nuyts et al. (2019).

(40) ik weet niet of ie ’t heel druk heeft maar ’t zal wel. [lit:] ‘I do not know whether he is very busy but it will [be] so.’

Moreover, as Nuyts et al. argue, instances of this type are very marginal: they only occur in (the data of) PDDS, in very small numbers, and only in combination with the particle wel (as in (40)). The evolution is statistically not significant. Maybe these few instances are a sign of the start of a re-autonomization process also in this verb, and of the same kind as in the other four modals. But if so, this process is starting later than in the other verbs, and is so far insignificant. Intuitively, willen, too, would seem to allow for ‘new’ autonomous uses of the type in (38) and (39), as e.g. in het wil maar niet [lit:] ‘It just won’t [work]’. (In this case, het ‘it’ has to be interpreted as an element referring to a state of affairs. But here one can fairly easily imagine the elided main verb lukken ‘work, succeed’. So, this would be a case in which the modal verb is still clearly an auxiliary, showing an implied main verb, cf. section 3.1.3.1.) However, no instances of this kind have been found in the data of Nuyts et al., in any period (not even in the spoken material). As indicated, willen does show autonomous uses in the data. Their frequency fluctuates, and like in the central modals they drop in END and rise again in PDD. But even in END they remain much more frequent than in the central modals. More importantly, however, the main verbal uses do not change in nature over time: they remain predominantly transitive, and do not feature a first argument referring to a state of affairs. Many of the autonomous instances of willen are of the type in (41) (taken from the spoken material of Nuyts et al. 2019). In these, the first argument refers to an entity (most often an animate or human being); the second argument is a complement clause referring to a state of affairs. LITERATURE 49

(41) ‘k wil gewoon heel graag dat je meegaat. ‘I just really want you to come along.’

This pattern does not occur in the other modals. But, most probably, it concerns a continuation of the original transitive pattern of willen, which could always take an entity as well as a state of affairs as its second argument. In any case, the pattern is already clearly present in the EMD data of Nuyts et al. too. On the basis of all the observations, then, Nuyts et al. (2019) conclude that the re- autonomization trend involves only part of the system of the modal auxiliaries in Dutch. It occurs in the three most central modals, kunnen, mogen and moeten, as well as the less central modal hoeven (in all four starting sometime between END and PDD), but not (really) in the two other less central modals, zullen and willen. As will be shown in section 1.2.2.2, this (mainly) seems to correlate with the extent to which the first four modals do show semantic similarity, of a kind not attested – or at least not to the same extent – in the latter two verbs.

1.2.1.2.3 The combination with a directional

Nuyts (et al.) did not venture into the question what is causing the re-autonomization process of the modal verbs in Dutch, but the issue is addressed by Honselaar & Olbertz (2016; Olbertz & Honselaar 2017). On the basis of entries of moeten in historical dictionaries, Olbertz & Honselaar (2017: 288) suggest that the process in Dutch is entirely due to the presence of a directional in the clause (cf. (36) above), and is therefore not a matter of degrammaticalization, as suggested by Nuyts (et al.). Honselaar & Olbertz (2016: 200), in a more substantial diachronic study of the same verb, leave open the origins of uses of the kind in (38) and (39), however. In what follows we will describe the tenor of their studies. Based on diachronic corpus research, Honselaar & Olbertz (2016) confirm that from the 18th to the 20th Century the relative frequency of the autonomous uses of moeten has drastically increased. In order to explain the upsurge, they refer to the fact that the oldest independent uses of moeten (nearly) exclusively occur in the presence of a directional element and with a first argument referring to an animate (usually human) being, as in the Early Middle Dutch example in (42) (from Olbertz & Honselaar 2017: 286).

50 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(42) Wi moeten int helsce vier. ‘We must [go] into the infernal fire.’

Therefore, Honselaar & Olbertz consider the combination with a directional element as the starting point of the process. They consider moeten in this pattern as a lexical verb, meaning ‘must move’ (or ‘be moved’ in (45) below). This contrasts with the analysis in Nuyts (et al.), in which moeten in (42) would be considered an auxiliary with an elided main verb gaan ‘to go’. Through time, the combination with a directional phrase strongly increases, to become really frequent in the 18th Century. An example from this time slot is (43) (from Honselaar & Olbertz 2016: 189).

(43) Johannes moest in de Gevankenis. ‘Johannes had to [go] to prison.’

From the 19th Century onwards, then, the types of directional phrases become more diverse: directional elements lacking a specific goal, such as weg ‘away’ in (44), become “increasingly popular” (cf. Honselaar & Olbertz 2016: 190). But also instances with a directional involving an inanimate entity as the first argument start to occur. An illustration of the latter is (45) (from Honselaar & Olbertz 2016: 195).

(44) daarom moest ik weg, weg met haar, zoover mogelijk weg! ‘this is why I had to [go] away, away with her, as far away as possible!’

(45) nu alles bij mij van zijne plaats moest. ‘now that everything in my house had to [be moved] from its place.’

Subsequently, the lexical use of moeten starts to expand to contexts other than directional movement, including ones featuring a telic adjective, as in examples (46a–b) (from Honselaar & Olbertz 2016: 192, 196, respectively). The adjectives dood ‘dead’ and leeg ‘empty’ share with the large majority of directionals (including those in (42) and (43)) that they mark an endpoint (hence are telic). So, we are dealing here with a “generalization of goal-oriented movement to the domain of goal-orientation as such”, according to Honselaar & Olbertz (2016: 193). LITERATURE 51

(46) a. je gaat nu blijkbaar uit van de veronderstelling, dat hij beslist dood moet… ‘apparently you assume that there is no doubt that he has to be killed’ (literally: ‘[that] he certainly must dead’)

b. Komaan, de kelder moet toch leeg; wie drinkt er eens met mij? ‘Come on, the cellar must [become] empty anyway; who drinks with me?’

The autonomous uses (with a directional element as well as with a telic adjective) even increase somewhat further in the course of the 19th and 20th Century. Also in the 19th Century, the first autonomous uses of moeten with a first argument referring to a state of affairs start to appear – (47) offers an illustration (from Honselaar & Olbertz 2016: 196). They further increase in the course of the 20th Century, (48) is an example (from Honselaar & Olbertz 2016: 197).

(47) zij vroeg zich af, of haar nog kracht genoeg overbleef het zware deksel in beweging te brengen. Maar het moest! ‘she asked herself, if she would have sufficient power left to get the heavy lid moved. But it had to [be]!’

(48) [about the rules of a game] Ik weet niet hoe het moet, zei ik. ‘I don’t know how to do it, I said.’ (literally: ‘I don’t know how it must’)

As mentioned, for Honselaar & Olbertz (2016: 200) it is not clear how the uses with a state of affairs have emerged, but Olbertz & Honselaar (2017: 288) ascribe them to analogy with the increasingly frequent combinations with a directional element. In this thesis we will investigate the role of directionals more thoroughly, not only in moeten, but in all four modals participating in the re-autonomization process, i.e. kunnen, mogen, moeten and hoeven. The results in Chapter 4 will indicate that, in spite of their prominent position in moeten, they cannot account for the re-autonomization process in the entire Dutch modal system.

52 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

1.2.2 Meaning evolutions in the modal verbs

1.2.2.1 The modals in Germanic11

The semantic evolution of the modal auxiliaries in (some of) the Germanic languages has also been investigated before. Here we will briefly discuss the main findings in the literature on the history of the English, German, Danish and Dutch modals. For English, Standop (1957) provides a detailed description of the various uses of the ancestors of may, must, shall, and will in a corpus of Old English (OE) prose and verse. Not only the different meanings, but also the uses of the modal verbs in different clause types are distinguished. He also makes a few diachronic remarks, however. Noting that OE *magan meant ‘can’ (possibility) rather than ‘may’ (permission), Standop (1957: 18) points out that can and may have shifted their meanings in a similar way, can taking up may’s earlier functions:

OE *magan (possibility) > may (permission) OE cunnan (‘know how to’) > can (possibility) > can (permission)

The variation and ‘competition’ between can and may in Middle and Early Modern English has been explored by Kytö (1987, 1988, 1991), using the Helsinki Corpus. Goossens (1992) focusses on the connections and relative diachrony of the various senses of can, including its epistemic meaning, which he relates to the use of can to express ‘external possibility’. The history of must (OE *motan) has attracted particular attention, both from scholars of OE and linguists investigating diachronic change. In OE, *motan could express both possibility and permission – see Standop (1957) and Goossens (1987) for suggestions about how it differed from *magan – and perhaps also necessity. While Solo (1977) argues that all apparent necessity cases in OE allow possibility readings, there seems to be agreement that *motan occasionally does mean ‘must’ in OE (see e.g. Goossens 1987 and Van Herreweghe 2000). Yanovich (2016) has proposed that *motan in OE was a ‘variable-force modal’ with a particular meaning which may both be understood as possibility and necessity, depending on the context. As for the diachronic development, the change from possibility to necessity has been suggested to happen in negative contexts (e.g. by Standop 1957 and Goossens 1987), or as a result of

11 This subsection was written in collaboration with Sune Gregersen (University of Amsterdam). LITERATURE 53 pragmatic inference (Traugott & Dasher 2002). Yanovich (2016) does not need either of these triggers, since in his scenario necessity was already one of the available readings of ‘variable- force’ *motan. What happened in Middle English (ME), according to this view, was a loss (namely of the ‘pure’ possibility reading) rather than an innovation. The development of epistemic out of ‘root’ meanings has also figured prominently in the literature. As already mentioned in section 1.1.2.3, Traugott (1989) characterizes the emergence of epistemic must, shall, and will as an example of subjectification in semantic change. She suggests that all three develop from deontic (which, in the case of will, also includes volition) to weakly epistemic and then to strongly (i.e. more subjective) epistemic meanings. Sweetser (1990) instead characterizes the development as an instance of metaphorical transfer from ‘real-world’ uses of the modals (cf. also section 1.1.2.4): The expression of obligation in the real world as in (49a) is extended to apply to the ‘knowledge world’ as well. Hence, in a case like (49b), “the available (direct) evidence compels me to the conclusion that you were home” (Sweetser 1990: 61).

(49) a. You must come home by ten. (Mom said so.) b. You must have been home last night.

A similar argument is made concerning the development of epistemic may, which Sweetser assumes developed from permission may. Bybee et al. (1994: 195–202), on the other hand, investigate the use of may in ME and conclude that the epistemic and permission readings are parallel developments out of the now rare ‘root possibility’ use.

ability > root possibility > epistemic possibility > permission

The ‘obligation-to-epistemic’ hypothesis has also been questioned. Furmaniak (2011) investigates the uses of must in a corpus of 16th- to 18th-Century texts and finds that the epistemic examples are often ambiguous with a reading which he terms ‘inevitability’, i.e. a dynamic modal meaning, as shown in (50) below (cf. Furmaniak 2011: 63).

54 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(50) My daughter is at present engag’d in a way, that to her must be more agreeable than entertaining either you or me. a. inevitability reading: ‘that to her is necessarily going to be more agreeable…’ b. epistemic reading: ‘that to her is surely going to be more agreeable…’

Most scholars, however, seem to agree that the epistemic meanings develop relatively late and that there are no unambiguous examples in the OE sources. The literature on the German modals has focussed predominantly on semantic changes. Klarén (1913), for instance, investigates the semantic development of können, mögen and müssen from Old High German (OHG, c. 800–1050) to the 18th Century. Similarly to Standop’s (1957) work on OE, Klarén distinguishes ‘maximally’ between different uses not just according to their meaning but also their occurrence in different clause types. A very different approach is presented by Bech (1951), who describes the possible meanings of the High German modal verbs (können, mögen, müssen, dürfen, sollen and wollen) in terms of three semantic opposi- tions: possible vs. necessary, intrasubjective vs. extrasubjective, and causal vs. autonomous, the latter of which has to do with causality vs. intentionality (see Bech 1951: 7). The semantic changes observed in the history of High German are conceptualized as shifts within this system with no reference to actual usage contexts. More recently, the contributors to Fritz & Gloning (1997) have investigated the history of the modals from various angles, paying particular attention to the way they are used in different text types and how these differ from spoken interaction, where most changes are assumed to have happened. The origin of epistemic mögen and müssen is investigated by Gamon (1993). The last two chapters of Diewald (1999) are devoted to changes in the six ‘core’ modals also investigated by Bech (1951). In OHG, the modal mögen is attested with a wide spectrum of meanings. Klarén (1913: 3–7) mentions various types of possibility, as well as the use of mögen as konjunktivvertreter, a meaning we would describe as ‘optative’, as in (51).

(51) oba wir sín nu thárben, ja mag iz gót irbarmen! ‘If we now lose him, well, may God have mercy!’

LITERATURE 55

In addition, Fritz (1997b: 9) suggests that permission and epistemic uses of magan are also found in OHG. However, he gives no examples of permission uses, and the examples cited in the Althochdeutsches Wörterbuch (AWb) with the meaning “Berechtigung, Erlaubnis haben” (s.v. magan, sense 6) all seem to allow a dynamic reading as well. As for the epistemic sense, Fritz (1997b: 95) quotes an ostensible example from Otfrid, but this is also probably better analysed as dynamic. Gamon (1993) discusses the problem of interpretation at length and quotes a number of examples, also from Otfrid, which according to him allow both a dynamic situational (‘propositional root possibility’) and an epistemic interpretation (Gamon 1993: 147– 149). An example is given in (52) (extracted from Otfrid, Evangelienbuch 2,14,89–91).

(52) Scal iz Kríst sin, fro min? […] Bi then gidóugnen séginin so thúnkit mih theiz megi sín ‘Will this be Christ, My Lord? […] From his secret powers it seems to me that it may be him.’

We agree that these are indeed ambiguous between a dynamic situational and an epistemic reading, suggesting that the potential for epistemic uses was already present in the OHG ancestor of mögen. In any event, epistemic instances are securely attested in the Middle High German (MHG, c. 1050–1350) period, and we also find examples which allow a directive reading. Klarén (1913: 19) also mentions two possible examples with concessive meaning from von der Wogelweide, but these also seem to allow a dynamic reading. The semantic field of mögen remains more or less the same during the Early New High German (ENHG, c. 1350–1650) and New High German (NHG, c. 1650–) period. All different meanings and uses attested in OHG and MHG are represented by this verb. Nevertheless, there are some differences in frequency between them. Whereas the non-dynamic (epistemic, optative, directive, and concessive) meanings are still generally rendered by NHG mögen, the different dynamic modal uses are increasingly expressed by können, which according to Fritz (1997b: 9) has evolved into the absolutely prototypical verb of possibility by the 18th Century. In Present Day German (PDG), most of the older uses of mögen have become rather marginal. Fritz (1997b: 9) only mentions the epistemic use (as in Er mag es gewußt haben ‘He may have known it’), which he describes as marginal. The directive use of permission has been taken over by dürfen (another modal in the system). The dominant use of mögen today appears to be ‘like’, 56 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH which has become increasingly frequent since the 16th Century, but – as already mentioned in section 1.2.1.1 – the early origins of this pattern are disputed. Kunnan was still used with its original main verbal meaning ‘know’ (wissen, verstehen, kennen) in OHG, but is also found with infinitives with a dynamic modal meaning (see Klarén 1913: 3). In MHG, the meaning ‘know’ decreases in frequency, and the dynamic use has become predominant (Klarén 1913: 13). As such, können gradually takes over the dynamic terrain of mögen in a way similar to what has been described for English can and may discussed above (cf. the example from Standop 1957). In ENHG, epistemic uses are attested as well, although they remain rather marginal. In addition, Peilicke (1997: 223–224) mentions the presence of a permission use of können in ENHG, but this meaning also appears to be marginal. Fritz (1997a) finds no examples in his newspaper corpus from 1609. Müssen in OHG expresses a possibility meaning which, according to Klarén (1913: 8), is never inherent in the subject, but contingent on the circumstances or someone else. In addition, OHG müssen is amply attested as konjunktivvertreter, i.e. with optative meaning. At some point in late OHG, müssen also starts appearing with a necessity meaning. Like for English must, the shift from possibility to necessity has been explained in different ways. On the one hand, it has been suggested that negation played a crucial role in the change (see e.g. Klarén 1913; Bech 1951). On the other hand, the shift has been interpreted as primarily pragmatically motivated (cf. Fritz 1997b; Diewald 1999; and Paul 2002 [1897]). No matter how the change to necessity meaning occurred, it is clear that it had become established as one of the primary meanings of müssen by the MHG period. The use of müssen as a necessity modal survives up through the medieval and early modern periods and until today. Moreover, müssen also develops an epistemic (evidential) use, perhaps towards the end of the Middle Ages (cf. Fritz 1997b: 100). The modal system of Early Middle Danish (EMDa, 13th Century) is described briefly in Bjerrum’s (1966) grammar of the Scanian law, which all subsequent works take as their starting point. Bjerrum’s description is in turn inspired by Bech’s (1951) work on German. According to Bjerrum’s analysis, EMDa did not formally distinguish ‘possibility’ from ‘permission’, or ‘necessity’ from ‘obligation’. The former meanings were both expressed by ma ‘can, may’, the latter by scal ‘must, have to’. (53) and (54) give an example of possibility and permission ma, respectively (both taken from dansksproghistorie.dk/74). LITERATURE 57

(53) Thænnæ steen ma æi eld skathæ ‘This stone cannot be harmed by fire’

(54) Ræfua unga ma ængin man grafua vp. af annars marko ‘No one is allowed to dig up fox cubs on another’s property’

The developments from Middle Danish (MDa, c. 1200–1500) to Modern Danish (ModDa, c. 1500–) have been investigated in Obe’s (2013) dissertation and more recently in the handbook chapter by Heltoft & Nielsen (2019). Hansen & Heltoft (2019 [2011]) also occasionally compare the older and modern systems in their grammar of contemporary Danish. Simply put, the meanings of permission and obligation remain unchanged from MDa to ModDa, while the dynamic modals change. Ma comes to replace scal as the modal of dynamic necessity, as in (55) (cf. dansksproghistorie.dk/74), whereas kan (< ‘know’) takes up ma’s earlier possibility function.

(55) hon giordh cors for sek. tha matte diæfolin ryma ‘She made the sign of the cross; then the devil had to flee’

The ‘competition’ between kan and ma, and between ma and scal, in three 15th-Century texts is investigated in detail by Obe (2013). Obe’s explanation for this change is similar to the one proposed by Fritz (1997b), Diewald (1999), and Paul (2002) for German müssen, namely that the meaning ‘possibility’ was reinterpreted as ‘necessity’ in affirmative contexts with only one option. Heltoft & Nielsen (2019) share this view. According to Heltoft & Nielsen (2019), epistemic (evidential) meanings of ma occur at least from c. 1700, as e.g. in (56) (cf. dansksproghistorie.dk/74).

(56) han maa være gaæt fra Forstanden ‘He must have gone out of his mind’

Kan is also described as having an epistemic use, both by Obe (2013) and Hansen & Heltoft (2019 [2011]). In many cases, however, a dynamic reading is possible as well, if not preferable. In this way, Danish kan seems more comparable to Dutch kunnen, which also usually allows a 58 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH dynamic reading alongside an epistemic one (cf. Van Ostaeyen & Nuyts 2004: 21), than to English may, which is often unambiguously epistemic (see Biber et al. 1999: 491–493). The semantic evolution of the modal auxiliaries in Dutch, finally, has been investigated in various studies of Nuyts (and colleagues). Their developments will be discussed more thoroughly in the following subsection(s).

1.2.2.2 The modals in Dutch

1.2.2.2.1 The central modals

As far as the ‘central’ modals, kunnen, mogen and moeten, are concerned, the semantic evolution from OD to PDD has been investigated in detail by Byloo & Nuyts (2014) and Nuyts & Byloo (2015). Like in their grammatical developments, the three verbs show a number of similarities in their meaning development, although there are considerable differences in the details. As mentioned in section 1.2.1.2.1, the original main verbal uses from which the three modal auxiliaries have emerged meant ‘to know’ in kunnen, ‘to have power’ in mogen and (most probably) ‘to measure’ in moeten. No matter how different those meanings are, in all three modals they have given rise to a full range of modal and related meanings which are typically associated with modal verbs in the languages of the world. This first involves the development of different dynamic modal meanings, from which a number of other meanings and uses have emerged, including a deontic, an epistemic or evidential, and a directive one. (See section 3.2.1 for definitions.) The three modals differ, however, in terms of when, how fast and how consistent these evolutions to subjective and intersubjective meanings (cf. section 1.1.2.3) have happened. In mogen and moeten, this evolution started long before the oldest OD documents. Even in the oldest data used in Byloo & Nuyts there is no trace anymore of their original main verbal uses (cf. also section 1.2.1.2.1) and they both already feature a full range of more subjective and intersubjective meanings from OD and EMD onwards. But mogen and moeten also differ, in the sense that in mogen the processes of subjectification and intersubjectification continue until today, while in moeten they do not. In mogen, there is a very gradual, continuing, decline through time of the least subjectified meanings, notably LITERATURE 59 the different dynamic modal ones. And there is a continuous increase of the more subjective and intersubjective meanings, most clearly of the deontic and the directive ones. The other intersubjective meanings and uses (i.e. volition, condition and concession, as indicated in the right hand column of Table 5 in section 1.1.2.3) only develop marginally. The epistemic meaning of mogen even disappears. Moeten, by contrast, has not advanced any further in terms of (inter)subjectification since OD and EMD. There is considerable fluctuation in the different subjective and intersubjective meanings through time, but the changes do not show a clear development pattern – they appear more or less random. The least subjectified (dynamic) meanings, for instance, do not diminish over time. Among the more subjectified meanings and uses only the deontic one gains some ground. The evidential meaning (which only shows up in the END data)12 remains marginal until today. The intersubjective categories, finally, even lose ground. Especially the meaning of volition shows a substantial decrease through time. Hence “moeten is in continuous semantic flux, yet without involving a clear development pattern of the kind predicted by the (inter)subjectification hypothesis” (Byloo & Nuyts 2014: 107). In kunnen, on the other hand, the process of subjectification appears to have started later. In OD and EMD the original main verbal meaning is still clearly present (there are even traces of it until PDD), next to the different dynamic modal meanings, which already strongly predominate in those periods. But it is only in END that the verb develops its more subjective deontic and epistemic meanings, while the intersubjective directivity meaning even only appears in PDD. And although they gain further ground towards the present, “all of these more subjective and intersubjective meanings remain infrequent today” (Byloo & Nuyts 2014: 102). The preceding concerns the general meaning profile of the modals. Nuyts (2013: 130–131) also briefly addresses the semantic profile of their ‘autonomous’ uses. It turns out that these do not return to their original main verbal meanings (as one would expect if they would involve ‘retraction’), but even seem to focus, more strongly than the purely auxiliary uses, on the (inter)subjective meanings and uses, especially the deontic and directive ones. The meaning profile of the ‘autonomous’ uses will be investigated in much more detail in Chapter 5.

12 As pointed out in Byloo & Nuyts (2014: 102), “given its low frequency, it cannot be excluded that it already existed before END and just happens to be missing in the older data.” 60 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

1.2.2.2.2 The less central modals

As for the ‘less central’ modals, the semantic evolution of hoeven has been studied in Nuyts et al. (2018). The evolutions in zullen and willen have not been studied in detail yet, but Nuyts et al. (2019) briefly discuss them, tentatively, on the basis of the information in the Oudnederlands Woordenboek (2012) (ONW; the ‘Old Dutch Dictionary’), the Vroegmiddelnederlands Woorden- boek (VMNW; Pijnenburg et al. red, 2000), the Middelnederlandsch Woordenboek (MNW; Verwijs & Verdam 1885–1929), and the Woordenboek der Nederlandsche Taal (WNT; De Vries & Te Winkel 1864–1998). As already mentioned in section 1.2.1.2.2, hoeven only emerged in END, as a reduction from the older verb behoeven, meaning ‘to need’. From the start it shows a considerable share of subjective and intersubjective meanings (which it has taken over from behoeven, even if they are far less prominent in the latter; this underlines the still very close association of the two forms in END). The dynamic meaning dominates by far, and the share of the original main verbal meaning is immediately very low (much lower than in behoeven). In the transition to PDD, the (inter)subjectification process of hoeven continues. The original main verbal meaning becomes marginal. And even the least subjectified dynamic modal meanings tend to decrease, in PDDS (i.e. the more progressive variant) even more so than in PDDW. The share of the other more subjective and intersubjective meanings, by contrast, strongly increases (especially the deontic, directive and volition uses). As Nuyts et al. (2018) also show, this evolution in PDD is even stronger in the ‘new’ main verbal uses of hoeven separately (as a subcategory of the ‘autonomous’ uses, cf. section 1.2.1.2.1). These lack the original main verbal meaning entirely (which is not very surprising in itself given its low prevalence in the general semantic profile), but also the dynamic meanings are less prominent, while the more subjective and intersubjective meanings together have a larger share than in the general meaning profile of the verb. So, like kunnen, mogen and moeten, hoeven also seems to show a tendency towards more subjective and intersubjective meanings in the ‘autonomous’ uses, and the deontic, but especially the directive meanings seem to play a central role here again, according to the data of Nuyts et al. (2018) (cf. Chapter 5). Nuyts et al. (2018) suggest that the very rapid and more or less simultaneously rise of a full range of subjective and intersubjective meanings in hoeven since its emergence in END may LITERATURE 61 be due to a strive of this verb to adapt to the semantic structure of the three most central modals, kunnen, mogen and moeten (i.e. ‘analogy’). On the basis of dictionary information, Nuyts et al. (2019) argue that zullen and willen show a very different picture, however. These verbs do not share the different subjective and intersubjective meanings and uses of kunnen, mogen, moeten and hoeven, or at least not to the same extent. The oldest meaning evolutions of zullen are not attested in the documents, but most probably this verb has never had any dynamic modal meaning (in any case there are no clear traces of such a use, not even in the oldest text material). Already in OD and EMD, zullen features a deontic and a directive use, but especially its use as the marker of the future tense is immediately predominant. According to Nuyts et al. many of the instances considered deontic or directive in the dictionaries may even just as well be interpreted temporally. (57) might be an example (see Oudnederlands Woordenboek, s.v. sullan, sense 2.1).

(57) Thu scalt hiro saluti congaudere [i.e., Jij zal je over hun heil verheugen] ‘You will rejoice in their salvation.’

Today, the deontic and directive meanings have largely disappeared (je zal dat doen ‘you shall do that’, for instance, has become fairly marginal as indication of an order). Instead, an epistemic meaning has emerged (as e.g. in Er wordt aan de deur geklopt, dat zal Jan zijn ‘Someone is knocking at the door, that will be John’, cf. Woordenboek der Nederlandsche Taal, s.v. zullen, sense 8), but this is no coincidence, according to Nuyts et al., since the future is inherently uncertain. The future tense remains absolutely dominant, however, and this seems to have kept zullen largely outside the system of the other modal verbs in Dutch. Willen, on the other hand, does not seem to have had any other meanings than the present one, namely an expression of a wish or desire (‘volition’). The other meanings and uses found in the dictionaries all seem to be variants of it, or related to it through conversational implicature. At no point in its history did it develop any of the prototypical modal and related meanings. Moreover, ‘volition’ is marginal among the modal categories, which according to Nuyts et al. may be why the verb is at the sideline of the system of the modal auxiliaries.

62 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

1.2.3 Summary of the findings relevant for the present corpus study

Before turning to the present corpus study, however, let us briefly summarize the main findings from sections 1.2.1.2 and 1.2.2.2 that are relevant for the current investigation. On the one hand, the studies by Nuyts (2011, 2013, 2014) and Nuyts et al. (2018) have shown that kunnen, mogen, moeten and hoeven started to re-autonomize sometime between Early New Dutch and Present Day Dutch. Moreover, the studies by Byloo & Nuyts (2014), Nuyts & Byloo (2015) and Nuyts et al. (2018) revealed that, in their evolution, the four modal verbs also developed a full range of modal and intersubjective meanings. (For the concrete figures of the four modals, see section 4.1 for the grammatical evolution, respectively section 5.1 for the semantic evolution.) On the other hand, the study by Nuyts et al. (2019) has demonstrated that zullen and willen, the two other modals in the system, do not (really) participate in the group, neither grammatically nor semantically. Therefore, in the present corpus investigation, only the first four modals have been further inquired – the last two verbs are not included. In the following chapter, we will describe and explain our methodology.

CHAPTER 2

Methodology13

As mentioned, the present corpus study only focusses on kunnen, mogen, moeten and hoeven, the four modal verbs which have been shown to reveal systematic re-autonomization behavior. Moreover, it focusses on the period from Early New Dutch onwards, since the earlier investigations have revealed that the re-autonomization process, in all four modals, has started sometime between Early New Dutch and Present Day Dutch (cf. sections 1.2.1.2.1 and 1.2.1.2.2), though without providing a more specific timing. In order to be able to determine the timing of developments within the New Dutch period, we have worked with materials from three maximally divergent stages of New Dutch, i.e. Early New Dutch (END, 1550–1650), New Dutch (ND, 1750–1850) and Present Day Dutch (PDD, 1950–). Ideally, we would also have examined the two intervening stages (1650–1750 and 1850–1950), but for reasons of feasibility (cf. the time limits on this project), these two stages were not included. As in the previous studies of Nuyts (et al.), for PDD we have moreover used (separately) data from both written (PDDW) and spoken (PDDS) language. In the frequency tables later on in this thesis (cf. Chapters 4 and 5), PDDW and PDDS will be presented in two consecutive columns, which may give the impression that they are to be considered consecutive stages in the language. This should and will be handled with care, though. As already mentioned in section 1.2.1.2.1, spoken language tends to be more progressive than written language, hence the spoken data may to some extent be taken to forecast a further stage in the evolution of Dutch as compared to the written data. Yet the distance between PDDW and PDDS is not the same as that between ND and PDDW, for instance. Moreover, there are many differences between spoken and written language which have nothing to do with the progressiveness of the variants.

13 Our methodology has also largely been explained in Caers & Nuyts (submitted). 64 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

In principle, we have analyzed 1000 instances per modal per period (and, for PDD, per genre), but for END and ND there were not enough instances in the available text materials of hoeven. So here we have included all instances we could find, i.e. 323 for both END and ND (cf. Table 16 in Chapter 4; the equal number for both is purely accidental). The choice for the sample size has been based on two elements. On the one hand, we have considered what is needed in view of our goal to gather a more complete picture of the re-autonomization process of the modals in Dutch. The data in the earlier studies of Nuyts (et al.) (samples of 200 instances) contained only very small absolute numbers of autonomous instances, and in order to get a better view of their properties, we need larger numbers of them, hence much larger samples. But, on the other hand, we also had to take into account what is practically feasible, given the available time. Therefore, we have kept the sample size at 1000, the more since even now our total data set covers 14,646 instances. For the written data, since there is no ready-made balanced diachronic corpus of Dutch available, we had to draw the samples from a self-compiled corpus of text materials covering the three different stages. For reasons of efficiency, we only made use of electronically available sources and corpora, since they allow for automatic queries. The main source is the Digitale Bibliotheek voor de Nederlandse Letteren (DBNL, see www.dbnl.org), which contains numerous texts of many different kinds or genres from all through the history of Dutch, including the three stages relevant to us. From this database we have used (nearly) all text material available – at the moment of consultation (in September 2016) – for each of the three (written) time periods, which in total came down to 779 different texts. To complement for underrepresented genres in the DBNL (cf. the need for balanced/ representative samples, explained below), we have used a few sources in addition. For END and ND, we have also made use of scans of 357 books provided by the Google books project for digitalizing parts of the University of Ghent library (see https://lib.ugent.be/), as well as of the Compilation Corpus Historical Dutch (2010) of Evie Coussé, in particular the part with administrative texts from the periods relevant for us. For PDDW, we have also made use of the newspaper parts of the ConDiv corpus (see Grondelaers et al. 2000), as well as of 27 texts which were carefully selected from Google (Scholar) websites. An overview of all text material extracted from the Digitale Bibliotheek voor de Nederlandse Letteren, the Google books project METHODOLOGY 65 of the University of Ghent, and the Google (Scholar) websites can be found in Appendix 1. The total number of words contained in the written data for the different periods is very hard to estimate. For the spoken data we have used the Corpus Gesproken Nederlands (CGN, see Nederlandse Taalunie 2004), a collection of 9 million words of Flemish and Netherlandic contemporary speech. The choice of materials and the selection of instances from them was guided by an attempt to be as representative as possible for each period/genre, in terms of regional spreading, discourse types and authors, as well as by the goal to achieve reasonable comparability in the (written) data across the different time periods. As for the regional spreading, we made a distinction between Northern (Netherlandic) and Southern (Flemish) Dutch. In the oldest language stages, Dutch was still a mix of all different kinds of dialects. But from END onwards, it started to show a tendency towards beginning standardization. Nevertheless, already in the 16th and 17th Century, this “language uniformization” got obstructed by “the [sub]division of the Netherlands into an independent [Northern] and a Spanish [Southern] part” (see Byloo & Nuyts 2011: 27–29), which ever since caused more differences between the two regions than within each of them. That is why, in this thesis, we have opted for an equal share of data from Northern and Southern Dutch in each sample (i.e. 500 instances of each). The choice between North and South was not always easy, however. It is not always obvious which criterion to use in allocating an instance (written or spoken) to one of these regions. Should one use the author’s (or speaker’s) place of birth? Or the place where (s)he was staying at the time of writing (or speaking)? Or maybe the place where (s)he has spent most of his/her life? Karel van Mander, for instance, lived from 1548 to 1606, was born in the South (in West-Flanders), but moved to the North (to Haarlem) in 1583, where his famous Schilder-boeck was published in 1604. So, to which region do we have to assign this work? Since in instances of this type we cannot know for sure whether the author has adapted his language to the place where he has written his text, we have systematically classified all text material according to the author’s place of birth (and, in case of early emigration, to the region (s)he has learned his/her language as a child). If the author is unknown, we have not included the text. 66 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

In addition to the regional spreading, we have also tried to achieve representativity in terms of discourse type. As for the written data, Table 6 provides an overview of the text types and the genres we consider to belong in each of them.

Table 6. Classification in text types

TEXT TYPES TEXT GENRES

rhymed/lyrical texts poetry, and all kinds of rhymed material

narrative texts prose (unrhymed)

dialogues drama (unrhymed)

personal texts autobiographies, diaries, letters,…

statutory texts administration, law, politics,…

religious description theology, sacred texts,…

secular description argumentation, information, history, lifestyle, linguistics, pedagogy, philosophy, science,…

Again, it was not always easy to assign a specific text to a certain genre and/or text type. In END, for instance, there are some treatises (i.e. argumentative texts) which are religiously inspired. Or in PDDW there are numerous autobiographies of politicians. In such cases we have always tried to assign the text to the most appropriate type: for instance, a text about the life and work of a politician is more a personal than a political (statutory) text; and a treatise is an argumentative text, no matter what subject it deals with. The precise share of each of these text types differs somewhat for the different time slots, however. This is due to the fact that they are not equally well represented in the materials for the different periods. The distribution of instances in the END, ND and PDDW data is indicated in Table 7. The fifty-fifty distribution between Northern and Southern Dutch has been applied within each of the slots in the table. These figures are valid for full samples of 1000 instances; they do not apply when we did not manage to collect a full sample of 1000 (cf. hoeven in END and ND).

METHODOLOGY 67

Table 7. Number of instances per text type

END ND PDDW

rhymed/lyrical texts 300 150 0

narrative texts 100 150 0

dialogues 0 50 0

personal texts 20 50 80

statutory texts 20 60 180

religious description 60 40 20

secular description 500 500 720

total 1000 1000 1000

As the table shows, in END and ND a considerable number of instances has been selected from rhymed/lyrical texts and literary narratives and dialogues, while in PDDW these text types are omitted entirely. Ideally, we would not have used any such instances at all, in any of the periods, as they are potentially subject to rhyme or rhythm pressure or other stylistic processes which distort the normal use of forms. But, given the availability of text materials in END and ND, it was impossible to exclude them. We have tried to limit their share to the extent possible, though, and to maximize the share of instances from non-literary texts. Within the latter, the differences between the time slots are also due to the relative presence of text types. As a consequence, the share of religious instances decreases through time, in favour of personal, statutory and secular ones. We also tried to distribute the data over as many different authors as possible, by not selecting more than 20 instances per text. Within the confines of all those criteria, then, the selection of the samples was random. For the spoken data we used all components present in the CGN, as listed in Table 8, except for “read speech” (the last one in the list). As argued in Byloo & Nuyts (2011: 31), “it contains read-aloud extracts from novels, i.e. written language that was not even intended to be read aloud.” Since this corpus is already well balanced, we did not fix a distribution of instances over the different text types, as has been done in the written samples. We simply made a random selection of 500 Northern and 500 Southern instances from the entire corpus. 68 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 8. CGN components a. Spontaneous conversations (‘face-to-face’) b. Interviews with teachers of Dutch c. Spontaneous telephone dialogues (recorded via a switchboard) d. Spontaneous telephone dialogues (recorded on MD via a local interface) e. Simulated business negotiations f. Interviews/discussions/debates (broadcast) g. (political) Discussions/debates/meetings (non-broadcast) h. Lessons recorded in the classroom i. Live (e.g. sports) commentaries (broadcast) j. Newsreports/reportages (broadcast) k. News (broadcast) l. Commentaries/columns/reviews (broadcast) m. Ceremonious speeches/sermons n. Lectures/seminars o. Read speech

The written data were queried by means of AntConc 3.4.4w (Windows, 2014). Since in END and ND the spelling was not entirely fixed yet, searches for these periods covered all possible spelling variants of the modals, as mentioned in De Vries and Te Winkel (1864–1998). For the queries in the CGN, we have used Corex (Version 6.1). Each instance (written and spoken) has been considered within its broader context. We selected at least 15 to 20 sentences before and 10 to 15 sentences after the actual clause (or more if necessary). (Only) interrupted cases, ungrammatical clauses, etc. that made the analysis of the modal impossible were excluded. The data have been tested statistically be means of Fisher’s Exact test. In line with common practice in linguistics, we assume statistical significance when the p-value is less than or equal to .05. In case of statistical insignificance, we do not mention the p-value. The testing has been done by means of the statistical software package R, version 3.3.2 (The R Foundation for Statistical Computing 2016). Bearing in mind our earlier comment regarding the special status of the spoken data, we have consistently measured all developments over time in two ways: one excluding and one including PDDS. Significancy figures for these developments always cover the data for each of the stages involved. ‘The significance of the evolution from END to PDDW’, for instance, covers the data for END, ND and PDDW. ‘The significance of the evolution from END to PDDS’ adds the PDDS data (separately from the PDDW data). When the METHODOLOGY 69 significance of a difference between only two stages is measured, this is stated explicitely. PDDS is thereby only compared to PDDW, never to older stages. In all examples in this thesis, the relevant modals are indicated in bold. In examples from our own corpus, the periods from which they come (END, ND, PDDW, PDDS), as well as a reference to their source text, are mentioned between brackets after the English translation. For example, “ND – Stijl (4) – North (DBNL)” means the instance is from the ND sample, and is taken from the fourth text of Simon Stijl, as listed in Appendix 1, i.e. Krispyn filozoof, which is part of the Northern Dutch material from the Digitale Bibliotheek voor de Nederlandse Letteren. For references to newspaper materials from the ConDiv Corpus, we use the following abbreviations: NRC (i.e. NRC Handelsblad), DS (i.e. De Standaard), HLN (i.e. Het Laatste Nieuws) and HBVL (i.e. Het Belang van Limburg). For spoken instances, finally, we mention their code number in the CGN (e.g. fn000250 for a Northern Dutch transcript, or fv701094 for a Southern Dutch one). We refrain from providing word-for-word glosses, but when relevant we render the critical elements faithfully in the English translation, even if this produces an ungrammatical pattern in English. Whenever the latter applies, we mark this by adding ‘[lit:]’ in front of the translation (cf. also (35a–f) in section 1.2.1.1). Corpus examples, especially from the CGN, are sometimes simplified (without formal marking) by omitting pause fillers, repetitions, back channel cues, or irrelevant parts of the utterance/text.

CHAPTER 3

Analytical categories

All instances first received a basic grammatical analysis in terms of whether they involve an auxiliary or a type of autonomous use. The autonomous instances received full further analysis in terms of a range of grammatical and semantic dimensions. For reasons of feasibility, the regular auxiliary uses were only partially analyzed further, however. They were not analyzed semantically, since the general meaning profile of each of the modals at stake is already known through the earlier studies reported in Byloo & Nuyts (2014), Nuyts & Byloo (2015), and Nuyts et al. (2018) (cf. section 1.2.2.2). Moreover, they were only analyzed for other grammatical features when necessary in order to understand the status of the autonomous uses. All semantic analyses, as well as most of the grammatical analyses, were double-checked by Jan Nuyts – cases of disagreement were resolved through discussion. In section 3.1 we present the grammatical categories used in the analyses. The central dimension is the distinction between ‘types of uses’ of the modals (discussed in section 3.1.3). This requires reference to a few more basic grammatical categories, however, which are therefore presented first. This concerns the notion of ‘transitivity’, including our conception of the notion of ‘an argument’ (3.1.1), as well as our classification of ‘types of arguments’ (3.1.2). In addition, we define our notion of ‘a directional element’ (3.1.4). In section 3.2 we introduce the semantic dimensions coded in the analyses. This concerns the ‘meaning categories’ expressed by the modals (3.2.1), as well as the category of ‘negative polarity’ (3.2.2).

3.1 The grammatical dimension

3.1.1 The notion of ‘transitivity’ and the concept of ‘an argument’

Verbs can be characterized in terms of their valency. Important for us is the distinction between intransitive and transitive verbs. This distinction is often defined in terms of whether the verb 72 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH takes a direct object (see e.g. Haeseryn et al. 1997: 50). Our present concept of ‘(in)transitivity’ is broader, however. It is defined in terms of the number of arguments a verb takes, whereby the notion of an argument is set widely. We consider as such any constituent that cannot be omitted without jeopardizing the grammaticality of the clause, irrespective of its formal properties (cf. also Vandeweghe 2013: 48–49, 96, 128–130). An intransitive verb takes only one argument, as in (58). A transitive verb takes two arguments. The second argument need not be a nominal phrase (realized as the direct object of the verb), as in (59), however. It can also be a prepositional phrase, or a pronominal adverb, if they function as an obligatory complement of the verb. For instance, one cannot omit the prepositional phrase buiten de vorm [lit:] ‘outside the form’ in (60), or the split pronominal adverb daer wt [lit:] ‘out of there’ in (61).

(58) Je zusje pesten, mag dat? [lit:] ‘Bullying your little sister, may that [i.e. is that acceptable/permitted]?’ (PDDW – Ghesquiere – South (DBNL))

(59) Wij hoeven op die wijs geen schepen meer. ‘We no longer need ships that way.’ (ND – Stijl (4) – North (DBNL))

(60) Het ging over een definitie van poëzie en hoe die niet buiten de vorm kan. [lit:] ‘It was about a definition of poetry and how that cannot outside the form.’ (PDDW – Brems – South (DBNL))

(61) soo es het ghewelt des waters doorghebroken ende heeft den keldere vervult met sulcker cracht dat noch hy noch die wateren daer niet wt en mochten. [lit:] ‘In that way, the violence of the water has broken through and has filled the basement with such a force that neither he nor those waters might [i.e. could] [get] out of there.’ (END – Lipsius et al. – South (UGent Library))

(61) illustrates that the second argument may be a directional (cf. section 3.1.4). This is even very common in our data: in all verbs of movement featuring a directional in our data the directional phrase cannot be omitted hence has to be considered the second argument. ANALYTICAL CATEGORIES 73

(62)–(63) offer two more examples (these are instances involving an elided main verb, see section 3.1.3, the implied verb is mentioned between square brackets after the clause).

(62) Zonder toelating van de eigenaar mag geen film de deur uit. [‘gaan’] [lit:] ‘Without the owner’s permission, no film may [go] out the door.’ (PDDW – DS – South (ConDiv))

(63) prei moet ‘r ook nog op d’rbij [i.e. bij in de pan]. [‘komen’] [lit:] ‘leeks must also [come] thereby [i.e. in the pan].’ (PDDS – fn000969 – North (CGN))

One type of non-omissible directional phrase we have not considered as a second argument, however, is the adverb weg ‘away’ in the pattern of the kind in (64a–b). Here one cannot imagine an elided main verb of movement (such as gaan ‘go’, for instance) anymore (this is a main verbal instance of the modal, see section 3.1.3), and we assume that the modal and the adverb have become a fixed combination (cf. also section 3.1.3.1). Therefore, in this pattern, we analyze the directional adverb as the non-verbal part and the modal as the verbal part of a compound verb (see e.g. Woordenboek der Nederlandsche Taal, s.v. wegmogen/wegmoeten). This is comparable to, for instance, the compound verbs (tegen iets) opkunnen [lit:] ‘can up (against something)’ [i.e., be able to stand up (to something)] (cf. WNT, s.v. opkunnen), (iets) kwijt kunnen [lit:] ‘can [get] rid (of something)’ and (ergens) terecht kunnen [lit:] ‘can [go] (somewhere)’ (cf. section 3.1.3.1).

(64) a. treinticketten dat mag ook allemaal weg zeker? [lit:] ‘Train tickets, that may all away too, right?’ (PDDS – fv701094 – South (CGN))

b. zonder ene twijfel weg moet het [i.e. elk oud, versleten kledingstuk]. [lit:] ‘Without any doubt, it [i.e. every old, worn piece of clothing] must away.’ (PDDS – fv700122 – South (CGN))

A special problem concerns constructions of the type in (65), in which the comparative adverb (niet) anders [lit:] ‘(not) otherwise’ is followed by a complement clause introduced by dan 74 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

‘than’. In this construction, one might consider the ‘that’-clause as the first argument of kunnen, but then the question arises how to analyse the phrase niet anders dan. Since the grammatical structure is not obvious here, we have treated the phrase as part of the ‘that’-clause, hence as part of the first argument as well. Whether this is the best possible analysis remains an open issue.

(65) Het kan bijna niet anders dan dat zijn onderzoek verregaande politieke repercussies met zich brengt. [lit:] ‘It almost cannot [be] otherwise than that his research entails far-reaching political repercussions.’ (PDDW – NRC – North (ConDiv))

3.1.2 Types of arguments

Our analyses also require that we make a distinction between types of arguments in terms of what they refer to, and how. We distinguish the following three categories:

o Nominal

In this category, the argument refers to an entity, usually an object or animate being, but occasionally also a place or a direction, for instance. This most commonly takes the form of a (pro)nominal group, as in (66). But it may also involve a prepositional phrase, a (split) pronominal adverb or an adverbial complement, as (the second argument) in (67), (68) resp. (69) (cf. the examples in (60–63) as well). (Note that in all three instances the second argument also expresses a directional – the elided main verb of movement is mentioned between square brackets.)

(66) Engels kunnen ze allemaal al. [lit:] ‘They all can [i.e. know] English already.’ (PDDS – fv400167 – South (CGN))

(67) Die prestatie moet in het Guinness Book of Records. [‘komen’] [lit:] ‘That accomplishment must [get] into the Guinness Book of Records.’ (PDDW – DS – South (ConDiv)) ANALYTICAL CATEGORIES 75

(68) Je kunt er toch geen uren naartoe. [‘gaan’] [lit:] ‘You cannot [go] there for hours, can you?’ (PDDS – fn000463 – North (CGN))

(69) Niemand mocht binnen. [‘gaan’] [lit:] ‘No one might [i.e. was allowed to] [go] inside.’ (PDDS – fv601222 – South (CGN))

o Deictic

In the second category, the argument refers to a state of affairs mentioned or implied in the context, by means of a (usually demonstrative) pronoun, or an equivalent, such as a pronominal adverb. (70) offers an example with a demonstrative pronoun as the first and only argument (dat ‘that’ deictically refers to beginnen wenen ‘starting to cry’; cf. also (58) above, with the pronoun dat ‘that’ deictically referring to je zusje pesten ‘bullying your little sister’). (71) is an illustration with a pronominal adverb as the second argument (d’r buiten [lit:] ‘out of it’ deictically refers to (buiten het) voor iedereen evenveel doen [lit:] ‘(out of) doing the same for everyone’).

(70) Dat hoeft niet. (beginnen wenen) [lit:] ‘That need not [i.e. is not what I want].’ ((you) starting to cry) (PDDS – fv700178 – South (CGN))

(71) Ge kunt d’r ook niet buiten hè. (buiten het voor iedereen evenveel doen) [lit:] ‘You also cannot out of it.’ (out of doing the same for everyone) (PDDS – fv700195 – South (CGN))

o State of affairs

In the third category, finally, the argument also refers to a state of affairs, like in the category ‘deictic’, but then clause-internally, by means of a complement clause, a free relative clause, a nominalization, or a nominal element that can be interpreted as signifying a complete state of affairs, as (the first and only argument) in (72a–d) respectively (in (72d), ‘ongoing traffic’ means as much as ‘to keep traffic moving’). 76 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(72) a. Het hoeft ook niet dat de scholen in antagonistische instellingen verdeeld zijn. [lit:] ‘It also need not [be] that the schools are divided into antagonistic institu- tions.’ (PDDW – Braem (2) – South (DBNL))

b. Waarom kon dit keer niet wat voor het eerste dossier tegen Di Rupo wel kon? [lit:] ‘Why could this time not [i.e. was this time not acceptable] what could for the first case against Di Rupo?’ (PDDW – DS – South (ConDiv))

c. Voor mij hoeft al dat bier zuipen niet zo. [lit:] ‘For me, drinking all that beer does not really need [i.e. is not what I want].’ (PDDS – fn008210 – North (CGN))

d. Doorgaand verkeer moet; stilstaand verkeer is altoos en overal een probleem. [lit:] ‘Ongoing traffic must [i.e. is (morally) obligatory]; stationary traffic is always and everywhere a problem.’ (PDDW – Laermans – South (DBNL))

3.1.3 Types and subtypes of uses of the modal verb14

3.1.3.1 Definitions

In line with Nuyts (2013) and Nuyts et al. (2018), we distinguished the following types of uses of the modal verb:

o Auxiliary (‘aux’)

In this category, the modal verb functions as a modifier of an explicit main verb elsewhere in the clause, as in (73).

(73) We moeten onze schoenen uitdoen in de klas. ‘We must take off our shoes in class.’ (PDDS – fn001054 – North (CGN))

14 This section is largely adopted from sections 2.2.1 and 2.2.2 in Caers & Nuyts (submitted). ANALYTICAL CATEGORIES 77

In all other categories, there is no explicit main verb elsewhere in the clause. Together they may be covered under the label ‘autonomous use’.

o Auxiliary with contextually elided main verb (‘aux V context’)

This type covers instances in which the main verb of the clause is elided because it is explicitely mentioned in the surrounding discourse. In (74), for instance, the verb doen ‘do’ is not repeated after the modal in the subordinate clause since it is already mentioned in the immediately preceding main clause. In these cases the modal is still clearly an auxiliary.15

(74) Men doet wat men kan; men speelt het spel mee voor zover mogelijk. ‘One does what one can [do], one plays the game as far as possible.’ (PDDW – Presser – North (DBNL))

o Auxiliary with implied main verb (‘aux V implicit’)

In this type, the main verb of the clause is not contextually given, but it is more or less clearly implied. Hence the modal verb can still be considered an auxiliary. In (75), e.g., one can easily imagine the elided main verb (te) gaan ‘go’ (as indicated between square brackets), although it has not been uttered in the preceding discourse.

(75) Je hoeft voor mij niet mee. [‘te gaan’] [lit:] ‘As far as I am concerned you need not [go] along.’ (PDDS – fn006846 – North (CGN))

Nearly all instances of this kind are special, though, in that making the implied main verb explicit sounds quite unnatural to native speakers. Also, note that main verb elision of this kind is unacceptable in English: the translation in (75) is impossible without the explicit presence of the main verb (te) gaan ‘go’. This indicates that these instances are qualitatively different from

15 Sometimes, the second argument has also been elliptically omitted, because of its contextual presence. We have nevertheless coded such instances as if the argument were present. For example, in Hij’s rijk, die, met zijn deel tevreden, daarvan, zoo veel hij kan, geniet ‘He is rich, who, satisfied with his part, enjoys it, as much as he can’ (ND – Loosjes and Walré – North (DBNL)), we code the instance as involving a second argument (the pronominal adverb daarvan [lit:] ‘thereof’) even if it is elided along with the main verb genieten ‘enjoy’. 78 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH those of the ‘aux V context’ type, in which elision of the verb is perfectly fine in English too (cf. Warner 1993, e.g., who uses the term “post-auxiliary ellipsis” for the phenomenon).

o Auxiliary with implied main verb or main verb itself (‘doubt’)

The third type covers instances in which it is not clear whether one can still assume an elided main verb or not, hence whether the modal is an auxiliary or is itself the main verb of the clause. In (76), for instance, one might possibly think of a vague or general verb like gebeuren ‘happen’, although adding it sounds very unnatural, even forced, and is normally not done in every day language use. Note again that this kind of use is unacceptable in English.

(76) Van ‘t Vaticaan mag ‘t niet. (het klonen van mensen) [lit:] ‘For the Vatican, it may not [happen].’ (the cloning of people) (PDDS – fn000967 – North (CGN))

o Main verb (‘main V’)

In the fourth type, finally, one cannot imagine an elided main verb anymore. Hence, the modal verb itself must be considered the main verb of the clause. In (77), e.g., one cannot insert an active main verb in the clause without causing a substantial change in meaning. Grammatically, one could insert a vague main verb like gebeuren ‘happen’, as in (76), but if one does, the meaning inevitably alters from ‘it is unacceptable’ (i.e. a deontic reading) to ‘it is not feasible’ (i.e. a dynamic reading; cf. section 3.2.1 for a detailed discussion of the different meaning categories). (The latter does not even make sense in the context: narrowing down the school system to the approach in focus is very well feasible and may even be bound to happen, and that is exactly what the author is reacting to and warning against.) Or one could add a verb in the passive voice (maybe even without a change in meaning), like gedaan worden ‘be done’, for instance (although it would sound highly unnatural), but that would introduce an additional, even if implicit, participant in the state of affairs, namely the subject of the active counterpart. In other words, this would be equivalent to adding an active verb plus an explicit subject in the clause. Such modifications go beyond mere ellipsis of a main verb, hence they are not accepted in our analyses. Only instances which allow for the addition of an active main verb without changing the meaning have been qualified as ‘aux V implicit’ or ‘doubt’ uses in our data. ANALYTICAL CATEGORIES 79

(77) Wat echter niet kan is dat men het onderwijs tot die aanpak verengt. [lit:] ‘What cannot [i.e. is not acceptable], however, is that one narrows down school education to that approach.’ (PDDW – Levrau and Timmerman – South (Google Scholar))

As mentioned in Nuyts (2013: 125), the four stages of ‘autonomous’ use described above are not sharply demarcated, but rather represent rough zones on a continuum, with gradual transitions. Hence, allocation of an instance to a specific stage is not always easy, particularly not when the possible presence of vague verbs such as zijn ‘be’ and gebeuren ‘happen’ is at stake. Nevertheless, the different stages are relevant for the present investigation, since one may assume that they correspond to the way in which the diachronic evolution between auxiliary and main verb proceeds (cf. section 4.3). The ‘main V’ type of autonomous use can be further divided into two major, grammatically different, categories (cf. also section 1.2.1.2.1), which may be regarded as the opposite ends of the gradual evolution just mentioned.

o Old main verb (‘old main V’)

The first type concerns the original main verbal use of the modals, from which the auxiliary use has emerged sometime in the more or less remote past, and which will be called the ‘old main V’ usage. (The verbs differ in terms of how far in the past the evolution has happened, cf. sections 1.2.1.2.1 and 1.2.1.2.2). Examples are rendered in (78a–b) below. As one can see, this use is transitive. This is well attested in our data (as well as in Nuyts 2013 and Nuyts et al. 2018) for kunnen (‘to know’) and hoeven (‘to need’). For mogen and moeten, however, there are no clear (pure) instances of the ‘old main V’ usage in the present corpus data (there is one for mogen in the data of Nuyts 2013, though, cf. footnote 16; and see also section 3.1.3.2 for complications). Nevertheless, their transitivity follows from the original main verbal meanings: mogen originally meant ‘to have power (over something)’, and moeten (most probably) ‘to measure (something)’ (cf. section 1.2.1.2.1), both imply an affected entity. The ‘old main V’ use typically has a (pro)nominal first argument referring to an animate, usually even human, entity, although inanimate, concrete or abstract, entities may occur as well (especially in hoeven, as illustrated in (89a) below). The second argument is typically (pro)nominal too, and refers to a concrete or 80 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH abstract entity.16 (78a–b) are thus prototypical instances of the ‘old main V’ usage (cf. also (59) in section 3.1.1, and (66) in section 3.1.2).

(78) a. Maar kunnen ze daar een beetje Engels? [lit:] ‘But can they [i.e. do they know] some English there?’ (PDDS – fn000547 – North (CGN))

b. De kreuple hoeft een stock. ‘The cripple needs a stick.’ (END – Vondel (8) – North (DBNL))

o New main verb (‘new main V’)

The other type concerns the main verbal use that has emerged from the auxiliary use (through the different intermediate stages) since END (cf. sections 1.2.1.2.1 and 1.2.1.2.2), and which will be termed the ‘new main V’ usage. This type can be divided even further into two different subtypes, which can be traced to comparable subtypes in the ‘aux’, ‘aux V context’, ‘aux V implicit’ and ‘doubt’ uses (as precursors for the respective ‘new main V’ uses). (These subtypes were not mentioned in Nuyts 2013, Nuyts et al. 2018, or Caers & Nuyts submitted.) On the one hand, in ‘new main V subtype 1’, the modal verb is intransitive, and its single argument refers to a state of affairs. Examples are rendered in (58) in section 3.1.1, in (70) and (72b–d) in section 3.1.2, as well as in (77) above. On the other hand, in ‘new main V subtype 2’, the modal verb can be both transitive and intransitive, but more importantly, the first argument expresses a (pro)nominal object referring to an entity, instead of a state of affairs. Intransitive instances of ‘new main V subtype 2’ only occur in mogen and moeten in our data and they all are of exactly the same type, featuring the directional adverb weg ‘away’, as

16 In mogen this may have been a prepositional phrase, though, as is suggested by the only ‘old main V’ instance found in the END data of Nuyts (2013): gheen vianden en moghen teghens dese machtighe stede van ierusalem [lit:] ‘no enemies may against [i.e. can concur] this mighty city of Jerusalem’. This pattern is also attested for the Old English cognate magan, as in ðeos eahsealf mæg wiÞ ælces cynnes broc on eagon ‘this eye salve can handle [lit: may with] any kind of annoyance of the eye’ (from the Dictionary of Old English Web Corpus, www.doe. utoronto.ca/pages/pub/web-corpus.html). ANALYTICAL CATEGORIES 81 illustrated in (79a–b) below (cf. also (64a–b) in section 3.1.1). (The absence of this pattern in our data for kunnen and hoeven may be accidental, though: intuitively it seems perfectly acceptable with them as well.)

(79) a. militair rationeel gezien mogen die dingen [i.e. nucleaire wapens] eigenlijk weg. [lit:] ‘From a military rational perspective these things [i.e. nuclear weapons] may [i.e. can] actually away.’ (PDDS – fv600063 – South (CGN))

b. dus ze [i.e. de gazonnekes] moeten weg. [lit:] ‘So, they [i.e. the lawns] must away.’ (PDDS – fv700028 – South (CGN))

The meaning of this idiomatic combination (of modal verb and adverb) is always the same: the entity at stake may (in cases with mogen) or has to (in cases with moeten) disappear from its current position, but what kind of disappearing it should involve is not really relevant. In (79a), for instance, the question of the fate of the nukes – should they be destroyed, or be moved to another location, and how – is not important. Similarly, in (79b) the way in which the lawns (eventually will) disappear does not matter, as long as they disappear. Although our data do not offer clues as to how this pattern has emerged, a reasonable hypothesis seems to be that it concerns an evolution from a combination in which the modal as an auxiliary supports a main verb that does specify the manner of disappearing. There are, in fact, also many instances of that type (with weg or other directionals) in the ‘aux V implicit’ category in our data for all four modals, see e.g. (91d) in section 3.1.4. Therefore, we do consider this pattern as ‘new main V subtype 2’. The transitive instances of ‘new main V subtype 2’ predominantly occur in our data for kunnen. It concerns a series of fixed (idiomatic) expressions in which the modal verb is part of a verbal compound, as in (80) and (81), or in which it combines with a prepositional phrase, as in (82).17

17 Mogen and moeten each feature one such instance in our data, namely: bij dit Leven O.H. mogen wij er niet buiten [lit:] ‘at this Life O.L. we may not [i.e. cannot] out of it’ (ND – Jonckbloet – North (DBNL)), resp. ‘k heb 82 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(80) wij konden aan niemand ons verhaal kwijt. [lit:] ‘We could not [get] rid of our story to anyone [i.e. We could not tell our story to anyone].’ (PDDS – fn000775 – North (CGN))

(81) hier [i.e. bij het bestuur voor landbouwvorming, innovatie en onderzoek] kunnen de over te dragen rijkslandbouw- en tuinbouwkundige ingenieurs terecht samen met de landbouwvorming uit het departement Economie. [lit:] ‘The state agricultural and horticultural engineers to be transferred can [go] here [i.e. to the committee for agricultural formation, innovation and investigation] together with the agricultural formation from the departement of Economics.’ (PDDS – fv600015 – South (CGN))

(82) ik kan niet tegen de hitte. [lit:] ‘I cannot against the heat [i.e. I cannot stand the heat].’ (PDDS – fn000639 – North (CGN))

At face value, these instances appear to belong to the ‘old main V’ pattern, since the latter, too, feature a transitive use with an animate entity as their first argument.18 Nevertheless, a search in the Woordenboek der Nederlandsche Taal (WNT) reveals that at least until the 18th or 19th Century combinations of the kind in (80) and (81) appeared with an explicit main verb present in the clause (even if we do not have instances of it in our own data). Thus, the WNT (s.v. kwijt, sense 2) offers a number of examples from the 18th and 19th Century of kwijt kunnen [lit:] ‘can [get] rid’ (cf. (80)), featuring the main verb (ge)raken ‘get’, including (83).

(83) Eenen teederhartigen lotgenoot, … daar ik mijne gesprekken over u altijd aan kon kwijt raken. (Feith, Werken 5, 36 [1824]; WNT, s.v. kwijt, sense 2) [lit:] ‘A soft-hearted companion,… to whom I could always get rid of my conversations about you [i.e. with whom I could always have conversations about you].’

negen kaartjes hoeveel moet ‘k er dan kwijt? [lit:] ‘I have nine cards, how many do I have to [get] rid of, then?’ (PDDS – fn000422 – North (CGN)). 18 In Caers & Nuyts (submitted), they have been considered as variants of the ‘old main V’ usage. ANALYTICAL CATEGORIES 83

Also the combination terecht kunnen ‘can [go] somewhere’, as in (81), could be accompanied by the main verb geraken ‘get’ in the 18th Century. The WNT (s.v. terecht, sense 2a) even explicitely relates it to the combination with geraken. An example is rendered in (84a). From the 19th Century onwards, the main verb in this combination seems to have disappeared, however, as illustrated by (84b).

(84) a. Zy konnen op Macassar … met geen ander geld, dan met Spaansche Realen te regte geraken. (Valentijn, O.-I. II, 1, 348 b [1724]; WNT, s.v. terecht, sense 2a) [lit:] ‘They could get on Macassar … with no other money than with Spanish Reals.’

b. Even als een reeder van Amsterdam of Antwerpen zijne O.I. schepen of ladingen te Hamburg of Londen doet verzekeren om dat hij bij de inlandsche-assurantie maatschappijen niet anders dan voor kleine gedeelten dier aanzienlijke waarden te regt kan. (Falck, Ambtsbr. 102 [1823]; WNT, s.v. terecht, sense 2a) [lit:] ‘Just like a shipowner from Amsterdam or Antwerp insures his O.I. ships or loads in Hamburg or London because he cannot [go] to the domestic insurance companies for anything other than small parts of their considerable values.’

Based on the WNT information, it seems reasonable to consider the constructions in (83) and (84a) the origins of the usage without the main verb expressed in the clause, as in (80) and (81) (and (84b)). Arguably, then, what has happened is that the main verb in the patterns in (83) and (84a) got omitted more and more often, and the non-verbal elements that used to be connected with the original main V got associated with kunnen. The same may apply to the construction in (82). Thus, in one of the WNT definitions of the adjective bestand (tegen) ‘resistant (to)’, the construction tegen iemand of iets kunnen [lit:] ‘can against someone or something’ is complemented by the verb standhouden ‘withstand’ (cf. WNT, s.v. bestand, sense 2). Moreover, the WNT (s.v. kunnen, sense 4b) suggests that other constructions in which kunnen combines with a prepositional phrase, such as, e.g., ergens bij, in, over, uit (enz.) kunnen [lit:] ‘can at, in, over, out of,… something’, can also be complemented by a main verb, namely komen ‘come’ and/or gaan ‘go’. Therefore, we consider the pattern of the kind in (82) an instance of ‘new main V subtype 2’ as well. 84 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

3.1.3.2 Complications

There are some complications, however, all pertaining to the ‘old main V’ usage in particular. First of all, as mentioned before, in our data there are no clear instances left of the original main verbal use of mogen and moeten. Nevertheless, there arguably are some relics of it (cf. also section 1.2.1.2.1). In the data of mogen there are a few instances of the type in (85a–b).

(85) a. ik mag ‘m wel. [lit:] ‘I may him [i.e. I like him / I can stand him].’ (PDDS – fn000537 – North (CGN))

b. als ge dat eten nu niet moogt gelijk dat gij zegt. [lit:] ‘If you may not [i.e. do not like] that food as you say.’ (PDDS – fv400423 – South (CGN))

In these cases, the first argument refers to a person, and the second argument either refers to a person, as in (85a), or to an object (e.g. food), as in (85b). Hence, grammatically, these instances do correspond to the ‘old main V’ pattern (cf. above). Their semantics is rather different from the original uses, though. The examples in (85a–b) no longer express the subject’s power over something, but rather indicate that the subject can(not) stand someone or something else. Yet, as mentioned in section 1.2.1.2.1, an evolution from the former to the latter meaning is not that hard to imagine: if someone has power over something, he is also able to handle it, hence in a way to stand it. Therefore, in line with the analyses in Nuyts (2013), we assume that instances of the kind in (85a–b) are a continuation of the ‘old main V’ pattern. A similar issue arises in the data for hoeven: there are some instances that have maintained the grammatical features of the original main verb but have changed to a ‘volition’ meaning, including (86).

(86) Ik hoef geen koffie. [lit:] ‘I need [i.e. want] no coffee.’ (PDDS – fn008522 – North (CGN))

ANALYTICAL CATEGORIES 85

These cases, too, have been classified as ‘old main V’. (The change from the original meaning ‘to need’ to ‘volition’ is very common.) More complicated is the (New Dutch) instance of moeten in (87) (this is the only case of this type in our data).

(87) Het schoon Land van Waes […], de welbevolkte Kempen en verschyde andere deelen van Vlaenderen en Brabant, moeten schier gants hun wezen aen onzen arbeyd. [lit:] ‘The beautiful Country of Waes […], the well-populated Kempen and several other parts of Flanders and Brabant must [i.e. owe] almost their entire being to our labor.’ (ND – Verlooy – South (DBNL))

This instance is semantically different from the original main verb (it means ‘to owe’, not ‘to measure’), and it also deviates from the latter in that it features a third argument (the prepositional phrase aen onzen arbeyd ‘to our labor’). How this use of moeten has emerged is unclear, our data offer no clues. Since it shares with the ‘old main V’ use that it takes a nominal first and second argument, however, we include it in the ‘old main V’ category. There are also a number of complications exclusively concerning grammatical deviations from the ‘old main V’ pattern. One involves instances in which the second argument refers to a full state of affairs, instead of an (abstract) entity. This occurs occasionally in all modals, except in mogen, but that may be an accidental whim of the data set. (88) offers an illustration.

(88) Maar dan moet ge met Excel werken om dat te doen en dat kan ik niet. [lit:] ‘But then you have to work with Excel to do that, and that I cannot.’ (PDDS – fv400246 – South (CGN))

In this example, the pronoun dat ‘that’ deictically refers to ‘working with Excel’. Nevertheless, there can be little doubt that this kind of use is a variant of the ‘old main V’ pattern (compare (88) with (78a), for instance), and therefore it has been handled as such in the analyses. The two remaining complications concern the ‘old main V’ uses of hoeven in particular. Firstly, this verb has (what in a somewhat deviant use of the notion may be called) an ‘ergative’ alternative to the regular pattern characterized above, in which the object of the need (i.e. 86 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH what is needed) appears as the first argument, while the experiencer of the need (i.e. what/who is in need) occurs as the second argument (in the regular pattern it is vice versa).19 The latter (i.e. the needer) sometimes appears as a prepositional phrase or a pronominal adverb, but even then it must be considered an argument, since it cannot be omitted. (89a) illustrates the regular pattern, (89b) exemplifies the ergative one (with the split pronominal adverb daertoe ‘to that’ as the second argument).

(89) a. Schoonheydt hoeft gheen schoon cieraet. ‘Beauty needs no adornments.’ (END – Jansen – North (DBNL))

b. Daer hoeft groot opmerck, en langhen tijt toe, om vol schoon stoffe, en vloeyende te wesen. [lit:] ‘To that [i.e., writing good lyrics] needs great attention and a long time, in order for it to be aesthetic and fluent.’ (END – Mander (8) – South (DBNL))

This ergative pattern occurs occasionally in END and ND, but largely disappears in PDD – there is only one remaining instance in our PDDW data. It is most probably a relic of the original (what is usually called) impersonal use of the verb behoeven ‘need’, from which hoeven has emerged in END (cf. section 1.2.1.2.2, again), in which the first argument (i.e. the needer) had dative case marking, and the second argument (i.e. the thing needed) genitive marking. If so, then the ergative pattern is probably older than the regular use in (89a), which would be an adaptation of the original impersonal use to the dominant nominative-accusative system in the language (a similar adaptation has taken place in the majority of other originally impersonal verbs in Dutch as well as in other languages, see van der Horst 1985: 34–35). Hence, instances of the ergative kind do belong in the category of ‘old main V’.

19 According to the traditional concept of the notion, ‘ergativity’ refers to a pattern in which “the subject of an intransitive clause is treated in the same way as the object of a transitive clause, and differently from [the] transitive subject” (see e.g. Dixon 1994: 1). Our use of the notion obviously does not match this definition (hence the quotation marks around the term), but it shares with it that the first argument (i.e. the subject) corresponds to what would be the second argument (i.e. the object) in a regular transitive verb. That is why we use it here, for lack of a better term. ANALYTICAL CATEGORIES 87

Secondly, transitive hoeven sometimes occurs with a first argument referring to a state of affairs, deictically or in a subject clause. Instances of this kind nearly always (and in END and PDD exclusively) feature the idiomatic expression XYZ hoeft geen betoog/uitleg ‘XYZ needs no argumentation/explanation’, whereby XYZ involves a claim or supposition. (90) offers an example.

(90) Het hoeft nauwelijks betoog dat deze grondrechten in dit calvinistische systeem veelal religieus van nature zijn. ‘It hardly needs argumentation that these basic rights in this calvinist system are mostly religious in nature.’ (PDDW – Kossmann (2) – North (DBNL))

Our data offer no indications as to how this pattern has emerged. At face value it appears to be of the regular type in (89a) (with the needer as the first argument and the thing needed as the second argument). Yet, it shares with the ‘ergative’ pattern that in the latter, too, the first argument (which corresponds to the second argument in the regular pattern) sometimes involves reference to a state of affairs (deictically or in a subject clause). Whether this signals a diachronic relationship is an open question. In any case, all instances of this kind systematically feature the meaning of the original main verbal use of behoeven (‘need’) (the pattern in (90) also commonly occurs with the latter verb), hence it seems appropriate to consider them variants of the ‘old main V’ pattern as well. All complicating cases are thus included in the ‘old main V’ category. Hence, in the following chapters, we will not discern any kinds of main V usage beyond the ‘old main V’, ‘new main V subtype 1’ and ‘new main V subtype 2’ categories defined in section 3.1.3.1. Of these, only the latter two are relevant for our present concern with the re-autonomization process in Dutch, along with the ‘aux V context’, ‘aux V implicit’ and ‘doubt’ categories. Hence, from now on we will use the label ‘new autonomous use’ to refer to this latter set of categories, i.e. at the exclusion of the ‘old main V’ uses.

88 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

3.1.4 Directionals20

Finally, as mentioned in section 1.2.1.2.3, we also marked the presence of a directional in the clause. In our definition, the category of directionals not only includes expressions of movement towards a specific end point, as in (91a–b) (cf. the instances in (63) in section 3.1.1, and in (67–69) in section 3.1.2 as well), but also, for instance, of movement across a landmark, as in (91c), or away from a starting point, as in (91d) (cf. also (61), (62) and (64a–b) in section 3.1.1), or even more vague elements of direction, as in (91e). Moreover, it not only includes physical movement, as in (91a) and (91c) (cf. (61), (63) and (68–69) as well), but also abstract movement, as in (91b) and (91e) (cf. also (67)) ((91d) is at the borderline between concrete and abstract movement, as is (62) above). (All examples in (91) are instances of the ‘aux V implicit’ type.)

(91) a. Buiten het stadion moet een groot scherm worden aangebracht, waarop de wed- strijd te zien is voor supporters die niet naar binnen mogen. [lit:] ‘Outside the stadium a big screen must be installed, on which the game can be watched by the supporters who may not [go] inside.’ (PDDW – DS – South (ConDiv))

b. Ik kan tijdens de blokuren niet op internet. [lit:] ‘I cannot [go] on the internet during block hours [meant is: working hours].’ (PDDS – fv901085 – South (CGN))

c. Eerst moest de vluchteling, die uit België vertrok, over de Frans-Belgische grens. [lit:] ‘First, the refugee, who left Belgium, must [go] across the French-Belgian border.’ (PDDW – Louyet – South (DBNL))

d. Ja maar als ik kwart voor acht kom hoeven we niet gelijk weg dan. [lit:] ‘Yes, but, if I arrive at a quarter to eight, then we need not [go] away immediately.’ (PDDS – fn000846 – North (CGN))

20 This section is largely adopted from section 2.2.3 in Caers & Nuyts (submitted). ANALYTICAL CATEGORIES 89

e. Dat project is opgehouden om de simpele reden dat we alle limieten van de samen- werking overschreden hadden en men kon gewoon op die basis niet meer verder. [lit:] ‘That project has ended for the simple reason that we had crossed all limits of the collaboration and one simply couldn’t [go] any further on that basis.’ (PDDS – fn000060 – North (CGN))

The limits of the category of directionals are not always obvious, however. For instance, an adverb such as verder ‘further’ (see (91e)) can also be used as a marker of continuitive (phasal) aspect in Dutch (as in doe maar verder ‘continue what you are doing’). That kind of use has not been included in our concept of a directional, of course. Yet, the difference between a directional and an aspectual use of this adverb is not always clear cut. In unclear cases (of this or any other type), we have given the instance the benefit of the doubt and have counted it as a directional.21

3.2 The semantic dimension 3.2.1 Meaning categories

The meaning categories used in this study are identical to those used in Byloo & Nuyts (2011, 2014), Nuyts & Byloo (2015), Nuyts (2016) and Nuyts et al. (2018). Most of them have been briefly mentioned in section 1.2.2.2. Here we provide a more detailed overview and definition of the categories relevant for the present study. Our first category concerns the original main verbal meanings ‘to know’ in kunnen, and ‘to need’ in hoeven, present in the examples in (92a–b).

(92) a. Hij kon al een beetje Spaans. [lit:] ‘He could [i.e. knew] already a bit of Spanish.’ (PDDS – fv901094 – South (CGN))

21 As mentioned in section 1.2.1.2.3, Honselaar & Olbertz (2016) also include telic adjectives in their analysis, such as dood ‘dead’ in hij moet dood [lit:] ‘he must dead [i.e. he must die]’. They consider these an analogical extension of the category of directionals, which they also assume to be telic in most cases. There are very few telic adjectives in the new autonomous uses in our data, however (none in kunnen, one in mogen, four in moeten and five in hoeven, across all periods). Even apart from that, since directionals are not necessarily telic ((91d–e), and maybe even (91c), are not, and instances of this kind are quite numerous in our data), the relevance of telic adjectives for the discussion is disputable. Therefore we have not coded them in this study. 90 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

b. Rudy Pevenage hoeft geen glazen bol, hij is er zeker van dat Jan Ullrich de Tour wint. ‘Rudy Pevenage needs no crystal ball, he is sure that Jan Ullrich wins the Tour.’ (PDDW – HBVL – South (ConDiv))

As indicated in section 3.1.3.2, we do not have instances with the original main verbal meaning of mogen and moeten, but our data do feature (what we have argued to be) semantically derived instances with the meaning ‘to like’, as in (93a), and ‘to owe’, as in (93b). (The examples are repeated from (85b) and (87) in section 3.1.3.2.)

(93) a. als ge dat eten nu niet moogt gelijk dat gij zegt. [lit:] ‘If you may not [i.e. do not like] that food as you say.’ (PDDS – fv400423 – South (CGN))

b. Het schoon Land van Waes […], de welbevolkte Kempen en verschyde andere deelen van Vlaenderen en Brabant, moeten schier gants hun wezen aen onzen arbeyd. [lit:] ‘The beautiful Country of Waes […], the well-populated Kempen and several

other parts of Flanders and Brabant must [i.e. owe] almost their entire being to our labor.’ (ND – Verlooy – South (DBNL))

Most prominent in the modal verbs are of course the modal meanings. We distinguish between four modality types: (i) dynamic, (ii) deontic, (iii) boulomaic, and (iv) epistemic.

(i) Dynamic modality involves “an indication of abilities/potentials or needs/necessities” in the state of affairs (SoA), and can be subdivided into three subtypes (Nuyts & Byloo 2015: 45):

(ia) Participant-inherent dynamic modality (‘dyn-inh’) refers to an ability or need to realize the SoA inherent in the first argument. Examples of participant-inherent ability are given in (94a–b) (but (94b) is ambiguous, also allowing the meaning of ‘dyn-imp’, see below); inherent need of the first argument is illustrated in (94c–d).

(94) a. De Discovery […] kan in zijn eentje binnen een etmaal twee maal heen en weer. [lit:] ‘The Discovery can [go] back and forth on his own twice within one day.’ (PDDW – NRC – North (ConDiv)) ANALYTICAL CATEGORIES 91

b. […] soo zal men die lucht soo luttel inhaelen als men mach. ‘In this way one will breathe that air as little as one may [i.e. can].’ (END – Pelsers – South (UGent Library))

c. Ze moest naar ‘t WC. [lit:] ‘She must [go] to the toilet.’ (PDDS – fv701305 – South (CGN))

d. ‘K hoefde bijna nooit. [lit:] ‘I almost never needed [to pee].’ (PDDS – fn007971 – North (CGN))

(ib) Participant-imposed dynamic modality (‘dyn-imp’) expresses a possibility or necessity for the first argument to realize the SoA, conditioned or imposed by the circumstances. (95a–b) offer an illustration of participant-imposed possibility, (95c–d) of participant-imposed necessity.

(95) a. We hadden geen auto, anders konden we dus de bergen in. [lit:] ‘We did not have a car, otherwise we could [go] into the mountains.’ (PDDS – fv400170 – South (CGN))

b. Soo es het ghewelt des waters doorgebroken ende heeft den keldere vervult met sulcker cracht dat noch hy noch die wateren daer niet wt en mochten. [lit:] ‘In that way, the violence of the water has broken through and has filled the

basement with such a force that neither he nor those waters might [i.e. could] [get] out of there.’ (END – Lipsius et al. – South (UGent Library))

c. Elk jong moet in het begin eenmaal naar de oppervlakte om lucht te happen voor de zwemblaas. [lit:] ‘Every young must [swim] to the surface once in the beginning in order to breathe air for the swim bladder.’ (PDDW – Hart – North (DBNL))

d. maar ik hoef gene postzegel ik woon bijna naast hem. [lit:] ‘but I need not [stick] a stamp, I almost live next to him.’ (PDDS – fv400060 – South (CGN)) 92 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(ic) Situational dynamic modality (‘dyn-sit’) involves a potential or inevitability which is not related to the first (or any other) argument in the SoA, but is inherent in the SoA as a whole. (96a–b) illustrate situational potential; (96c–d) exemplify situational inevitability.

(96) a. Het kon allemaal niet in één keer wat men bedacht. [lit:] ‘It could not [happen] all in one go what one came up with.’ (PDDS – fn000286 – North (CGN))

b. Ick heb sijne woorden genomen in sulcken sin alsse bequamelijck ghenomen konnen worden, heeft hy se anders verstaen, dat mach wel, maer ’t en gaet my niet teghen. [lit:] ‘I have taken [i.e. interpreted] his words in such a way as they can easily be taken; if he has intended them differently, that may well [be], but it does not go against me [i.e. it does not seem to me].’ (END – Cupus – North (UGent Library))

c. Als het moest, zou hij zelfs naar Zeeland rijden. [lit:] ‘If it had to, he would even drive to Zeeland.’ (PDDW – Tindemans – South (DBNL))

d. Gek zijn hoeft niet, maar het helpt. [lit:] ‘Being crazy need not, but it helps.’ (PDDW – HBVL – South (ConDiv))

The three subtypes of dynamic modality mentioned here are not rendered separately in the hierarchy of qualificational categories as presented in Figure 3 in section 1.1.2.3, since they have the same scope relations to the categories above and below them (see e.g. Nuyts & Byloo 2015: 46–47). This does not mean, however, that they do not differ mutually in terms of their scope. As argued in Nuyts & Byloo (2015: 47), we assume that “the order of presentation [given] above does reflect an increase in scope, with a [growing] role [for] the contextual and situational elements in the different variants.”

ANALYTICAL CATEGORIES 93

(ii) Deontic modality (‘deo’) involves “an indication of the degree of moral acceptability of the SoA mentioned in the clause” (Nuyts & Byloo 2015: 47). Examples are given in (97a–d).

(97) a. Zo doormodderen kan natuurlijk niet. [lit:] ‘Muddling along like that obviously cannot [i.e. is not acceptable].’ (PDDW – NRC – North (ConDiv))

b. (Mogen wy nu ons karakter van afgezanten bezoedelen?) Neen, dit mogen wy niet. [lit:] ‘(May we besmirch our character of ambassadors now?) No, we may not [do] this [meant is: that is not acceptable].’ (ND – Clercq – South (DBNL))

c. Ook het uitzicht van publieke gebouwen gehoorzaamt aan de norm dat een beetje grandeur mag, zelfs moet. [lit:] ‘The appearance of public buildings too obeys the standard that a bit of grandeur may, even must [i.e. is morally acceptable/preferable].’ (PDDW – Laermans – South (DBNL))

d. ze kan ooit opmerkingen plaatsen dat je denkt van uh Riet dat hoeft helemaal niet. [lit:] ‘Sometimes she can make remarks that make you think, uh, Riet, that needs not at all [i.e. that is not acceptable].’ (PDDS – fn000779 – North (CGN))

An important difference with most ‘traditional’ classifications is that deontic modality in this framework is considered distinct from directivity, which is the expression of permission and obligation (see below for reasons, as discussed in Nuyts et al. 2010: 17–18). Deontic modality in the present view is thus narrower than in most other classifications of modality (cf. section 1.1.2.1).

(iii) Boulomaic modality (‘boulo’) concerns “an indication of the degree of the speaker’s liking or disliking of the SoA” (Nuyts 2016: 39). It is close to the category of deontic modality (cf. the fact that it is situated at the same level as deontic modality in the scope hierarchy presented in Figure 3 in section 1.1.2.3), yet “(dis)liking something is not the same thing as (dis)approving of 94 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH something” (cf. Nuyts 2016: 39). Hence both categories are kept apart here. (98a–c) offer some illustrations. ((98a) is ambiguous with ‘deo’; (98b) is ambiguous with ‘vol’, cf. below.)

(98) a. d’r zou ook wel ‘ns een momentje sfeer in mogen. [lit:] ‘Sometimes a moment of ambiance might [come] in there too.’ (PDDS – fn007371 – North (CGN))

b. die [foto] moet d’r ook in. [lit:] ‘That [picture] must [come] in there too.’ (PDDS – fn000438 – North (CGN))

c. leest u graag een boek in één ruk uit of uh hoeft dat niet? [lit:] ‘Do you like to read a book from cover to cover or, uh, need it not?’ (PDDS – fv400129 – South (CGN))

(iv) Epistemic modality (‘epi’), finally, involves “an indication of the degree of likelihood of the SoA mentioned in the clause” (Nuyts & Byloo 2015: 48). An example is (99) (which is ambiguous with the meaning of ‘dyn-sit’).

(99) wel zou het kunnen ik zeg maar wat dat dat u misschien manisch-depressief bent en dat men dat uit uw genetica kan afleiden. [lit:] ‘Nevertheless, it could [be], I’m just saying, that perhaps you are manic- depressive and that one may deduce that from your genetics.’ (PDDS – fv400567 – South (CGN))

The difference between situational dynamic and epistemic modality is that the former indicates the presence of a potential, necessity or inevitability for the state of affairs to occur without it involving an assessment of chances, while the latter does involve a speaker assessment of chances. A category closely related to epistemic modality is evidentiality, although in many classifications only the former is considered part of the domain of modality (see, e.g., Nuyts 2017 for a discussion on the status of evidentiality). Whereas epistemic modality involves an ANALYTICAL CATEGORIES 95 estimation of the probability of the SoA, evidentiality (‘evi’) expresses “an indication of the source of information about the SoA” (Nuyts & Byloo 2015: 48). A number of subtypes can be distinguished. Relevant for us is the subtype of inferential evidentiality: an indication “that the hypothetical SoA is inferred from other, known or perceived, facts” (Nuyts & Byloo 2015: 48). (100) offers an illustration.

(100) ja dat moet. (blij zijn wanneer je dertien kilometer per uur haalt) [lit:] ‘yes, that must.’ (being happy when you reach thirteen miles an hour) (PDDS – fn000541 – North (CGN))

The modal verbs not only feature modal and related meanings, however; they can also have a (wide) range of non-modal meanings and uses (see Nuyts & Byloo 2015: 48). We distinguish three of them, all of which are considered intersubjective in the studies by Nuyts (et al.) (cf. the right hand column of Table 5 in section 1.1.2.3). Most prominent is the category of directivity (‘dir’). It concerns “the marking of a permission, obligation, interdiction, advice, etc. to realize the SoA expressed in the clause” (Nuyts & Byloo 2015: 49). As mentioned, we do not follow the traditional view of deontic modality as encompassing permission and obligation, since they are illocutionary/speech act related categories rather than expressions of modality (cf. Nuyts & Byloo 2015: 49). In fact, permission and obligation “can be argued to be precisely the same categories as expressed by the imperative mood” (Byloo & Nuyts 2014: 90; see also Nuyts et al. 2010: 17–18). (101a–d) offer some examples.

(101) a. Ik vraag aan Van Agt of ik met hem kan meerijden. Dat kan. [lit:] ‘I ask Van Agt whether I can drive with him. That can [i.e. is permitted].’ (PDDW – Thijn – North (DBNL))

b. Dat mag niet: om tien uur dient men binnen te zijn. [lit:] ‘That may not [i.e. is prohibited]: at ten o’clock one has to be inside.’ (PDDW – Mechanicus – North (DBNL))

96 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

c. Dat moet ze ook vanwege haar werk. [lit:] ‘She also must [i.e. is obliged to] [do] that because of her job.’ (PDDS – fn000693 – North (CGN))

d. Aan de Italiaanse unief hoeft namelijk niks. [lit:] ‘At the Italian university nothing namely needs [i.e. nothing is required].’ (PDDW – HBVL – South (ConDiv))

The two other non-modal uses, which are closely related to each other and to directivity (see Nuyts 2008: 197ff), are volition and intention. Volition (‘vol’) involves “the marking of a wish or desire that the SoA in the clause become real” (Nuyts & Byloo 2015: 49–50), whereas intention (‘int’) “marks the aim to realize the SoA” (Nuyts & Byloo 2015: 50). Volition has been illustrated in (102a–d) (kunnen is ambiguous with ‘deo’ here); intention is exemplified in (103a–b) (moeten is ambiguous with ‘vol’).

(102) a. Ik snap niet hoe iemand dat ooit kan. (poseren bovenop een surfplank) [lit:] ‘I do not understand how someone can ever [do] that [meant is: why someone would ever want to do that, i.e. posing on top of a surfboard].’ (PDDS – fn000382 – North (CGN))

b. (Toen heb je onder mijn rok zitten kijken.) En dat mocht niet. [lit:] ‘(At that moment you were looking under my skirt.) And that might not [meant is: I didn’t want that].’ (PDDS – fn000847 – North (CGN))

c. Moet jij ook een banaan? [lit:] ‘Must you also [have] a banana [meant is: do you also want a banana]?’ (PDDS – fn000790 – North (CGN))

d. Voor Jan Boskamp hoeft de komst van Gunther Schepens niet meer. [lit:] ‘For Jan Boskamp the arrival of Gunther Schepens needs no longer [meant is: he doesn’t want it anymore].’ (PDDW – HLN – South (ConDiv)) ANALYTICAL CATEGORIES 97

(103) a. hij loopt sneller dan ie zou mogen. [lit:] ‘He runs faster than he might [meant is: faster than he is meant to do].’ (PDDS – fn006851 – North (CGN))

b. Deutsche Telekom moest naar de beurs. [lit:] ‘Deutsche Telekom must [go] to the stock exchange [meant is: it was intended to happen].’ (PDDW – NRC – North (ConDiv))

Finally, there are some instances in which it is difficult or impossible to determine the meaning of the modal itself. Such uses, as in (104a–b), are covered under the label ‘other’ (cf. Byloo & Nuyts 2011: 20).

(104) a. Mag ik nog wat thee? [lit:] ‘May I still [have] some tea?’ (PDDS – fn000739 – North (CGN))

b. Training moet voor de volle honderd procent geconcentreerd zijn, anders hoeft het niet. [lit:] ‘Training must be focused for the full hundred percent, otherwise it needs not [i.e. there is no sense].’ (PDDW – Schenk – North (DBNL))

We have only used this category if no other meaning could be assigned, though. The idiomatic expressions with kunnen, mentioned at the end of section 3.1.3.1 (cf. (80–82)), for instance, are all classified as involving (subtypes of) dynamic modality, since the modal verb recognizably expresses an ability/possibility (in (82), e.g., the first argument is inherently unable to cope with the heat). Although all these semantic categories are clearly different, in practice they are not always easy to discern (cf. also Byloo & Nuyts 2011: 20–21). We have paid close attention to the phenomenon of ambiguity, by not forcing instances showing it into one specific semantic category, but by coding them for all their possible meanings (see Byloo & Nuyts 2014: 88). 98 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

3.2.2 Negative polarity

Finally, we also marked whether the modals occur in a negative polarity context, in order to see whether this has an effect on the evolution in their ‘autonomous’ use. (Cf. the fact that hoeven is a negative polarity item.) What may count as a negative polarity context is a matter of dispute (see, e.g., van der Wouden 1996 for one perspective on the issue, specifically related to hoeven). Examples of explicit negation in the clause undoubtedly do. They have been labeled ‘neg’ in our data. This not only encompasses regular clausal negation, but also negation incorporated in a (pro)- nominal group, as in (105a–b). And it also covers negation in the main clause when the modal verb is part of a subordinate clause, as illustrated in (105c).

(105) a. Niemand mocht binnen. [lit:] ‘Nobody might [go] inside.’ (PDDS – fv601222 – South (CGN))

b. Ik hoef geen kapstok. [lit:] ‘I need [i.e. want] no coat rack.’ (PDDS – fn000358 – North (CGN))

c. Ik snap niet hoe iemand dat ooit kan. [lit:] ‘I do not understand how someone can [i.e. would] ever [do] that.’ (PDDS – fn000382 – North (CGN))

There are, however, also many types of contexts lacking explicit negation that may nevertheless count as involving negative polarity. We will not go into the controversies over which ones do or do not (in general and with reference to hoeven specifically), the more since no problems have emerged in our analyses in this respect. We only have marked instances as ‘neg-pol’ appearing in contexts of which the negative polarity status is more or less uncontroversial. This concerns occurrences in questions, as well as ones featuring markers of what van der Wouden calls ‘weak negation’ in the clause (forms like nauwelijks ‘hardly’, slechts ‘merely’, pas ‘only’, etc.), as exemplified in (106). ANALYTICAL CATEGORIES 99

(106) Timmer hoeft pas vrijdag 13 februari het ijs op. [lit:] ‘Timmer only needs [to go] on the ice on Friday February 13th.’ (PDDW – NRC – North (ConDiv))

CHAPTER 4

The grammatical evolution

Since we want to be able to compare our present results in some detail with those obtained in the earlier investigations into the grammatical evolution of the modal verbs as reported in Nuyts (2013) and Nuyts et al. (2018) (see section 1.2.1.2) we start this chapter by presenting the relevant frequency data from those earlies studies (4.1). Then we offer the results of our own corpus investigation (4.2). Finally, we discuss whether the re-autonomization process in the Dutch modal system may be considered a case of (collective) degrammaticalization (4.3).

4.1 Relevant frequency data from Nuyts (et al.)

The relevant parts of the frequency data obtained in the studies by Nuyts (et al.) are presented in Tables 9 to 12. They only contain the data for the four modal verbs at stake in the present study, and only for the periods considered here (END, PDDW and PDDS; ND was not considered in the earlier studies). The presentation in the tables is modified (making use of the original data sheets of the earlier studies), though, in order to facilitate the comparison with our current investigation. Thus, in the tables in Nuyts (2013) the ‘aux V context’ and ‘aux V implicit’ uses were joined into one ‘aux + elided main’ category, but they are presented separately here. In the tables in Nuyts et al. (2018) cases of ‘doubt’ were included in the ‘aux V implicit’ category, but they are presented as a separate category here. And neither of the earlier articles presented separate frequency data for the old and new ‘main V’ categories, but we do so here. The tables show the absolute frequencies of the different grammatical categories, as well as their share in the full samples. Sample size is indicated in the top row.

102 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 9. Frequency of types of uses of kunnen (based on Nuyts 2013)

END PDDW PDDS n = 200 n = 200 n = 200 old main V 3 1.5% 0 0.0% 5 2.5% aux 195 97.5% 180 90.0% 160 80.0% aux V context 1 0.5% 0 0.0% 4 2.0% aux V implicit 1 0.5% 10 5.0% 18 9.0% doubt 0 0.0% 7 3.5% 6 3.0% new main V 0 0.0% 3 1.5% 7 3.5% total new auton. 2 1.0% 20 10.0% 35 17.5%

Table 10. Frequency of types of uses of mogen (based on Nuyts 2013)

END PDDW PDDS n = 200 n = 200 n = 200 old main V 1 0.5% 0 0.0% 1 0.5% aux 194 97.0% 189 94.5% 151 75.5% aux V context 4 2.0% 1 0.5% 2 1.0% aux V implicit 1 0.5% 4 2.0% 27 13.5% doubt 0 0.0% 1 0.5% 1 0.5% new main V 0 0.0% 5 2.5% 18 9.0% total new auton. 5 2.5% 11 5.5% 48 24.0%

Table 11. Frequency of types of uses of moeten (based on Nuyts 2013)

END PDDW PDDS n = 200 n = 200 n = 200 old main V 0 0.0% 0 0.0% 0 0.0% aux 196 98.0% 190 95.0% 179 89.5% aux V context 0 0.0% 0 0.0% 2 1.0% aux V implicit 4 2.0% 9 4.5% 14 7.0% doubt 0 0.0% 1 0.5% 4 2.0% new main V 0 0.0% 0 0.0% 1 0.5% total new auton. 4 2.0% 10 5.0% 21 10.5%

Table 12. Frequency of types of uses of hoeven (based on Nuyts et al. 2018)

END PDDW PDDS n = 86 n = 200 n = 200 old main V 17 19.8% 5 2.5% 6 3.0% aux 65 75.6% 179 89.5% 138 69.0% aux V context 1 1.2% 0 0.0% 0 0.0% aux V implicit 0 0.0% 6 3.0% 15 7.5% doubt 2 2.3% 2 1.0% 5 2.5% new main V 1 1.2% 8 4.0% 36 18.0% total new auton. 4 4.7% 16 8.0% 56 28.0%

THE GRAMMATICAL EVOLUTION 103

4.2 Results of the present study 4.2.1 Timing of the re-autonomization process

4.2.1.1 Different types of uses22

The frequency data of our present corpus study are rendered in Tables 13–16, in the same way as in Tables 9–12 above.23 They do not separate between the two subtypes of ‘new main V’ (cf. section 3.1.3.1), we turn to this distinction in section 4.2.1.2.

Table 13. Frequency of types of uses of kunnen

END ND PDDW PDDS n = 1000 n = 1000 n = 1000 n = 1000 old main V 8 0.8% 2 0.2% 1 0.1% 19 1.9% aux 971 97.1% 975 97.5% 931 93.1% 801 80.1% aux V context 16 1.6% 14 1.4% 9 0.9% 2 0.2% aux V implicit 1 0.1% 9 0.9% 14 1.4% 73 7.3% doubt 1 0.1% 0 0.0% 15 1.5% 28 2.8% new main V 3 0.3% 0 0.0% 30 3.0% 77 7.7% total new auton. 21 2.1% 23 2.3% 68 6.8% 180 18.0%

Table 14. Frequency of types of uses of mogen

END ND PDDW PDDS n = 1000 n = 1000 n = 1000 n = 1000 old main V 1 0.1% 0 0.0% 2 0.2% 6 0.6% aux 990 99.0% 992 99.2% 962 96.2% 770 77.0% aux V context 4 0.4% 1 0.1% 0 0.0% 1 0.1% aux V implicit 4 0.4% 6 0.6% 17 1.7% 121 12.1% doubt 1 0.1% 0 0.0% 3 0.3% 10 1.0% new main V 0 0.0% 1 0.1% 16 1.6% 92 9.2% total new auton. 9 0.9% 8 0.8% 36 3.6% 224 22.4%

Table 15. Frequency of types of uses of moeten

END ND PDDW PDDS n = 1000 n = 1000 n = 1000 n = 1000 old main V 0 0.0% 1 0.1% 0 0.0% 2 0.2% aux 984 98.4% 992 99.2% 965 96.5% 888 88.8% aux V context 1 0.1% 0 0.0% 0 0.0% 1 0.1% aux V implicit 14 1.4% 3 0.3% 15 1.5% 86 8.6% doubt 0 0.0% 4 0.4% 14 1.4% 12 1.2% new main V 1 0.1% 0 0.0% 6 0.6% 11 1.1% total new auton. 16 1.6% 7 0.7% 35 3.5% 110 11.0%

22 This subsection is largely adopted from section 3.1 in Caers & Nuyts (submitted). 23 In Caers & Nuyts (submitted) the instances of ‘doubt’ were included in the ‘aux V implicit’ category. And the instances in which kunnen is part of a verbal compound or takes a prepositional phrase (cf. (80–82) in section 3.1.3.1) were classified as ‘old main V’, while they are included in the ‘new main V’ category here. 104 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 16. Frequency of types of uses of hoeven

END ND PDDW PDDS n = 323 n = 323 n = 1000 n = 1000 old main V 76 23.5% 38 11.8% 28 2.8% 23 2.3% aux 243 75.2% 258 79.9% 891 89.1% 696 69.6% aux V context 1 0.3% 0 0.0% 0 0.0% 0 0.0% aux V implicit 0 0.0% 1 0.3% 26 2.6% 68 6.8% doubt 1 0.3% 0 0.0% 20 2.0% 67 6.7% new main V 2 0.6% 26 8.0% 35 3.5% 146 14.6% total new auton. 4 1.2% 27 8.3% 81 8.1% 281 28.1%

The tables confirm that through time there is a clear increase in the new autonomous uses in all four modals (in line with Nuyts et al.): the overall development for the total number of new autonomous uses versus the ‘old main V’ and ‘aux’ uses together is highly significant in all of them (.000, no matter whether PDDS is included or excluded). A closer look at the tables reveals some differences between the modals, however. Specifically, it shows that hoeven differs substantially from the other three modals. In kunnen, mogen and moeten, the upsurge of the new autonomous uses turns out to be very recent: in all three, it is situated after ND (i.e. after 1850). New autonomous uses are already present in END and ND, but they are fairly marginal then, the ‘doubt’ and ‘new main V’ uses even more so than the ‘aux V context’ and ‘aux V implicit’ ones. The few instances of the former are moreover all of ‘subtype 2’, with an (in)animate entity as the first argument, see below. It is not until PDD that the ‘doubt’ and ‘new main V’ categories assume a significant role in our data, along with a considerable increase of ‘aux V implicit’ instances (cf. also Tables 9– 11 above).24 (This evolution is in line with the assumption that the ‘aux V context’ and ‘aux V implicit’ uses are a lead way towards the ‘new main V’ uses through the instances of ‘doubt’, cf. section 3.1.3.1.) In addition, the data also show – even more convincingly than the earlier studies of Nuyts – that the phenomenon is not exclusively confined to the spoken material, but is present in the written language as well. As in the earlier studies, new autonomous uses are significantly more frequent in PDDS than in PDDW in all three modals (.000 in each in the current data). In the data of Nuyts (2013) the difference between END and PDDW was only significant in kunnen, however, not in mogen and moeten, but in the present study the

24 There is also an upsurge of ‘old main V’ uses of kunnen in PDDS. These are mainly of the type exemplified in (88) in section 3.1.3.2, featuring a deictic second argument. THE GRAMMATICAL EVOLUTION 105 difference between ND and PDDW is highly significant in all three modals (.000 in each; the difference between END and ND is not significant in any of them).25 As argued in Nuyts (2013), it is unlikely that instances of the ‘new main V’, ‘doubt’ and ‘aux V implicit’ types are due to ‘casual’ or ‘sloppy’ language use, and for that reason more likely to occur in spoken language. An utterance with a modal verb in which no other main verb is possible (‘new main V’), or in which one is (‘aux V implicit’) or might be (‘doubt’) possible, but is intuitively dispreferred by native speakers (cf. section 3.1.3.1), can hardly be the result of a strive to minimize effort in speaking, for instance. The higher frequency of these uses in PDDS may rather be related to the more progressive nature of spoken language, and may signal the direction in which these forms are evolving (cf. section 1.2.1.2.1). These three modals do differ mutually, however, in terms of how frequent the new autonomous uses are. In END there is (only) a significant difference between kunnen and mogen (.04). But in ND and PDDW kunnen has significantly more such uses than the two other modals (.01 in ND and .002 in PDDW for kunnen vs mogen, and .005 in ND and .001 in PDDW for kunnen vs moeten; there is no significant difference between mogen and moeten in these periods). In PDDS, finally, all three verbs differ substantially, with the highest frequency in mogen and the lowest in moeten (kunnen vs moeten and mogen vs moeten both .000, kunnen vs mogen .02). This is again very comparable to the situation in Tables 9–11 from Nuyts (2013). Hoeven, on the other hand, as a much younger modal (cf. section 1.2.1.2.2), still shows a very high share of ‘old main V’ uses in END, which then decrease drastically towards PDD. Yet, this modal also develops new autonomous uses right from its inception in END, even if their numbers are still minimal then. Their frequency increases very significantly towards ND (.000), a time slot earlier than in the other three modals (the difference with each of the other modals in ND is highly significant, .000). Also in PDDW, when the upsurge in the other modals happens, hoeven maintains the highest number of new autonomous uses (although the difference with kunnen is not significant, but the difference with mogen and moeten is very much so, .000). Remarkable is also that the rise of autonomous uses in hoeven in ND is nearly exclusively a

25 The relative frequencies of the different categories in our present study are very comparable to those in Nuyts (2013), as rendered in Tables 9–11 above, but the differences in statistical significance are due to the sample size (200 in the earlier studies, 1000 here). 106 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH matter of the ‘new main V’ type. In PDDW this use decreases again, but the ‘aux V implicit’ and ‘doubt’ uses compensate for the loss (the total number of new autonomous uses does not differ significantly between ND and PDDW). Finally, hoeven shares with the other modals that the new autonomous uses are significantly more frequent in PDDS than in PDDW (.000). It even has by far the highest number of them among the four verbs in PDDS (the difference with mogen is significant at .004, with kunnen and moeten at .000). The picture emerging from these facts, then, is that hoeven – in spite of being the junior member of the team – was the first to launch on the re-autonomization path, and is most radical in pursuing it. Kunnen follows rather closely, while mogen is much slower in joining but does peak in PDDS. Moeten, however, participates only hesitantly in the trend, even in PDDS.26 These observations might be taken to suggest that hoeven plays a leading role in the process, and that the other modals follow its example by analogy. That hoeven should be able to assume this role would be remarkable, however, in view of its very low frequency in the language as compared to the other modals. In the lemmatized Corpus Gesproken Nederlands (which counts app. 9 million words), for instance, the incidence of hoeven is only 3.39 per 10,000 words, as compared to 61.10 for kunnen, 58.55 for moeten, and 10.56 for mogen. We have no frequency information for the other stages in the present investigation, but it is significant that for END and ND we have not even remotely managed to collect full samples for hoeven (cf. Table 16), whereas this was no problem for the other three modals.27 Hence, we should be careful not to jump to quick conclusions about the role of hoeven in the re-autonomization process.

26 A key element to explain why the re-autonomization process is strongest in hoeven and weakest in moeten might be negation. For the role of negative polarity, cf. section 5.2.2. 27 We also know (see Nuyts & Byloo 2015: 62) that in Early Middle Dutch (around 1200–1300, i.e. preceding the time range covered in the present study), behoeven (as the predecessor of hoeven) had an incidence of 0.29 per 10,000 words, as compared to 5.83 for kunnen, 14.11 for moeten, and 27.50 for mogen (these counts are based on all known text materials for that period, covering app. 1.6 million words). The overall lower frequencies as compared to PDDS may be due to the difference between spoken and written language, but what can be compared are the proportions within each of the time slots/genres. Hence, while the relative frequency of the other three modals has changed considerably since then, behoeven was by far the least frequent form even then. This, too, suggests that it has been far less frequent than the other verbs all through. THE GRAMMATICAL EVOLUTION 107

4.2.1.2 Two subtypes of ‘new main V’ uses

As indicated in section 3.1.3.1, within the ‘new main V’ category one can discern two grammatically quite different subtypes (a distinction not made in the studies by Nuyts et al.). These turn out to behave very differently in the re-autonomization process. Tables 17–20 show the absolute numbers of instances in the ‘new main V subtype 1’ versus ‘new main V subtype 2’ categories, as well as their share in the total number of ‘new main V’ uses (as indicated in the top row, cf. Tables 13–16).

Table 17. Frequency of ‘new main V’ uses of subtype 1 vs. subtype 2 in kunnen

END ND PDDW PDDS n = 3 n = 0 n = 30 n = 77 subtype 1 (deictic/SoA 26 86.7% 53 68.8% first argument) subtype 2 ((pro)nominal 3 100% 4 13.3% 24 31.2% first argument)

Table 18. Frequency of ‘new main V’ uses of subtype 1 vs. subtype 2 in mogen

END ND PDDW PDDS n = 0 n = 1 n = 16 n = 92 subtype 1 (deictic/SoA 16 100% 89 96.7% first argument) subtype 2 ((pro)nominal 1 100% 3 3.3% first argument)

Table 19. Frequency of ‘new main V’ uses of subtype 1 vs. subtype 2 in moeten

END ND PDDW PDDS n = 1 n = 0 n = 6 n = 11 subtype 1 (deictic/SoA 6 100% 6 54.5% first argument) subtype 2 ((pro)nominal 1 100% 5 45.5% first argument)

Table 20. Frequency of ‘new main V’ uses of subtype 1 vs. subtype 2 in hoeven

END ND PDDW PDDS n = 2 n = 26 n = 35 n = 146 subtype 1 (deictic/SoA 2 100% 26 100% 35 100% 146 100% first argument) subtype 2 ((pro)nominal

first argument)

As the tables show, the few ‘new main V’ instances of kunnen, mogen and moeten in our END and ND data are all of ‘subtype 2’, with a (pro)nominal first argument. But in the significant 108 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH upsurge of new autonomous uses in PDD the ‘new main V’ uses of ‘subtype 1’, with the first (and only) argument referring to a state of affairs, absolutely predominate. Instances of ‘subtype 2’ play a minor role in that evolution.28 Our data for hoeven do not feature any ‘new main V subtype 2’ uses (although intuitively they would seem perfectly possible as well). But it already shows some ‘new main V subtype 1’ uses right from its inception in END, two time slots earlier than the three other modals. This again adds to the picture that emerged in section 4.2.1.1, that this verb is a forerunner in the re-autonomization process (in spite of the provisos formulated there). It seems, then, that the two alternative pathways towards the ‘new main V’ uses take a different position in the re-autonomization process. The ‘type 2’ pathway (as argued in section 3.1.3.1, emerging from auxiliary uses with the same type of valency pattern) seems to have started earlier, but plays a very moderate role in the process. The ‘type 1’ pathway (no doubt emerging from auxiliary uses featuring a first argument referring to a state of affairs, see section 4.2.2.2), however, is the main responsible in the upsurge of the re-autonomization process in PDD or (in hoeven) in ND.

4.2.2 Directionals in the re-autonomization process29

As indicated in section 1.2.1.2.3, the earlier studies by Nuyts (et al.) did not offer any suggestions as to what might be causing the re-autonomization process in the Dutch modals, but Olbertz & Honselaar (2017) have argued that it might be due to a generalization from autonomous uses in the presence of a directional (a phenomenon common in many Germanic languages). Let us see whether our present data offer support for this assumption.

4.2.2.1 The role of directionals

Tables 21 to 24 show the absolute and relative frequencies of all instances with a directional in the clause in the different types of uses of all four modals. For the autonomous uses (‘old’ and ‘new’ ones) this is based on the full set of instances of each type as mentioned in Tables 13–16. The set size is repeated in the present tables, after the slash in the columns with the absolute

28 This is except for the strong presence of ‘new main V subtype 2’ uses in kunnen in PDDS. Why instances of this type are so frequent in PDDS but not in PDDW or the earlier stages remains an open issue. 29 This section is a modified version of section 3.2 in Caers & Nuyts (submitted). THE GRAMMATICAL EVOLUTION 109 frequencies. For the ‘aux’ uses the frequency data are based on subsamples of 200 instances selected randomly from the full set of cases of this type, as also indicated in the tables. (For the sake of simplicity, the tables do not mention the difference between ‘new main V’ subtypes 1 and 2, and between their precursors in the other types of new autonomous uses, see section 4.2.2.2.)

Table 21. Frequency of directionals in types of uses of kunnen

END ND PDDW PDDS old main V 0/8 0.0% 0/2 0.0% 0/1 0.0% 0/19 0.0% aux 12/200 6.0% 9/200 4.5% 12/200 6.0% 10/200 5.0% aux V context 1/16 6.3% 1/14 7.1% 0/9 0.0% 0/2 0.0% aux V implicit 0/1 0.0% 3/9 33.3% 6/14 42.9% 22/73 30.1% doubt 0/1 0.0% - - 2/15 13.3% 2/28 7.1% new main V 0/3 0.0% - - 0/30 0.0% 2/77 2.6%

total new auton. 1/21 4.8% 4/23 17.4% 8/68 11.8% 26/180 14.4%

Table 22. Frequency of directionals in types of uses of mogen

END ND PDDW PDDS old main V 0/1 0.0% - - 0/2 0.0% 0/6 0.0% aux 20/200 10.0% 6/200 3.0% 11/200 5.5% 7/200 3.5% aux V context 0/4 0.0% 0/1 0.0% - - 0/1 0.0% aux V implicit 2/4 50.0% 1/6 16.7% 13/17 76.5% 49/121 40.5% doubt 1/1 100% - - 1/3 33.3% 2/10 20.0% new main V - - 0/1 0.0% 0/16 0.0% 3/92 3.3%

total new auton. 3/9 33.3% 1/8 12.5% 14/36 38.9% 54/224 24.1%

Table 23. Frequency of directionals in types of uses of moeten

END ND PDDW PDDS old main V - - 0/1 0.0% - - 0/2 0.0% aux 12/200 6.0% 7/200 3.5% 7/200 3.5% 10/200 5.0% aux V context 0/1 0.0% - - - - 0/1 0.0% aux V implicit 14/14 100% 1/3 33.3% 11/15 73.3% 56/86 65.1% doubt - - 3/4 75.0% 6/14 42.9% 5/12 41.7% new main V 1/1 100% - - 0/6 0.0% 4/11 36.4%

total new auton. 15/16 93.8% 4/7 57.1% 17/35 48.6% 65/110 59.1%

Table 24. Frequency of directionals in types of uses of hoeven

END ND PDDW PDDS old main V 0/76 0.0% 0/38 0.0% 0/28 0.0% 0/23 0.0% aux 8/200 4.0% 12/200 6.0% 7/200 3.5% 10/200 5.0% aux V context 0/1 0.0% ------aux V implicit - - 0/1 0.0% 15/26 57.7% 21/68 30.9% doubt 1/1 100% - - 8/20 40.0% 14/67 20.9% new main V 0/2 0.0% 0/26 0.0% 0/35 0.0% 0/146 0.0%

total new auton. 1/4 25.0% 0/27 0.0% 23/81 28.4% 35/281 12.5% 110 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

As the tables show, the situation in the usage types not at stake in the re-autonomization trend is fairly simple. In the ‘old main V’ uses, directionals are entirely absent in all time slots/genres. This is not surprising, since these main verbal predecessors of the modals do not involve movement. And also in the regular ‘aux’ uses, in which the main verb may or may not involve movement, the incidence of directionals is overall very low (roughly around 5%). It fluctuates somewhat through time in kunnen, moeten and hoeven, but the evolutions in each of these verbs are statistically not significant (no matter whether PDDS is included or not). In mogen the overall evolution is significant (.01, both with and without PDDS), due to the somewhat higher frequency of directionals in END (only END vs ND is significant, .007; ND vs PDDW and PDDW vs PDDS are not). The share of directionals in the ‘aux’ uses is moreover more or less equal in all four verbs. None of the differences between any two modals in any of the periods is statistically significant, except for mogen vs hoeven in END (.03; an effect of the higher incidence of directionals in mogen in END). But in a comparison of all four modals jointly, even the latter effect disappears: the variation within the full set of verbs is not significant in any of the time slots/genres, not even in END. We have no information regarding the incidence of directionals in language use in general, but there is no reason to assume that they are abnormal in the ‘aux’ uses of these four modals. In the new autonomous uses, however, the situation is very different and much more complex. In absolute numbers (i.e. in terms of their share in the full samples for each of the modals), the frequency of instances with a directional increases substantially in PDD in all four modals, along with the general increase of new autonomous uses in them. But in terms of the share of directionals within the set of new autonomous uses in particular, we see another picture. In spite of some fluctuations (observed in fairly small numbers of instances in the older stages, though), their frequency remains more or less constant through time in kunnen and mogen. The global evolution for all new autonomous uses together is not significant in these two verbs, no matter whether one includes PDDS or not. In moeten, however, the situation is unclear. The overall evolution is significant (.01 if PDDS is included, .005 if it is not), and this would mean a decrease in the share of the directionals. Yet there is no significant difference between any two successive slots (not even between END and ND, although the drop there would seem to be the cause for the significance of the overall evolution). In hoeven, finally, the THE GRAMMATICAL EVOLUTION 111 overall evolution is significant as well (.000 if PDDS is included, .002 if it is not), but it involves fluctuation. The difference between END and ND is not significant, but the increase of directionals from ND to PDDW (.001), and their decrease between PDDW and PDDS (.001), are very much so. How to explain this complex set of observations is not obvious. Also relevant is the fact that, in all four modals, directionals nearly exclusively occur in ‘aux V implicit’ and ‘doubt’ uses, hence their incidence in (the combination of) these two types separately is slightly higher than in the total set of new autonomous uses. With rare exceptions, they are absent in the ‘aux V context’ and ‘new main V’ instances. The few ‘new main V’ cases in kunnen, mogen and moeten are moreover exclusively of subtype 2, with a (pro)nominal first argument (cf. below). There are huge differences between the four modals, however, in how frequent directionals are in the new autonomous uses, hence in how much their share differs from that in the ‘old main V’ and ‘aux’ uses. Directionals are relatively more frequent in the new autonomous contexts in all four. But in kunnen this is only moderately so, and they never reach 20% of the new autonomous instances. The difference with the ‘old main V’ and ‘aux’ uses together is only significant in ND (.03) and PDDS (.001), but it is not in END and PDDW. In hoeven and mogen the differences appear somewhat larger, but even so directionals in new autonomous contexts never reach 30% in hoeven and 40% in mogen. Moreover, these somewhat higher frequencies do not always mean much: in both verbs, the difference with the ‘old main V’ and ‘aux’ uses together is only significant in PDDW and PDDS (.002 in hoeven in PDDS, .000 in hoeven in PDDW and in mogen in PDDW and PDDS), but it is not in END and ND. Only in moeten, the incidence of directionals is convincingly higher across the board in the new autonomous uses: it ranges from just below 50% to over 90%, and the difference with the ‘old main V’ and ‘aux’ uses together is highly significant (.000) in all periods. These observations, then, show that there is some connection between directionals and new autonomous uses in the modals, even if the extent to which this is the case differs considerably between the individual verbs. This finding should not come as a surprise, since it corresponds to the situation in other Germanic languages (cf. section 1.2.1.1).30

30 Our corpus data demonstrate, though, that the Dutch modals do not follow the trend in (some) other Germanic languages to license autonomous uses in the presence of directionals more or less exclusively in strong 112 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

4.2.2.2 Limits to the role of directionals

There are, however, several elements in our data which indicate that the directionals are not the decisive factor in causing the general re-autonomization trend in the Dutch modals (pace Olbertz & Honselaar 2017). One part of the evidence emerges from a more careful look at the frequency facts presented in the preceding (sub)section, another part from a further analysis of the role of directionals in the different ‘new main V’ instances and in their presumable ancestors. Let us first turn to the frequency facts again. If one only considers moeten in the data (as Olbertz and Honselaar have done), the high share of directionals in the new autonomous uses of this verb may make it seem plausible that instances featuring them have triggered a generalization, through analogy, towards autonomous uses in the absence of directionals. Even in this narrow perspective, however, this assumption is undermined by the fact that the share of autonomous uses with directionals is very small if considered in the context of the full range of uses of moeten. In END, for instance, they constitute only 1.5% (cf. Table 23: 15 instances) of the full set of 1000 instances. Or in PDDW, when the upsurge of new autonomous uses happens, they only constitute 1.7%. One may wonder whether that is enough to trigger an analogy effect even in moeten separately. More importantly, this analogy argument cannot be extended straightforwardly to the other modals, since the frequency of directionals in their new autonomous uses is much lower, and only in part significantly above that in the ‘aux’ uses. Moreover, while in mogen the periods with a significantly higher incidence of directionals correspond to those witnessing the upsurge of new autonomous uses (i.e. PDDW and PDDS, cf. Tables 14 and 22), this is not really the case in kunnen and hoeven. In kunnen (cf. Tables 13 and 21), the significantly higher incidence of directionals in ND is matched by a slightly more prominent role of the new autonomous uses in that period, but without this involving a real breakthrough. The real upsurge of the new autonomous uses happens in PDDW, but in that period these uses do not show a significantly

modals, not in weak ones (cf. section 1.2.1.1). Instances of this kind do occur most frequently in strong moeten. But they are also relatively somewhat more frequent in weak mogen, and they are not particularly prominent in the other strong modal in Dutch, hoeven. Whatever the reason for the exclusive link with strong modals for instance in the Scandinavian languages, then, the situation in Dutch implies that modal strength as such is not a key element in causing the relation between directionals and autonomous uses of the modals. THE GRAMMATICAL EVOLUTION 113 higher share of directionals than the other usage types. And in hoeven (cf. Tables 16 and 24), the breakthrough of new autonomous uses happens in ND, but, again, the directionals are not significantly more prominent in these uses in that period. All of this makes analogy even less obvious as a cause for the developments in each of these verbs. Moreover, one cannot really consider the process in these other modals to be a spill-over from the developments in moeten either. Moeten features a much smaller number of new autonomous uses than the other modals in PDD (cf. Tables 13–16), and the re-autonomization process starts earlier in time in hoeven and kunnen than in moeten. Hence, moeten is not leading in the re-autonomization process, but is only a backbencher which reluctantly follows the other modals. That, too, is not really compatible with the assumption that the directionals in moeten would be the cause for the process in the entire modal system. An even more important argument against the assumption that the directionals are the trigger for the entire re-autonomization process in the Dutch modals emerges if we take a closer look at their presence in the different (sub)types of new autonomous uses, and particularly in the ‘new main V’ uses and what may be assumed to be their auxiliary precursors. As observed in section 4.2.2.1, directionals are more or less confined to new autonomous uses of the ‘aux V implicit’ and ‘doubt’ type. In nearly all instances featuring a directional one can imagine a main verb expressing movement, most often the verb gaan ‘to go’ or close equivalents, which brings along the directional. (The examples of main verb omission in the context of directionals in other Germanic languages, as in (35a–f) in section 1.2.1.1, are also more or less always of this kind.) However, as mentioned, there are only very few directionals in the ‘new main V’ category, and all of these are moreover of subtype 2, with a (pro)nominal first argument (see below). Yet, ‘new main V’ is the end stage of the re-autonomization process. This observation is further strengthened if we look at the presence of directionals in the auxiliary instances of the ‘aux’, ‘aux V context’, ‘aux V implicit’, and ‘doubt’ types which may be considered to be the precursors for the ‘new main V’ uses of subtypes 1 and 2. As mentioned, we may assume that the ‘new main V subtype 1’ use (the by far most common ‘new main V’ type) has mainly emerged from auxiliary instances with a first argument referring to a state of affairs. Tables 25–28 show the frequencies of instances featuring such a first argument, respectively in the ‘aux’, ‘aux V context’, ‘aux V implicit’ and ‘doubt’ categories. The tables offer 114 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH the absolute numbers, as well as their share, in Tables 26, 27 and 28 in the total number of contextual/implicit/doubt instances per slot as mentioned in Tables 13–16 and repeated here after the slash, and in Table 25 in the subsamples of 200 purely auxiliary instances per slot that were also used for the directionals in Tables 21–24.

Table 25. Frequency of ‘aux’ instances with a first argument referring to a state of affairs

END ND PDDW PDDS kunnen 3/200 1.5% 3/200 1.5% 5/200 2.5% 5/200 2.5% mogen 5/200 2.5% 2/200 1.0% 3/200 1.5% 5/200 2.5% moeten 2/200 1.0% 4/200 2.0% 4/200 2.0% 4/200 2.0% hoeven 1/200 0.5% 2/200 1.0% 15/200 7.5% 11/200 5.5%

Table 26. Frequency of ‘aux V context’ instances with a first argument referring to a state of affairs

END ND PDDW PDDS kunnen 0/16 0.0% 0/14 0.0% 2/9 22.2% 0/2 0.0% mogen 0/4 0.0% 0/1 0.0% - - 0/1 0.0% moeten 1/1 100% - - - - 0/1 0.0% hoeven 0/1 0.0% ------

Table 27. Frequency of ‘aux V implicit’ instances with a first argument referring to a state of affairs

END ND PDDW PDDS kunnen 0/1 0.0% 2/9 22.2% 2/14 14.3% 14/73 19.2% mogen 1/4 25.0% 1/6 16.7% 1/17 5.9% 2/121 1.7% moeten 0/14 0.0% 1/3 33.3% 2/15 13.3% 5/86 5.8% hoeven - - 0/1 0.0% 9/26 34.6% 10/68 14.7%

Table 28. Frequency of ‘doubt’ instances with a first argument referring to a state of affairs

END ND PDDW PDDS kunnen 0/1 0.0% - - 9/15 60.0% 22/28 78.6% mogen 0/1 0.0% - - 2/3 66.7% 6/10 60.0% moeten - - 0/4 0.0% 7/14 50.0% 7/12 58.3% hoeven 0/1 0.0% - - 12/20 60.0% 38/67 56.7%

These tables cover both transitive and intransitive instances, but the large majority of them is intransitive, a fact which is in line with the assumption that cases of this subtype are the source for the intransitive ‘new main V’ uses of subtype 1. (For all four modals in all periods covered in Tables 25–28 together, only seven ‘aux’ and one ‘aux V implicit’ instances feature a transitive verb.) Also in line with this assumption is the fact that from ND onwards cases with a first argument referring to a state of affairs are considerably more frequent in the ‘aux V implicit’ THE GRAMMATICAL EVOLUTION 115 category, and in PDD also in the ‘doubt’ category, as compared to the ‘aux’ and ‘aux V context’ types – after all, the ‘aux V implicit’ and ‘doubt’ uses are the presumed transitory cases towards the ‘new main V’ instances (cf. above). The differences are particularly clear in PDD, simultaneously with the upsurge of the new autonomous uses (at least in kunnen, mogen and moeten; only in hoeven the link is less perfect, since the upsurge of new autonomous uses in that verb already starts in ND, see section 4.2.1.1). The differences between the four usage types, for all verbs together, are not significant in END, but they are in ND (.001) and in PDDW and PDDS (.000 in both). As for the presence of directionals in the precursors for the ‘new main V subtype 1’ uses, then, it is significant that there is not a single instance covered by Tables 25–28, in any of the periods in any of the four modals, which features one. All instances in the ‘aux’, ‘aux V context’, ‘aux V implicit’, and ‘doubt’ categories containing a directional in our data, without exception, have a (pro)nominal first argument referring to an entity, nearly always an animate/human being (cf. below). So, directionals are not only absent in the ‘new main V’ uses of subtype 1 (i.e. the by far most common subtype of ‘new main V’ instances), but also in all types of uses constituting the direct developmental pathway towards them. As indicated, there are directionals in the few instances of ‘new main V subtype 2’. Tables 29–32 compare their absolute and relative frequency to that in the precursors of the pattern. These figures only differ from those in Tables 21–24 in section 4.2.2.1 in the omission of the instances with a first argument referring to a state of affairs (cf. Tables 17–20 and 25–28). (The ‘old main V’ uses are not included in the tables since there are no directionals in them, cf. section 4.2.2.1.)

Table 29. Frequency of directionals in types of uses of kunnen with a (pro)nominal first argument

END ND PDDW PDDS aux 12/197 6.1% 9/197 4.6% 12/195 6.2% 10/195 5.1% aux V context 1/16 6.3% 1/14 7.1% 0/7 0.0% 0/2 0.0% aux V implicit 0/1 0.0% 3/7 42.9% 6/12 50.0% 22/59 37.3% doubt 0/1 0.0% - - 2/6 33.3% 2/6 33.3% new main V 0/3 0.0% - - 0/4 0.0% 2/24 8.3%

total new auton. 1/21 4.8% 4/21 19.0% 8/29 27.6% 26/91 28.6%

116 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 30. Frequency of directionals in types of uses of mogen with a (pro)nominal first argument

END ND PDDW PDDS aux 20/195 10.3% 6/198 3.0% 11/197 5.6% 7/195 3.6% aux V context 0/4 0.0% 0/1 0.0% - - 0/1 0.0% aux V implicit 2/3 66.7% 1/5 20.0% 13/16 81.3% 49/119 41.2% doubt 1/1 100% - - 1/1 100% 2/4 50.0% new main V - - 0/1 0.0% - - 3/3 100%

total new auton. 3/8 37.5% 1/7 14.3% 14/17 82.4% 54/127 42.5%

Table 31. Frequency of directionals in types of uses of moeten with a (pro)nominal first argument

END ND PDDW PDDS aux 12/198 6.1% 7/196 3.6% 7/196 3.6% 10/196 5.1% aux V context ------0/1 0.0% aux V implicit 14/14 100% 1/2 50.0% 11/13 84.6% 56/81 69.1% doubt - - 3/4 75.0% 6/7 85.7% 5/5 100% new main V 1/1 100% - - - - 4/5 80.0% total new auton. 15/15 100% 4/6 66.7% 17/20 85.0% 65/92 70.7%

Table 32. Frequency of directionals in types of uses of hoeven with a (pro)nominal first argument

END ND PDDW PDDS aux 8/199 4.0% 12/198 6.1% 7/185 3.8% 10/189 5.3% aux V context 0/1 0.0% ------aux V implicit - - 0/1 0.0% 15/17 88.2% 21/58 36.2% doubt 1/1 100% - - 8/8 100% 14/29 48.3% new main V ------total new auton. 1/2 50.0% 0/1 0.0% 23/25 92.0% 35/87 40.2%

Since there are only very few ‘aux’ instances with a first argument referring to a state of affairs in our data (cf. Table 25), the figures for this category hardly differ from those in Tables 21–24, hence we can refer to the discussion there regarding the role of directionals in this type of use. The situation in the new autonomous uses is also largely comparable to that in Tables 21– 24 (especially for END and ND), except that the share of directionals is much higher in PDD (the time slot with by far most subtype 1 instances), in PDDW even more than in PDDS, in all four modals. Consequently, there are significantly more directionals in the new autonomous than in the ‘aux’ uses of subtype 2 in PDD (written and spoken), not only in mogen, moeten and hoeven (.000, cf. Tables 22–24 in section 4.2.2.1), but also in kunnen (.001 in PDDW and .000 in PDDS, unlike in Table 21). Moreover, this not only applies to the ‘aux V implicit’ and ‘doubt’ uses, but also to the ‘new main V’ instances (the differences between the latter and the ‘aux V implicit’ and ‘doubt’ uses of type 2 in kunnen, mogen and moeten are not significant in any of the time slots, except THE GRAMMATICAL EVOLUTION 117 for ‘aux V implicit’ vs. ‘new main V’ in PDDS kunnen, .008). (For hoeven, there are no ‘new main V subtype 2’ uses in our data.) All the elements together – the detailed frequency facts and the situation of directionals in the different grammatical (sub)types of instances – indicate, then, that, even if directionals do play a role in triggering autonomous uses of the modals, specifically of subtype 2, they cannot be considered the determining factor in causing the general re-autonomization trend in the Dutch modal verbs. As mentioned in fn. 26 in section 4.2.1.1, another element that might play a role in the re- autonomization process in the Dutch modals might be negation. We will look into this factor in Chapter 5, after the semantic analysis, since the latter may be relevant for understanding the facts about negative polarity.

4.3 Reflection and discussion31

In sum, our data confirm that kunnen, mogen, moeten and hoeven are involved in a process of re-autonomization. They all show a wide range of new autonomous uses in PDD (and especially in the spoken variant of the language), although the upsurge of them happened somewhat earlier in hoeven (already in ND) than in kunnen, mogen and moeten (only after ND). Moreover, our study reconfirms that the ‘new main V’ uses in particular are predominantly of subtype 1. There is also an evolutionary pathway leading to ‘new main V’ uses of subtype 2, yet the number of instances in the end stage of it (the ‘new main V’ uses of this kind) is very small – autonomous instances of this kind are mainly of the ‘aux V implicit’ and (to a lesser extent also of the) ‘doubt’ types. Our data show that directionals do play a role in this alternative pathway, but this cannot explain the developments along the pathway towards ‘new main V’ uses of subtype 1. The question remains whether the re-autonomization process may be called a case of (collective) degrammaticalization. As already mentioned in section 1.2.1.2.3, Olbertz and Hon- selaar (2017) argue that it is not, based on their assumption that the entire process is due to instances without a main verb in the presence of a directional. Autonomous instances without a directional would then have emerged as a generalization from (i.e. in analogy with) instances

31 This section is largely adopted from section 4 in Caers & Nuyts (submitted). 118 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH with a directional. Section 4.2.2 has shown that this assumption is disputable. Yet even if it were plausible, this would not undermine the idea that we are dealing here with a case of (collective) degrammaticalization. As argued in the earlier investigations of Nuyts (2013) and Nuyts et al. (2018) on the re-autonomization of the Dutch modals (cf. sections 1.2.1.2.1 and 1.2.1.2.2), as well as in a few recent studies on their semantic evolution (see Byloo & Nuyts 2014, Nuyts & Byloo 2015, and Nuyts et al. 2018, as discussed in sections 1.2.2.2.1 and 1.2.2.2.2), analogy does play a crucial role in the evolution of the modals, semantically as well as structurally, even if in other ways than in the directionals hypothesis. Yet analogy and (de)- grammaticalization are not mutually exclusive factors – on the contrary, analogy appears to be an element steering grammaticalization (along with other factors at work in grammatical systems, such as the no-synonymy principle, cf. Nuyts & Byloo 2015), in the sense that it co- determines when, how, and how fast the process sets in or evolves (see also Fischer 2007). The reason why the re-autonomization process in the Dutch modals has to be considered a case of (collective) degrammaticalization, then, is that the available evidence indicates that it involves an evolution starting from a regular auxiliary use and ending in a ‘(new) main V’ use. The grammatical properties of the vast majority of ‘new main V’ uses (i.e. those of subtype 1) are too different from those of the ‘old main V’ uses for the former to be plausibly considered a continuation of the latter (in which case we would be dealing with retraction, not with degrammaticalization; cf. Haspelmath 2004, Norde 2009). Moreover, it is not hard to formulate a plausible hypothesis, partly on the basis of empirical facts, regarding what may have been the developments resulting in the emergence of the ‘new main V subtype 1’ pattern. Let us spell this out. As argued in section 4.2.2, the most plausible scenario is that the ‘new main V’ use of subtype 1 has evolved from regular auxiliary uses featuring a first argument referring to a state of affairs via the ‘aux V context’, ‘aux V implicit’ and ‘doubt’ uses of that kind. The evolutions leading up to this process are according to a normal grammaticalization scenario, combined with an (inter)subjectification process (see Hopper & Traugott 2003, Traugott & Dasher 2002). Auxiliary uses with this kind of pattern have become possible due to the semantic change towards more (inter)subjective meanings (in Traugott & Dasher’s 2002 sense), such as deontic, epistemic and directive uses, in all these modals (cf. section 3.2.1.1 for definitions). The original THE GRAMMATICAL EVOLUTION 119 meanings of these verbs, as well as their oldest dynamic modal meanings (ability or context- imposed possibility, or need or context-imposed necessity), pertain to the first argument participant in the clause, which is therefore typically agentive. But more subjective meanings such as deontic and epistemic modality involve a loosening and loss of that tie and pertain to the state of affairs in the clause as a whole, without any particular connection with one of the participants in it. The result of this widening of the scope, then, is that the restrictions on the nature of the first argument participant disappear, and that other types of first arguments become possible, including ones referring to states of affairs. In line with this, auxiliary uses with the first argument referring to a state of affairs are already present in Early Middle Dutch in the data for kunnen, mogen and moeten used in Byloo & Nuyts (2014) and Nuyts & Byloo (2015), when the developments towards (inter)subjective meanings are on their way in kunnen and well advanced in mogen and moeten (cf. section 1.2.2.2.1). And they first emerge in END in the data for (be)hoeven used in Nuyts et al. (2018), at the moment when the (inter)subjectification process sets in in this verb (cf. section 1.2.2.2.2). The further evolutions, however, are not in line with a grammaticalization perspective. In the Early Middle Dutch data for kunnen, mogen and moeten used in Nuyts (2013), all instances with a first argument referring to a state of affairs, without exception, feature an explicit main verb in the clause. But, as our present data also show (see Tables 25–28 above), starting reluctantly in END, and increasingly towards ND and PDD, this use starts to lose the explicit main verb.32 (The picture in hoeven is much more complex due to the very rapid changes from the ‘old main V’ use to the auxiliary use and then on to the ‘new main V’ use, which all started in END.)33

32 One might argue that main V uses with a first argument referring to a state of affairs may already have existed in OD and EMD but have gone unnoticed due to the fairly small samples in Nuyts (2013). Yet they are also entirely absent in our present much larger END and ND samples for kunnen, mogen and moeten, making this argument rather unlikely. 33 In hoeven one might also think of the ergative instances of the ‘old main V’ use as the source for the ‘new main V’ uses, since these, too, occasionally involve a first argument referring to a state of affairs (cf. section 3.1.3.2). If so, then in that verb the rise of the ‘new main V’ use would be a case of retraction, since through the history of this verb the ergative pattern exclusively occurs as a main V. What has happened in this verb, then, would only be a matter of losing the second argument. There are a few caveats, though. Firstly, ergative instances of hoeven are infrequent, and among them instances with a first argument referring to a state of affairs are a minority. And secondly, such ergative uses do not exist in the other three modals. Hence, in this 120 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

There may even be an explanation for why this evolution has happened, though. As argued in Nuyts (2014), the trigger for the process may have been iconicity. After all, the result is a pattern in which the semantic scope relationship between the modal meanings and the state of affairs affected by them is rendered clearly in linguistic structure. The evolution has moreover been made possible by the fact that in the resulting pattern there still is an affected main verb, even if it is now situated at an embedded level, subordinate to rather than grammatically heading/controlling the modal verb (i.e. iconicity), either directly, as the main verb of the subordinate clause, or indirectly, as the verb expressing the event referred to in a deictic first argument – compare (107a) with (107b) and (107c).

(107) a. Jan zou verhuisd kunnen zijn. ‘John might have moved.’

b. Het zou kunnen dat Jan verhuisd is. [lit:] ‘It might be that John has moved.’

c. Dat zou kunnen. [lit:] ‘That [i.e. John has moved] might.’

The process as described above does not fit neatly into the list of types of degrammaticalization proposed by Norde (2009). As discussed in Nuyts (2013), the structural evolution does fully correspond to what Norde (2009) calls ‘degrammation’ (cf. section 1.1.1.5), but the process deviates from the latter in that it does not involve a return to a ‘lexical’ meaning: as (briefly) shown in Nuyts (2013), the new autonomous uses have (inter)subjective meanings – cf. section 5.2 for an elaborate overview, based on the present corpus data. This, then, might be taken to mean that we are dealing with an additional type of degrammaticalization, beyond those postulated by Norde. The question is, however, whether it is appropriate to impose a semantic

scenario, one would have to assume not only that this small number of cases has ‘infected’ other uses in hoeven, but also that it has spread from this very infrequent verb to the other, much more frequent, modals (see section 4.2.1.1). This would seem far less likely than the scenario sketched in the previous and present section. Also, even if this would be the scenario, the evolution in the other modals would still start from regular auxiliary uses, hence would not be a matter of retraction in them. THE GRAMMATICAL EVOLUTION 121 criterion on the definition of types of degrammaticalization (cf. also section 1.1.1.5). After all, grammaticalization is a process of structural evolution, (inter)subjectification one of semantic evolution, and there is increasing evidence that these two, even if often coinciding, are not inherently linked – they may also occur independently of each other (cf. section 1.1.1.5). If so, then the notion of degrammaticalization should also be defined in purely structural terms, without invoking a semantic requirement. And as far as the structural evolution is concerned, if the present process in the Dutch modal system may not be called a case of (collective) degrammaticalization, then the notion of (unidirectionality in) grammaticalization loses scientific value since it becomes irrefutable (see, for instance, also Campbell (2001a), Janda (2001) and Newmeyer (2001), discussed in section 1.1.1.5, for much earlier warnings in this regards). In the following chapter, we will focus on the different meanings of the new autonomous uses of the modal verbs.

CHAPTER 5

The semantic evolution

The semantic evolution of the modal verbs has been analyzed in Byloo & Nuyts (2014), Nuyts & Byloo (2015) and Nuyts et al. (2018) (cf. sections 1.2.2.2.1 and 1.2.2.2.2). As in the previous chapter, we will first present the relevant figures of the earlier investigations (5.1). Then we will offer the results of our own corpus study, including a discussion of the role of negative polarity (5.2). Finally, we will discuss what these tell us about the re-autonomization of the modals in Dutch in terms of the processes of subjectification and intersubjectification (5.3).

5.1 General meaning profile according to Nuyts et al.

The figures from the studies by Nuyts et al. relevant for the present investigation are presented in Tables 33 to 36. They show the absolute frequencies, and the share in the full samples, of the different meanings expressed by, respectively, kunnen, mogen, moeten and hoeven. The categories are ordered in terms of their position on the scale of subjectification and intersubjectification as proposed in the earlier studies (see section 1.1.2.3). First in the list are the original, ‘objective’, main V meaning (only present in kunnen and hoeven) and the different qualificational meanings, ordered from the top down in terms of increasing subjectivity (cf. Figure 3 in section 1.1.2.3). Then follow the different intersubjective meanings, with first the three illocution-related categories (cf. Table 5 in section 1.1.2.3), and then some other intersubjective categories (‘other IS’) that were attested in the data of mogen and moeten in Byloo & Nuyts (2014) and Nuyts & Byloo (2015) (these are lacking in the present corpus material). ‘Other’ (cf. section 3.2.1) figures at the end of the table for lack of insight into the nature of the meaning of instances in this category. For the sake of simplicity, ambiguous instances in the data have been classified according to their meaning highest on the (inter)subjectification scale, 124 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH i.e. according to a ‘progressive’ count. (The picture hardly differs for a ‘conservative’ count, in which the ambiguous instances are classified according to their least (inter)subjective reading, though, see e.g. Byloo & Nuyts 2014). To the extent that the figures in the tables are required for the comparison with the present corpus data, we will refer to them in the following (sub)- sections.

Table 33. Meanings of kunnen (progressive count) (based on Byloo & Nuyts 2014; Nuyts & Byloo 2015)

END PDDW PDDS n = 200 n = 200 n = 200 ‘know’ 1 0.5% dyn-inh 87 43.5% 55 27.5% 33 16.5% dyn-imp 85 42.5% 71 35.5% 104 52.0% dyn-sit 22 11.0% 33 16.5% 25 12.5% deo 4 2.0% 24 12.0% 21 10.5% epi 1 0.5% 10 5.0% 4 2.0% dir 7 3.5% 13 6.5%

Table 34. Meanings of mogen (progressive count) (based on Byloo & Nuyts 2014; Nuyts & Byloo 2015)

END PDDW PDDS n = 200 n = 200 n = 200 dyn-inh 8 4.0% dyn-imp 72 36.0% 19 9.5% 23 11.5% dyn-sit 43 21.5% 11 5.5% 2 1.0% deo 5 2.5% 27 13.5% 26 13.0% epi 5 2.5% vol 22 11.0% 10 5.0% 9 4.5% dir 39 19.5% 115 57.5% 132 66.0% other IS 5 2.5% 18 9.0% 4 2.0% other 1 0.5% 4 2.0%

Table 35. Meanings of moeten (progressive count) (based on Byloo & Nuyts 2014; Nuyts & Byloo 2015)

END PDDW PDDS n = 200 n = 200 n = 200 dyn-inh 1 0.5% 3 1.5% dyn-imp 66 33.0% 47 23.5% 79 39.5% dyn-sit 35 17.5% 39 19.5% 9 4.5% deo 25 12.5% 38 19.0% 27 13.5% evi 5 2.5% 7 3.5% 7 3.5% vol 15 7.5% 5 2.5% 7 3.5% int 1 0.5% 12 6.0% 4 2.0% dir 52 26.0% 45 22.5% 58 29.0% other IS 2 1.0% 2 1.0% other 1 0.5% 4 2.0% 4 2.0%

THE SEMANTIC EVOLUTION 125

Table 36. Meanings of hoeven (progressive count) (based on Nuyts et al. 2018)

END PDDW PDDS n = 86 n = 200 n = 200 ‘need’ 19 22.1% 1 0.5% 2 1.0% dyn-imp 42 48.8% 81 40.5% 71 35.5% dyn-sit 7 8.1% 49 24.5% 35 17.5% deo 9 10.5% 20 10.0% 31 15.5% boulo 4 2.0% 6 3.0% vol 2 2.3% 8 4.0% 21 10.5% dir 5 5.8% 31 15.5% 33 16.5% other 2 2.3% 6 3.0% 1 0.5%

5.2 Results of the present study 5.2.1 Meaning profile of the new autonomous uses

5.2.1.1 A brief overview of all meanings and uses

The results of our present corpus study are rendered in Tables 37 to 40. They present an overview of the meanings of all types of new autonomous uses together of, respectively, kunnen, mogen, moeten and hoeven in the different time slots investigated. The meanings are ordered in terms of the same (inter)subjectification logic as in Tables 33–36. The frequencies of all meanings are expressed in terms of their relative share in the total number of new autonomous uses per time slot, as indicated in the top row of each table. The absolute numbers are mentioned between brackets, after the relative counts. Taking into account the phenomenon of ambiguity, we have made a further subdivision between meanings in ambiguous and unambiguous cases. As in Byloo & Nuyts (2014: 95ff), the central column in each period represents the frequencies of the latter, while the left and right columns demonstrate those of the former. The left column (i.e. the “most (I)S” one) indicates how often an ambiguous meaning shows up as the most (inter)subjective reading; the right column (or the “least (I)S” one) features its frequency as the least (inter)subjective alternative.34 The “total” row, finally, presents – within each period – the total share of meanings (rounded to an integer) in unambiguous respectively ambiguous cases, separated by a slash.35

34 In case of ambiguity with three or four different readings, all meanings “in the middle” have been counted both as most and as least (inter)subjective, cf. Byloo & Nuyts (2014: 99). 35 Sometimes, the total share of ambiguous instances (mentioned in the “total” row at the bottom of the table) is less than the total sum of the frequencies in the left or in the right column. This is due to the fact that – within instances showing three or four alternative readings – the meanings “in the middle” have been counted both as 126 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

In order to make the data in the tables easier to interpret, they are also rendered schematically in Graphs 1 to 4, showing the meaning evolutions of all types of new autonomous uses together of the four modal verbs investigated. The graphs only represent the ‘progressive’ counts, however (as defined in section 5.1). ‘Conservative’ graphs are not given, since their picture hardly differs from the ‘progressive’ one.

most and as least (I)S (cf. fn. 34), while in the “total” row they are only counted once. Moreover, there may be other small differences as well – both within the ambiguous and the unambiguous cases – due to rounding effects. Table 37. Meanings of all ‘new autonomous’ uses of kunnen

END ND PDDW PDDS n = 21 n = 23 n = 68 n = 180 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S dyn-inh 23.8 (5) 9.5 (2) 8.7 (2) 13.0 (3) 11.8 (8) 2.9 (2) 12.2 (22) 2.8 (5) dyn-imp 9.5 (2) 66.7 (14) 13.0 (3) 65.2 (15) 2.9 (2) 23.5 (16) 4.4 (3) 1.1 (2) 28.3 (51) 2.2 (4) dyn-sit 8.7 (2) 29.4 (20) 7.4 (5) 32.2 (58) 11.7 (21) deo 4.3 (1) 5.9 (4) 13.2 (9) 1.5 (1) 4.4 (8) 7.2 (13) 1.1 (2) epi 1.5 (1) 7.8 (14) vol 0.6 (1) dir 5.9 (4) 5.9 (4) 3.9 (7) 2.2 (4) Total 90/10 87/13 84/16 82/18

dir

vol

epi

deo

dyn-sit

dyn-imp

dyn-inh END ND PDDW PDDS

Graph 1. Meanings of all 'new autonomous’ uses of kunnen (progressive count) Table 38. Meanings of all ‘new autonomous’ uses of mogen

END ND PDDW PDDS n = 9 n = 8 n = 36 n = 224 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S dyn-inh 11.1 (1) dyn-imp 11.1 (1) 66.7 (6) 25.0 (2) 12.5 (1) 4.5 (10) dyn-sit 11.1 (1) 0.4 (1) deo 25.0 (2) 11.1 (4) 22.2 (8) 5.4 (12) 4.9 (11) boulo 0.4 (1) 1.3 (3) vol 2.8 (1) 0.4 (1) 2.2 (5) int 0.9 (2) dir 11.1 (1) 12.5 (1) 25.0 (2) 19.4 (7) 66.7 (24) 4.0 (9) 78.1 (175) other 12.5 (1) 2.2 (5) Total 89/11 88/13 78/22 95/5

dir

int

vol

boulo

deo

dyn-sit

dyn-imp END ND PDDW PDDS

Graph 2. Meanings of all 'new autonomous' uses of mogen (progressive count) Table 39. Meanings of all ‘new autonomous’ uses of moeten

END ND PDDW PDDS n = 16 n = 7 n = 35 n = 110 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S dyn-inh 2.7 (3) dyn-imp 62.5 (10) 18.8 (3) 28.6 (2) 14.3 (1) 11.4 (4) 5.7 (2) 29.1 (32) 10.9 (12) dyn-sit 6.3 (1) 22.9 (8) 2.9 (1) 3.6 (4) 13.6 (15) 3.6 (4) deo 14.3 (1) 14.3 (1) 2.9 (1) 22.9 (8) 5.7 (2) 1.8 (2) 2.7 (3) 0.9 (1) boulo 1.8 (2) evi 0.9 (1) vol 6.3 (1) 14.3 (1) 2.9 (1) 2.9 (1) 1.8 (2) 8.2 (9) int 2.9 (1) dir 18.8 (3) 12.5 (2) 28.6 (2) 8.6 (3) 25.7 (9) 10.0 (11) 27.3 (30) Total 75/25 86/14 83/17 85/15

dir int vol evi deo dyn-sit dyn-imp dyn-inh END ND PDDW PDDS

Graph 3. Meanings of all 'new autonomous' uses of moeten (progressive count) Table 40. Meanings of all ‘new autonomous’ uses of hoeven

END ND PDDW PDDS n = 4 n = 27 n = 81 n = 281 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S ‘need’ 25.0 (1) 7.4 (2) dyn-inh 0.7 (2) dyn-imp 16.0 (13) 6.2 (5) 7.8 (22) 2.1 (6) dyn-sit 25.0 (1) 50.0 (2) 7.4 (2) 63.0 (17) 25.9 (7) 22.2 (18) 7.4 (6) 23.8 (67) 11.7 (33) deo 7.4 (2) 8.6 (7) 4.9 (4) 3.7 (3) 4.3 (12) 9.3 (26) 9.3 (26) boulo 2.5 (2) 4.9 (4) 2.1 (6) 0.7 (2) 11.7 (33) vol 25.0 (1) 18.5 (5) 3.7 (1) 8.6 (7) 11.1 (9) 22.4 (63) 10.7 (30) 0.7 (2) dir 4.9 (4) 17.3 (14) 6.8 (19) 14.6 (41) other 3.7 (3) Total 75/25 67/33 78/22 68/32

dir

vol

boulo

deo

dyn-sit

dyn-imp

dyn-inh END ND PDDW PDDS

Graph 4. Meanings of all 'new autonomous' uses of hoeven (progressive count) THE SEMANTIC EVOLUTION 131

The tables and graphs indicate that there is no return to ‘objective’ meanings (unlike what Norde’s 2009 analysis of degrammaticalization would predict).36 In all four modals there is a clear increase of (inter)subjective meanings in the ‘new autonomous’ uses (in line with what was observed in Nuyts 2013). Nevertheless, there are some differences in the frequency of meanings in the different types of ‘new autonomous’ uses. In the following subsections we will discuss the profile of the ‘aux V implicit’ type (5.2.1.2) and the ‘new main V’ type (5.2.1.3), and we will compare them mutually and with the general meaning profile of the modals as presented in Tables 33–36 in section 5.1. We do not present the figures for the ‘doubt’ uses, because of their ambivalent status in between ‘aux V implicit’ and ‘new main V’ (cf. section 3.1.3.1). And we also leave aside the ‘aux V context’ uses, since they are too infrequent in most modals.

5.2.1.2 Meanings of the ‘aux V implicit’ uses

The semantic profile of the ‘aux V implicit’ uses is rendered in Tables 41–44, respectively for kunnen, mogen, moeten and hoeven. In order not to complicate matters too much, we do not distinguish between the precursors of ‘new main V’ subtypes 1 and 2. We will, however, mention relevant facts about them in the discussion below.

36 One might object that some ‘new autonomous’ uses of hoeven in END and ND still feature the original, objective meaning of ‘need’, albeit only in ambiguous instances (cf. Table 40). This actually concerns instances of the ‘new main V’ type (see section 5.2.1.3). The presence of this meaning, however, could be explained by the fact that in END and ND, simultaneously with the start of the re-autonomization/degrammaticalization process, hoeven was still grammaticalizing towards an auxiliary. Hence its original main verbal uses were still clearly present (cf. section 4.2.1.1), alongside the ‘new autonomous’ uses, and the ‘new main V’ ones in particular. This may have caused occasional interference between the two. In PDD the risk for this interference has become minimal (cf. the absence of ‘need’ in the ‘new main V’ uses there), due to the strong rise of ‘new main V’ uses and the strong decline of the ‘old main V’ ones. Supporting this analysis is the fact that the ‘new main V’ instances expressing ‘need’ in END and ND are all more or less of the type meer hoefde niet ‘more was not needed’, featuring a first argument at the edge between being nominal (referring to an object) or deictic (referring to a state of affairs). This also explains the ambiguity in these instances between the meaning of ‘need’ and ‘(situational dynamic) necessity’. Table 41. Meanings of ‘aux V implicit’ uses of kunnen

END ND PDDW PDDS n = 1 n = 9 n = 14 n = 73 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S dyn-inh 28.6 (4) 6.8 (5) 2.7 (2) dyn-imp 100 (1) 66.7 (6) 42.9 (6) 14.3 (2) 1.4 (1) 57.5 (42) 5.5 (4) dyn-sit 22.2 (2) 14.3 (2) 11.0 (8) 9.6 (7) deo 11.1 (1) 2.7 (2) 1.4 (1) 2.7 (2) epi 9.6 (7) vol 1.4 (1) dir 14.3 (2) 5.5 (4) 2.7 (2) Total 100/0 100/0 86/14 79/21

Table 42. Meanings of ‘aux V implicit’ uses of mogen

END ND PDDW PDDS n = 4 n = 6 n = 17 n = 121 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S dyn-inh dyn-imp 75.0 (3) 16.7 (1) 4.1 (5) dyn-sit 25.0 (1) 0.8 (1) deo 33.3 (2) 11.8 (2) 3.3 (4) boulo 0.8 (1) 0.8 (1) vol 0.8 (1) int 0.8 (1) dir 16.7 (1) 33.3 (2) 88.2 (15) 2.5 (3) 85.1 (103) other 16.7 (1) 4.1 (5) Total 100/0 83/17 100/0 97/3

Table 43. Meanings of ‘aux V implicit’ uses of moeten

END ND PDDW PDDS n = 14 n = 3 n = 15 n = 86 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S dyn-inh 2.3 (2) dyn-imp 71.4 (10) 14.3 (2) 66.7 (2) 26.7 (4) 13.3 (2) 36.0 (31) 11.6 (10) dyn-sit 6.7 (1) 6.7 (1) 3.5 (3) 10.5 (9) 2.3 (2) deo 33.3 (1) 6.7 (1) 13.3 (2) 6.7 (1) 1.2 (1) 1.2 (1) 1.2 (1) boulo 2.3 (2) evi vol 7.1 (1) 6.7 (1) 2.3 (2) 5.8 (5) int dir 7.1 (1) 14.3 (2) 13.3 (2) 26.7 (4) 10.5 (9) 27.9 (24) Total 86/14 100/0 73/27 84/16

Table 44. Meanings of ‘aux V implicit’ uses of hoeven

END ND PDDW PDDS n = 0 n = 1 n = 26 n = 68 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S ‘need’ dyn-inh 2.9 (2) dyn-imp 46.2 (12) 3.8 (1) 29.4 (20) 4.4 (3) dyn-sit 100 (1) 19.2 (5) 3.8 (1) 13.2 (9) 7.4 (5) deo 1.5 (1) 2.9 (2) 2.9 (2) boulo 1.5 (1) 1.5 (1) vol 7.7 (2) 4.4 (3) 13.2 (9) 1.5 (1) dir 7.7 (2) 19.2 (5) 8.8 (6) 22.1 (15) other Total 100/0 92/8 85/15

134 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

The global picture emerging from these tables is that in all four modals the semantic profile of the ‘aux V implicit’ uses is comparable to the general meaning profile of the verbs. Thus, in kunnen, in the ‘aux V implicit’ uses – like in the general meaning profile – there is a clear dominance of the dynamic modal meanings. Especially the ‘dyn-imp’ uses are highly represented (see Table 41). The share of ‘dyn-imp’ is even (slightly) higher than in the general meaning profile, but the difference is not significant, in any of the stages (compare Tables 33 and 41).37 The percentages of ‘dyn-inh’ and ‘dyn-sit’ are comparable to those in the general meaning profile as well – only in PDDS the number of ‘dyn-inh’ readings is significantly lower in the ‘aux V implicit’ uses (.05). The share of the more subjective and intersubjective meanings in the ‘aux V implicit’ uses deviates somewhat from that observed in the general meaning profile, but there are hardly any significant differences between them. The ‘deo’ uses, for instance, are clearly represented in PDDW and PDDS in the general meaning profile of kunnen (see Table 33), but they hardly occur in the ‘aux V implicit’ uses (cf. Table 41) – but the differences are not significant, neither in PDDW nor in PDDS. ‘Dir’ is more frequent in the PDDW ‘aux V implicit’ uses than in the general meaning profile, but the differences are not significant. ‘Epi’ does not occur in the ‘aux V implicit’ uses in PDDW, but it does in the PDDS data, and significantly more often than in the general meaning profile (.01).38 The share of ‘dir’ in the ‘aux V implicit’ uses in PDDS is comparable to that in the general meaning profile, the differences are not significant. Finally, we also have an instance of ‘vol’ among the ‘aux V implicit’ uses of kunnen. This case is ambiguous with ‘deo’, of the type illustrated in (103a) in section 3.2.1. In mogen the ‘aux V implicit’ uses in END especially feature a ‘dyn-imp’ reading, and to a lesser extent also a ‘dyn-sit’ reading. The differences with the general meaning profile are not significant (compare Tables 34 and 42). In the evolution towards PDD, then, ‘dyn-imp’ and ‘dyn- sit’ drastically decrease, while the ‘deo’ and in particular the ‘dir’ readings increase. ‘Deo’ is less represented in the ‘aux V implicit’ uses than in the general meaning profile, but only significantly so in PDDS (.000), not in PDDW. ‘Dir’ is more frequent in the ‘aux V implicit’ instances than in

37 The significancy values (here and below) are calculated according to the ‘progressive’ count, in order to allow comparison with the general meaning profile of the modals (see Tables 33–36). Significances according to the ‘conservative’ count are only mentioned if they differ substantially from those for the ‘progressive’ count. 38 In our PDDS data, ‘epi’ uses are always ambiguous with ‘dyn-sit’. Moreover, they are always of subtype 1, with a first argument referring to a state of affairs. THE SEMANTIC EVOLUTION 135 the general meaning profile, the difference is highly significant in both PDDW (.02) and PDDS (.000). Finally, in PDDS there are also some ‘boulo’, ‘vol’ and ‘int’ uses among the ‘aux V implicit’ instances, but they are very marginal. In the ‘aux V implicit’ uses of moeten in END, as in the general meaning profile of the verb, the focus lies on the ‘dyn-imp’ and, to a lesser extent, the ‘dir’ readings. ‘Dyn-imp’ is even significantly more frequent than in the general meaning profile (.007; compare Tables 35 and 43). In the evolution to PDDW, the share of ‘dyn-imp’ uses drastically decreases, although it remains fairly high, at a level comparable to that in the general meaning profile. The share of ‘dyn-sit’ and ‘deo’ slightly increases, the differences with the general meaning profile are not significant. ‘Dir’ also increases towards PDDW, unlike in the general meaning profile, but the difference with the latter is not significant. The share of ‘vol’ in the ‘aux V implicit’ uses remains fairly marginal in PDDW, comparable to the situation in the general meaning profile. In PDDS, ‘dyn-imp’ is more frequent again, while ‘deo’ is less frequent – the difference with the general meaning profile is only significant for ‘deo’ (.003), not for ‘dyn-imp’. ‘Dyn-sit’ is significantly more frequent in the ‘aux V implicit’ uses than in the general meaning profile in PDDS (.01). The share of ‘vol’ and ‘dir’ in the ‘aux V implicit’ uses in PDDS is more or less the same as in PDDW, and does not differ significantly from that observed in the general meaning profile. Finally, in PDDS we also have some ‘dyn-inh’ uses among the ‘aux V implicit’ instances, but – as in the general meaning profile – they are very marginal. ‘Evi’ and ‘int’ uses are not attested in the ‘aux V implicit’ instances of moeten. In hoeven, finally, the ‘aux V implicit’ uses especially feature a ‘dyn-imp’ reading, in PDDW even more than in PDDS. The difference with the general meaning profile is not significant, neither in PDDW nor in PDDS (compare Tables 36 and 44). ‘Dir’ is highly represented as well in PDD; in PDDS it is even significantly more frequent than in the general meaning profile (.01), in PDDW the difference is not significant. ‘Dyn-sit’ especially occurs in PDDW, ‘vol’ in PDDS – the differences with the general meaning profile are not significant. ‘Deo’, however, is much less represented in the ‘aux V implicit’ uses – the difference with the general meaning profile is significant in PDDS (.02), but not in PDDW. Finally, in PDDS there are also some ‘boulo’ and ‘dyn-inh’ uses among the ‘aux V implicit’ instances of hoeven, but they are very marginal.

136 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

5.2.1.3 Meanings of the ‘new main V’ uses

The semantic profile of the ‘new main V’ uses of kunnen, mogen, moeten and hoeven is rendered in Tables 45–48, respectively. Parallel to Tables 41–44, we do not distinguish between ‘new main V’ uses of subtype 1 and 2.

Table 45. Meanings of ‘new main V’ uses of kunnen

END ND PDDW PDDS n = 3 n = 0 n = 30 n = 77 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S dyn-inh 100 (3) 18.2 (14) 2.6 (2) dyn-imp 13.3 (4) 9.1 (7) dyn-sit 33.3 (10) 13.3 (4) 45.5 (35) 6.5 (5) deo 13.3 (4) 26.7 (8) 3.3 (1) 6.5 (5) 15.6 (12) epi vol dir 3.3 (1) 10.0 (3) 2.6 (2) 2.6 (2) Total 100/0 83/17 91/9

Table 46. Meanings of ‘new main V’ uses of mogen

END ND PDDW PDDS n = 0 n = 1 n = 16 n = 92 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S dyn-inh dyn-imp 100 (1) 3.3 (3) dyn-sit deo 12.5 (2) 43.8 (7) 10.9 (10) 7.6 (7) boulo 2.2 (2) vol 1.1 (1) 4.3 (4) int dir 43.8 (7) 43.8 (7) 6.5 (6) 71.7 (66) other Total 100/0 56/44 92/8

Table 47. Meanings of ‘new main V’ uses of moeten

END ND PDDW PDDS n = 1 n = 0 n = 6 n = 11 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S dyn-inh dyn-imp 100 (1) 9.1 (1) dyn-sit 66.7 (4) 27.3 (3) deo 16.7 (1) 16.7 (1) 18.2 (2) boulo evi 9.1 (1) vol 18.2 (2) int dir 100 (1) 16.7 (1) 9.1 (1) 18.2 (2) Total 0/100 83/17 91/9

Table 48. Meanings of ‘new main V’ uses of hoeven

END ND PDDW PDDS n = 2 n = 26 n = 35 n = 146 most least most least most least most least unambiguous unambiguous unambiguous unambiguous (I)S (I)S (I)S (I)S (I)S (I)S (I)S (I)S ‘need’ 50.0 (1) 7.7 (2) dyn-inh dyn-imp 2.9 (1) 0.7 (1) dyn-sit 50.0 (1) 50.0 (1) 7.7 (2) 61.5 (16) 26.9 (7) 22.9 (8) 8.6 (3) 22.6 (33) 16.4 (24) deo 7.7 (2) 11.4 (4) 2.9 (1) 8.6 (3) 6.8 (10) 13.0 (19) 13.0 (19) boulo 2.9 (1) 8.6 (3) 3.4 (5) 0.7 (1) 21.9 (32) vol 19.2 (5) 3.8 (1) 17.1 (6) 20.0 (7) 36.3 (53) 6.2 (9) 0.7 (1) dir 14.3 (5) 6.2 (9) 10.3 (15) other 8.6 (3) Total 50/50 65/35 71/29 53/47

THE SEMANTIC EVOLUTION 139

In all four modals, the semantic profile of the ‘new main V’ uses differs from the general meaning profile (cf. section 5.1) and the ‘aux V implicit’ profile (cf. section 5.2.1.2), in that they focus more on the strongly subjectified and intersubjective meanings (cf. Figure 3 and Table 5 in section 1.1.2.3). Thus, in kunnen, the share of ‘dyn-inh’ and especially of ‘dyn-imp’ uses in PDD is drastically lower in the ‘new main V’ category than in the general meaning profile or in the ‘aux V implicit’ uses (compare Tables 33, 41 and 45). The differences are (highly) significant in PDDW for ‘dyn- inh’ (.000 relative to the general meaning profile; .007 in comparison to ‘aux V implicit’) and ‘dyn-imp’ (respectively .02 and .05), and in PDDS for ‘dyn-imp’ (.000 for both semantic profiles). For ‘dyn-inh’ in PDDS, the difference with the general meaning profile is not significant, but that with the ‘aux V implicit’ profile is (.05). This involves a higher share of ‘dyn-inh’ uses than in the ‘aux V implicit’ category, though, due to the relatively high frequency of ‘new main V subtype 2’ instances (cf. (82) in section 3.1.3.1), featuring a (pro)nominal first argument, which are nearly always dynamic modal.39 In fact, all instances of ‘new main V’ with a ‘dyn-inh’ reading in our data, and the large majority of those expressing a ‘dyn-imp’ use, are of type 2. That there should be hardly any ‘new main V subtype 1’ instances with these meanings should not come as a surprise, since ‘dyn-inh’ and ‘dyn-imp’ pertain to the first argument in the clause and normally require a controlling, hence agentive, entity in that role, not a state of affairs (cf. also section 4.3). On the other hand, the more subjective ‘dyn-sit’ and ‘deo’ readings are more prominent in the ‘new main V’ uses, as compared to the general meaning profile and the ‘aux V implicit’ uses. Only for ‘dyn-sit’ in PDDW, the difference between ‘new main V’ and ‘aux V implicit’ is not significant. The other differences are (highly) significant, for both readings (for ‘dyn-sit’ in PDDW .04 relative to the general meaning profile, in PDDS .000 relative to both the general meaning profile and ‘aux V implicit’; for ‘deo’ .000 in PDDW and .02 in PDDS in comparison with the general meaning profile, and .008 in PDDW and .001 in PDDS relative to ‘aux V implicit’). ‘Dir’ is only more prominent in the ‘new main V’ uses than in the general meaning profile in PDDW (.04). It is slightly less represented in PDDS in the general meaning

39 There is only one ‘new main V subtype 2’ instance in our data that does not feature a dynamic reading, namely hij kan van mij de pot op [lit:] ‘As far as I am concerned, he can on the pot [i.e., he can go to hell]’ (from PDDS – fn006759 – North (CGN)), which is directive. 140 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH profile, and in PDDW and PDDS in the ‘aux V implicit’ profile, but the differences with the ‘new main V’ uses are not significant. ‘Epi’ and ‘vol’ uses are not attested in the ‘new main V’ instances of kunnen. In mogen, the focus of the ‘new main V’ uses strongly – and in PDDW even exclusively – lies on the more subjective ‘deo’ and even more so on the intersubjective ‘dir’ reading (see Table 46). This situation is comparable to the picture in the general meaning profile in PDD (cf. Table 34), although, ‘dir’ is even more strongly represented in the ‘new main V’ uses. The difference with the general meaning profile in the share of ‘dir’ is significant in PDDW (.02) and in PDDS (.04).40 The difference in the share of ‘deo’ is not significant, neither in PDDW nor in PDDS. As compared to ‘aux V implicit’ (see Table 42), the ‘dir’ uses are somewhat less represented in the ’new main V’ category, but the differences are not significant, neither in PDDW nor in PDDS. The ‘deo’ uses are more frequent in the ‘new main V’ instances, the difference is significant in PDDS (.000), but not in PDDW. In PDDS, there are also some ‘boulo’ and ‘vol’ uses in the ‘new main V’ instances, although they remain rather marginal (the differences with the general meaning profile and the ‘aux V implicit’ uses are not significant). ‘Dyn-imp’ is rather marginal as well in ‘new main V’ mogen (the difference with the general meaning profile is significant, .03; that with ‘aux V implicit’ is not), and it only occurs in subtype 2 with a (pro)nominal first argument. More specifically, all instances in PDDS are of the type with the directional adverb weg ‘away’ illustrated in (79a) in section 3.1.3.1.41 The other dynamic meanings, as well as the ‘other IS’ uses, are entirely absent. In the ‘new main V’ uses of moeten especially ‘dyn-sit’ is more strongly present than in the general meaning profile and ‘aux V implicit’ in PDD, while ‘dyn-imp’ is barely represented (compare Tables 35, 43 and 47). For ‘dyn-sit’, the difference with the general meaning profile is significant in PDDW and PDDS (.02 in both); the difference with ‘aux V implicit’ is significant in PDDW (.01), but not in PDDS. For ‘dyn-imp’, the difference is only significant in PDDS (.008 relative to the general meaning profile; .01 as compared to ‘aux V implicit’); in PDDW it is not. ‘Dyn-inh’ is entirely absent in ‘new main V’ moeten. ‘Evi’ and ‘vol’ seem to be more represented

40 This is only true according to the ‘progressive’ count. In the ‘new main V’ uses, the difference between the ‘progressive’ and the ‘conservative’ count in the share of ‘dir’ uses is significant in PDDW (.02; in PDDS it is not). 41 The single case in ND (expressing ‘dyn-imp’, cf. Table 46) features the prepositional phrase er buiten [lit:] ‘out of it’, as illustrated in fn. 17 in section 3.1.3.1. THE SEMANTIC EVOLUTION 141 in the PDDS ‘new main V’ uses of moeten, but the difference with the general meaning profile and the ‘aux V implicit’ uses is not significant. The share of ‘deo’ and ‘dir’, finally, is comparable to that in the general meaning profile, both in PDDW and in PDDS – the differences are not significant. As compared to ‘aux V implicit’, the ‘deo’ uses are more frequent in the ‘new main V’ instances, while the ‘dir’ uses are less frequent, but the differences are not significant, neither in PDDW nor in PDDS. The END and PDDS instances with a ‘dir’ reading (including those ambiguous with a ‘dyn-imp’ meaning) in our data are all of subtype 2 with a (pro)nominal first argument (of the type with the directional adverb weg ‘away’).42 An ‘int’ meaning is not attested in the ‘new main V’ uses of moeten in our data. In hoeven, finally, the share of ‘dyn-imp’ is drastically lower in the ‘new main V’ uses than in the general meaning profile and ‘aux V implicit’, it is barely present (cf. Tables 36, 44 and 48). The differences are not significant in END (relative to the general meaning profile; there are no ‘aux V implicit’ uses in our END data) and in ND (relative to the ‘aux V implicit’ profile; we have no ND data for the general meaning profile), but they are highly significant in PDDW and in PDDS (.000 relative to both the general meaning profile and ‘aux V implicit’). On the other hand, the share of ‘dyn-sit’ uses is significantly higher in the ‘new main V’ uses than in the general meaning profile already in END (.01). Also in ND, when the re-autonomization process breaks through in hoeven (cf. section 4.2.1.1), the ‘dyn-sit’ uses remain dominant (the difference with ‘aux V implicit’ is not significant). Moreover, ‘deo’ and ‘vol’, which were already present in END in the general meaning profile, emerge in the ‘new main V’ uses of hoeven in ND. In PDD, then, the number of ‘dyn-sit’ uses decreases in the ‘new main V’ instances, to become equally (in)- frequent as in the general meaning profile and ‘aux V implicit’ (the differences in PDDW and PDDS are not significant). The more subjective and intersubjective meanings, however, increase in the ‘new main V’ uses. The percentages of ‘deo’ and ‘dir’ are highly comparable to those in the general meaning profile, the differences are not significant in PDDW and PDDS. As compared to the ‘aux V implicit’ uses, the percentage of ‘deo’ is significantly higher in the ‘new main V’ uses in PDDS (.003), but just below significance in PDDW (.07), while the share of ‘dir’ is significantly lower in the ‘new main V’ uses in PDDS (.02), but not in PDDW. The percentages

42 The same applies to one ‘dyn-sit’ and one ‘vol’ use in PDDS, as illustrated in (79b) in section 3.1.3.1, respectively in (64b) in section 3.1.1. 142 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH of ‘boulo’ are overall very low, the differences with the general meaning profile and the ‘aux V implicit’ uses are not significant, neither in PDDW nor in PDDS. The share of ‘vol’, finally, is much higher in the ‘new main V’ instances than in the general meaning profile (.000, both in PDDW and PDDS) and ‘aux V implicit’ (.01 in PDDW, .000 in PDDS).43 In sum, while the ‘aux V implicit’ uses, like the general meaning profile of the modals, especially focus on the dynamic modal meanings, and most particularly on the ‘dyn-imp’ uses (together with a very high share of ‘dir’ uses in mogen), in the ‘new main V’ uses there is a clear tendency towards more subjective and intersubjective readings in all four modals. Especially the meanings of ‘dyn-sit’, ‘deo’, and ‘dir’, together with ‘vol’ in hoeven, play a central role. (On the implications for our understanding of the re-autonomization process, see section 5.3.)

5.2.2 The role of negative polarity in the re-autonomization process

As mentioned at the end of section 4.2.2.2, in this section we will investigate the role of negative polarity in the re-autonomization process of the Dutch modal verbs. On the one hand, we know from earlier studies that hoeven has evolved into a negative polarity item very gradually, and that moeten occurs much less often in negative polarity contexts than kunnen and mogen (see Nuyts et al. 2018). Moreover, we also know that hoeven is much more common in Northern than in Southern Dutch, and that, correspondingly, moeten dislikes negative polarity contexts more in the former than in the latter (cf. Diepeveen et al. 2006). On the other hand, the present corpus study suggests that hoeven is leading the re-autonomization process, and that kunnen, mogen and (to a lesser extent) moeten follow its example by analogy (cf. section 4.2.1.1). In view of these observations, one may wonder whether negative polarity has caused the general re-autonomization trend in the four modal verbs. The absolute and relative frequency of instances of the different types of uses of the four modal verbs involving a negative polarity context is rendered in Tables 49–52. For the ‘autonomous’ uses (both ‘old’ and ‘new’ uses) this concerns the full set of instances of each type as indicated in Tables 16 to 19 in section 4.2.1.1 (the set size is repeated in Tables 49–52, after the

43 This only applies according to the ‘progressive’ count. In the ‘new main V’ uses, the difference between the ‘progressive’ and the ‘conservative’ count in the share of ‘vol’ uses is highly significant in PDDS (.000; in PDDW it is not significant). ‘Vol’ uses in PDDS especially occur as the “most IS” reading in hoeven, particularly often ambiguously with ‘boulo’ and – to a lesser extent – ‘dyn-sit’ and ‘deo’. THE SEMANTIC EVOLUTION 143 slash in the columns with the absolute frequencies). For the ‘aux’ uses this is based on subsamples of 200 instances selected randomly from the full set of cases of this type (as also indicated in the tables) (the instances are not necessarily the same as those used in studying the role of directionals in Tables 21–24 in section 4.2.2.1). (The difference between ‘new main V’ subtypes 1 and 2, and between their precursors in the other types, will be mentioned below.)

Table 49. Frequency of negative polarity in types of uses of kunnen

END ND PDDW PDDS old main V 3/8 37.5% 2/2 100% 1/1 100% 6/19 31.6% aux 98/200 49.0% 65/200 32.5% 49/200 24.5% 48/200 24.0% aux V context 1/16 6.3% 2/14 14.3% 2/9 22.2% 0/2 0.0% aux V implicit 1/1 100% 7/9 77.8% 3/14 21.4% 27/73 37.0% doubt 1/1 100% - - 8/15 53.3% 8/28 28.6% new main V 2/3 66.7% - - 18/30 60.0% 40/77 51.9% total new auton 5/21 23.8% 9/23 39.1% 31/68 45.6% 75/180 41.7%

Table 50. Frequency of negative polarity in types of uses of mogen

END ND PDDW PDDS old main V 1/1 100% - - 1/2 50.0% 4/6 66.7% aux 41/200 20.5% 49/200 24.5% 51/200 25.5% 50/200 25.0% aux V context 0/4 0.0% 0/1 0.0% - - 0/1 0.0% aux V implicit 3/4 75.0% 3/6 50.0% 8/17 47.1% 40/121 33.1% doubt 0/1 0.0% - - 0/3 0.0% 2/10 20.0% new main V - - 1/1 100% 9/16 56.3% 47/92 51.1% total new auton 3/9 33.3% 4/8 50.0% 17/36 47.2% 89/224 39.7%

Table 51. Frequency of negative polarity in types of uses of moeten

END ND PDDW PDDS old main V - - 0/1 0.0% - - 0/2 0.0% aux 10/200 5.0% 14/200 7.0% 12/200 6.0% 13/200 6.5% aux V context 0/1 0.0% - - - - 0/1 0.0% aux V implicit 0/14 0.0% 0/3 0.0% 0/15 0.0% 2/86 2.3% doubt - - 0/4 0.0% 0/14 0.0% 2/12 16.7% new main V 0/1 0.0% - - 0/6 0.0% 4/11 36.4% total new auton 0/16 0.0% 0/7 0.0% 0/35 0.0% 8/110 7.3%

Table 52. Frequency of negative polarity in types of uses of hoeven

END ND PDDW PDDS old main V 42/76 55.3% 25/38 65.8% 28/28 100% 23/23 100% aux 169/200 84.5% 160/200 80.0% 200/200 100% 200/200 100% aux V context 0/1 0.0% ------aux V implicit - - 1/1 100% 26/26 100% 68/68 100% doubt 1/1 100% - - 20/20 100% 67/67 100% new main V 1/2 50.0% 26/26 100% 35/35 100% 146/146 100% total new auton 2/4 50.0% 27/27 100% 81/81 100% 281/281 100%

144 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

As the tables show, the picture is quite complicated. Nevertheless, a few clear patterns can be discerned. In the ‘weak’ modals, kunnen and mogen, the share of instances appearing in a negative polarity context, per type of use, is very comparable in and between both modals during their evolution from END to PDD. In the ‘old main V’ uses, the incidence of negative polarity fluctuates through time, but the differences are not significant, neither within nor between the two modals (no matter whether PDDS is included or not). In the ‘aux’ uses, however, there is a significant decrease in the share of negative polarity instances in kunnen (.000, with and without PDDS). But this is entirely due to their much higher frequency in END. The difference between END and ND is significant (.001).44 From ND onwards the frequency of negative polarity more or less stabilizes, fluctuating around 24 to 33 percent. The differences between ND and PDDW, and between PDDW and PDDS are not significant. In mogen, the share of negative polarity instances also remains more or less stable in the purely auxiliary uses. The overall evolution is not significant. The frequencies fluctuate between 20 and 26 percent, comparable to the percentages in kunnen in ND, PDDW and PDDS. The difference between kunnen and mogen in the purely auxiliary uses is only significant in END (.000), due to the much higher frequency of negative polarity instances in kunnen in that time period (cf. above). In the ‘new autonomous’ uses, finally, the incidence of negative polarity also fluctuates through time, from 24% to 46% in kunnen and from 33% to 50% in mogen, but, here too, the differences are not significant, neither within nor between the two modals (no matter whether PDDS is included or not). In line with expectation, the share of negative polarity instances in the ‘strong’ modals moeten and hoeven is completely different. In moeten there are only very small numbers of them in the ‘aux’ uses (the fluctuations through time are not significant), and in the ‘new autonomous’ uses in PDDS (but the difference between PDDW and PDDS is not even significant). In hoeven, by contrast, negative polarity is already very prominent in END, and it increases to 100% in all uses in PDD (the evolution is significant, in the ‘old main V’ and ‘aux’ uses at .000, with and without PDDS, in the ‘new autonomous’ uses at .000 with PDDS, but .001

44 As also explained in Nuyts et al. (2018: 44, note 28), it seems that kunnen originally has tended to become a negative polarity item as well. THE SEMANTIC EVOLUTION 145 without PDDS). In the ‘old main V’ and ‘aux’ uses the turn to an exclusively negative polarity status happens in the transition from ND to PDD (.000), in the ‘new autonomous’ it happens a time slot earlier, in the transition from END to ND (.01), i.e. when hoeven also shows a significant increase in its ‘new autonomous’ uses (cf. section 4.2.1.1). As for the comparison of the ‘weak’ vs. the ‘strong’ modals, then, kunnen and mogen feature more negative polarity than moeten, but less than hoeven, in all types of uses, in all different time slots. In the ‘old main V’ uses none of the differences are significant, except for that in PDDS between hoeven vs. kunnen (.000), and hoeven vs. mogen (.04). In the ‘aux’ uses, however, the differences are highly significant (.000) in all time slots, both with moeten and hoeven. In the ‘new autonomous’ uses, finally, the periods in which differences are significant correspond to those in which there are significant increases in ‘new autonomous’ uses. Thus, in spite of the absence of negative polarity in moeten, the difference with kunnen in END and ND, and with mogen in ND is just below significance (maybe the small number of ‘new autonomous’ uses in these periods plays a role here). In END the difference with mogen is significant, though (.04). But in PDDW and PDDS, when the upsurge of ‘new autonomous’ uses happens in the ‘central’ modals (cf. section 4.2.1.1), kunnen and mogen feature significantly more negative polarity than moeten (.000). The situation is more or less the same with reference to hoeven. In END, there is no significant difference with kunnen and mogen. But from ND onwards, when the ‘new autonomous’ uses significantly increase in hoeven (cf. section 4.2.1.1), this verb has significantly more negative polarity than kunnen and mogen (.000). Most important for the story of negative polarity, however, is the comparison between the different types of uses in each of the modals. In kunnen and mogen, there are no significant differences between the ‘new autonomous’ uses and the ‘old main V’ and ‘aux’ uses together in END and ND, except in kunnen in END (.04; this is due to the higher incidence of negative polarity in the ‘aux’ uses). In PDD, however, at the time of the upsurge of the ‘new autonomous’ uses, the difference between both sets of uses is significant (for kunnen, .002 in PDDW and .000 in PDDS; for mogen, .02 in PDDW and .003 in PDDS). The situation is similar in hoeven. In END there is no significant difference in negative polarity instances between the ‘new autonomous’ vs. the ‘old main V’ and ‘aux’ uses together. But in ND, when there is a significant increase in ‘new autonomous’ instances, the difference is 146 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH significant (.004). (In PDD there is no difference at all since hoeven has become exclusively a negative polarity item.) In moeten, finally, the differences between ‘new autonomous’ vs. ‘old main V’ and ‘aux’ together are never statistically significant. These observations indicate that there is a relatively strong connection between negative polarity and ‘new autonomous’ uses. The picture is even more outspoken if we look at the situation in the ‘aux V implicit’ and ‘new main V’ uses in particular (we disregard the ‘aux V context’ and ‘doubt’ uses for the same reason as explained at the end of section 5.2.1.1). As the tables show, in spite of some fluctuations, the presence of negative polarity is only slightly higher in the ‘aux V implicit’ uses, but more substantially higher in the ‘new main V’ uses, as compared to the regular ‘aux’ uses. In kunnen and mogen, there are no significant differences in END and ND between any of the three types of uses, except for ‘aux’ vs. ‘aux V implicit’ in ND kunnen (.009) and in END mogen (.03). More important, however – in view of the re-autonomization process – are the data in PDDW and PDDS. In both kunnen and mogen the ‘aux V implicit’ uses do not feature significantly more negative polarity than the ‘aux’ uses, except for kunnen in PDDS (.05). The ‘new main V’ uses, by contrast, do feature significantly more negative polarity than the ‘aux’ uses (in kunnen, .000 in PDDW and PDDS; in mogen, .02 in PDDW and .000 in PDDS). Moreover, in PDDW kunnen and in PDDS mogen, the ‘new main V’ uses also feature significantly more negative polarity than the ‘aux V implicit’ uses (.02 resp. .01). The difference in PDDS kunnen and in PDDW mogen is not significant, though. In moeten, there are – because of the lack of negative polarity in the ‘new autonomous’ uses in our written data – no significant differences between each of the three types of uses, in any of the three written stages (END, ND and PDDW). In PDDS, however, the stage in which we do find negative polarity instances also in the ‘new autonomous’ uses, the difference between ‘aux’ and ‘aux V implicit’ is not significant, but that between ‘aux’ and ‘new main V’ (.007) as well as that between ‘aux V implicit’ and ‘new main V’ (.001) is highly significant. In hoeven, finally, there is (only) a significant difference in the share of negative polarity between ‘aux’ and ‘new main V’ in ND (.006); the other differences (between any of the three types of uses) are not significant, in any of the periods. So, in spite of major differences in the share of negative polarity between the four modal verbs (cf. above), in all of them there appears to be a tendency towards more negative polarity THE SEMANTIC EVOLUTION 147 in the ‘new main V’ uses than in the ‘aux’ uses, at the time of the upsurge of the ‘new autonomous’ uses, i.e. in PDD for kunnen, mogen and moeten, and in ND for hoeven. The ‘aux V implicit’ uses take a position in between, but are closer to the regular ‘aux’ uses. Let us see whether the higher share of negative polarity in the ‘new main V’ uses of the modals is connected to instances of ‘subtype 1’ vs. ‘subtype 2’, and how this relates to their precursors in the ‘aux’ and ‘aux V implicit’ types. Tables 53–55 show the frequency of negative polarity instances in both subtypes in, respectively, the ‘aux’, ‘aux V implicit’ and ‘new main V’ categories. The frequencies are rendered in absolute numbers and in terms of their share in the total number of instances in each category, as indicated after the slash.

Table 53. Frequency of negative polarity in ‘aux’ instances of subtype 1 vs. subtype 2

END ND PDDW PDDS kunnen subtype 1 5/9 55.6% 2/3 66.7% 1/7 14.3% 1/9 11.1% subtype 2 93/191 48.7% 63/197 32.0% 48/193 24.9% 47/191 24.6% mogen subtype 1 2/6 33.3% 0/4 0.0% 1/6 16.7% 4/7 57.1% subtype 2 39/194 20.1% 49/196 25.0% 50/194 25.8% 46/193 23.8% moeten subtype 1 0/3 0.0% 1/3 33.3% 1/7 14.3% 1/4 25.0% subtype 2 10/197 5.1% 13/197 6.6% 11/193 5.7% 12/196 6.1% hoeven subtype 1 2/3 66.7% 2/2 100% 15/15 100% 6/6 100% subtype 2 167/197 84.8% 158/198 79.8% 185/185 100% 194/194 100%

Table 54. Frequency of negative polarity in ‘aux V implicit’ instances of subtype 1 vs. subtype 2

END ND PDDW PDDS kunnen subtype 1 2/2 100% 0/2 0.0% 1/14 7.1% subtype 2 1/1 100% 5/7 71.4% 3/12 25.0% 26/59 44.1% mogen subtype 1 0/1 0.0% 0/1 0.0% 1/1 100% 0/2 0.0% subtype 2 3/3 100% 3/5 60.0% 7/16 43.8% 40/119 33.6% moeten subtype 1 0/1 0.0% 0/2 0.0% 0/5 0.0% subtype 2 0/14 0.0% 0/2 0.0% 0/13 0.0% 2/81 2.5% hoeven subtype 1 9/9 100% 10/10 100% subtype 2 1/1 100% 17/17 100% 58/58 100%

148 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 55. Frequency of negative polarity in ‘new main V’ instances of subtype 1 vs. subtype 2

END ND PDDW PDDS kunnen subtype 1 15/26 57.7% 24/53 45.3% subtype 2 2/3 66.7% 3/4 75.0% 16/24 66.7% mogen subtype 1 9/16 56.3% 47/89 52.8% subtype 2 1/1 100% 0/3 0.0% moeten subtype 1 0/6 0.0% 4/6 66.7% subtype 2 0/1 0.0% 0/5 0.0% hoeven subtype 1 1/2 50.0% 26/26 100% 35/35 100% 146/146 100% subtype 2

The percentages strongly fluctuate, but the differences between the two subtypes, per modal verb, are not significant. (Only in the ‘aux V implicit’ uses of kunnen in PDDS, there is significantly more negative polarity in ‘subtype 2’ than in ‘subtype 1’ uses, .01). In other words, negative polarity instances are not particularly connected to ‘subtype 1’ or ‘subtype 2’ uses. One may wonder whether the higher share of negative polarity instances in the ‘new main V’ uses might have something to do with the meanings. Could it be that, quite in general, ‘dyn- sit’, ‘deo’ and ‘dir’, as the dominant meanings of the ‘new main V’ uses in PDD (cf. Tables 45– 48 above) occur more often in negative polarity contexts than the other meanings of the modals? (We do not consider ‘vol’ here, since it only occurs in hoeven as a dominant meaning.) Let us focus on PDD, as the time slot with the by far largest number of ‘new main V’ uses, and on kunnen and mogen (in view of the special situation in moeten and hoeven in regards negative polarity). We have no semantic analyses for the ‘aux’ uses in the present investigation, but we can use the facts about the incidence of negative polarity in the different meanings across usage types obtained in the study by Byloo & Nuyts (2014).45 They are presented in Tables 56–57 for kunnen and mogen in PDDW and PDDS. These show the absolute frequencies, as well as their share in the total number of instances with each meaning, as indicated in Tables 33 and 34 in section 5.1, and repeated here after the slash. To see more clearly whether negative polarity is

45 The same criteria for classifying an instance as involving negative polarity as used in the present investigation (cf. section 3.2.2) have also been applied in the data from Byloo & Nuyts. There are no significant differences between the latter and the present study in the overall incidence of negative polarity contexts. THE SEMANTIC EVOLUTION 149 more frequent in the meanings predominating in the ‘new main V’ uses, viz. ‘dyn-sit’, ‘deo’ and ‘dir’, these are set apart in the tables.

Table 56. Frequency of negative polarity in all different meanings of kunnen (Byloo & Nuyts; progr. count)

PDDW PDDS Meanings dominant in 24/64 37.5% 21/59 35.6% ‘new main V’ dyn-sit 11/33 33.3% 3/25 12.0% deo 12/24 50.0% 16/21 76.2% dir 1/7 14.3% 2/13 15.4% Meanings marginal or absent 42/136 30.9% 54/141 38.3% in ‘new main V’ dyn-inh 22/55 40.0% 15/33 45.5% dyn-imp 20/71 28.2% 39/104 37.5% epi 0/10 0.0% 0/4 0.0% Grand total 66/200 33.0% 75/200 37.5%

Table 57. Frequency of negative polarity in all different meanings of mogen (Byloo & Nuyts; progr. count)

PDDW PDDS Meanings dominant in 61/153 39.9% 51/160 31.9% ‘new main V’ dyn-sit 8/11 72.7% 2/2 100% deo 20/27 74.1% 13/26 50.0% dir 33/115 28.7% 36/132 27.3% Meanings marginal or absent 14/47 29.8% 12/40 30.0% in ‘new main V’ dyn-imp 5/19 26.3% 5/23 21.7% vol 7/10 70.0% 7/9 77.8% other IS 2/18 11.1% 0/4 0.0% other - - 0/4 0.0% Grand total 75/200 37.5% 63/200 31.5%

As the tables show, the share of negative polarity in the meanings dominating in ‘new main V’ is very comparable to that in the other meanings. None of the differences, in any of the slots (kunnen or mogen, PDDW or PDDS), is significant. This means that the higher share of negative polarity in the ‘new main V’ uses cannot be explained by reference to the meaning profile of the latter. Since it is not obvious what other elements may have caused the higher incidence of negative polarity contexts in the ‘new autonomous’ uses, and especially the ‘new main V’ uses, then, it seems we must conclude that negative polarity does correlate in some direct way with the re-autonomization process in the Dutch modals. The question how precisely it may play a role remains an open issue. 150 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

5.3 Conclusion

In sum, our data confirm (cf. Nuyts 2013: 130–131) that the ‘new autonomous’ uses of the Dutch modals tend to focus on strongly (inter)subjectivized meanings (in Traugott & Dasher’s 2002 sense, i.e. on the meanings and uses highest in the (inter)subjectification hierarchy, in Byloo & Nuyts’ 2011, 2014, and Nuyts & Byloo’s 2015 interpretation of Traugott’s notions, cf. section 1.1.2.3), even more so than the regular auxiliary uses. In all four modals, this tendency is most outspoken in the ‘new main V’ uses, as a subcategory of the ‘new autonomous’ uses. The semantic profile of the ‘aux V implicit’ uses is more comparable to the general meaning profile of the verbs. Hence the ‘new main V’ uses do not return to ‘objective’ main verbal meanings (unlike what is generally assumed in a degrammaticalization scenario, cf. Norde 2009). This strengthens the assumption that they have emerged from auxiliary uses with these meanings, and not from the original main verbal uses. Our data also show that in all four modals there are more negative polarity instances in the ‘new autonomous’ uses, and especially in the ‘new main V’ ones, than in the ‘aux’ uses. This is especially true at the moment of the upsurge of the ‘new autonomous’ uses (i.e. in PDD in the ‘central’ modals, and in ND in hoeven). It seems that negative polarity plays a role in the re- autonomization process of the modal verbs in Dutch. This could also explain why the process is (first and) strongest in hoeven (as a negative polarity item), and weakest in moeten (as a verb that ‘dislikes’ negation) (cf. section 4.2.1.1). The reason for the higher share of negative polarity instances in the ‘new main V’ uses is not clear, however. Although the more subjective (‘dyn- sit’, ‘deo’) and intersubjective (‘dir’) readings (have been shown to) stand out as the dominant meanings of the ‘new main V’ uses, there is no particular connection between them and negative polarity. The question why and how negative polarity plays a role in the process remains an issue for further research. In the following chapter, we will consider the re-autonomization process in its broader context, by (briefly) focussing on the presence of this phenomenon in a few other West Ger- manic languages/varieties.

CHAPTER 6

Re-autonomization beyond (standard) Dutch46

As known from the literature, the modal verbs in (standard) German, English and the Scan- dinavian languages cannot be used independently in the same way the modal verbs in Dutch can, in particular not with a first argument referring to a state of affairs (see e.g. Mortelmans et al. 2009). This does not mean, however, that this kind of new autonomous use is not also attested in other Germanic languages. It has been observed in a number of lesser-studied West Germanic varieties, both in the Low Countries and elsewhere. In this chapter we present an overview of the situation in the latter. For convenience we confine the survey to the cognates of the three most ‘central’ Dutch modals, i.e. kunnen, mogen and moeten (a study of the cognates of hoeven is subject matter for a later investigation). First, we provide an overview of the grammars, dictionaries, and other sources we have consulted, and discuss a few problems with the reliability and coverage of this material (6.1). Then we present our findings (6.2). Finally, we offer a conclusion as to the areal distribution of the re-autonomization phenomenon (6.3).

6.1 Material and methods

Point of departure for this cross-linguistic study was the observation that the Dutch construction has parallels in contemporary (Westerlauwers) Frisian and is also recorded in the traditional dialect of Dortmund (Schleef 1967). Hence we decided to investigate whether the new autonomous use with a first argument referring to a state of affairs is also attested in other Low German dialects and in the traditional dialects of Flanders and The Netherlands. There are a number of obstacles to such an endeavour, however. First and foremost, the languages and varieties under investigation are – or were – primarily spoken vernaculars, standard Dutch or

46 This chapter is largely adopted from sections 3, 4 and 5 in Caers & Gregersen (2019: 399–417). 152 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(High) German being the traditional written languages. Although Frisian, Low German, and to a lesser extent dialects of Dutch have been and are used as written languages, the available resources for doing grammatical research are much more modest than for standard Dutch and German. Because of the lack of comprehensive and comparable corpora for the investigated languages and dialects, a large-scale quantitative study is inconceivable. We have had to restrict ourselves to consulting whichever resources were available, such as smaller corpora, dictionaries, and grammatical descriptions. This task, however, has been made much easier in recent years because of the increased availability of digitised resources online. A substantial number of dialect grammars and dictionaries from the last two centuries are freely available through online collections. For the Dutch dialects of Flanders and The Netherlands, we primarily consulted sources in the Digitale Bibliotheek voor de Nederlandse Letteren (DBNL) and in the online collections of the Meertens Instituut, chiefly the Elektronische Woordenbank van de Nederlandse Dialecten (eWND). For Frisian, we searched the text corpus Yntegrearre Taaldatabank Frysk (YTF) and the spoken-language Korpus Sprutsen Frysk (KSF), both compiled and published by the Fryske Akademy. In lieu of a corpus of older spoken Frisian, we searched the Frisian portion of the Nederlandse Volksverhalenbank (NVb), which contains a large number of traditional folk tales and anecdotes recorded in the mid-twentieth century. The Low German material was consulted primarily through the Internet Archive (IA) and the Münchener DigitalisierungsZentrum (MDZ). A substantial number of works on Low German were only available in print; in locating these various sources, we have made use especially of the bibliography published by the Verein für niederdeutsche Sprachforschung (VndS 2018) and the online bibliography compiled by Bordasch (2018). For Afrikaans, finally, we had to limit ourselves to a number of relatively recent printed grammars and dictionaries, such as the Genootskap van Regte Afrikaners (1882 [1876]), the Handwoordeboek van die Afrikaanse Taal (HAT, 2015) and the Woordeboek van die Afrikaanse Taal (WAT, 1996). Since the re-autonomization process of the three most ‘central’ modals in Dutch started only after 1850, we have primarily consulted works (from the available sources mentioned above) dating from the second half of the 19th Century and afterwards. RE-AUTONOMIZATION BEYOND (STANDARD) DUTCH 153

Before presenting our findings in section 6.2, a few caveats have to be mentioned, though. Firstly, it should be kept in mind that the available dictionaries and grammars vary considerably in their scope and coverage. Some, such as the Niedersächsisches Wörterbuch (NW) or the dictionaries of the Groningen dialect by Molema (1887) and ter Laan (1929), give copious examples of various patterns and expressions, whereas others, such as most 19th-Century grammar sketches, restrict themselves to the phonology (Lautlehre) and inflectional morphology (Formenlehre) of the dialect in question. Many of the earliest dialect dictionaries and word collections (‘idioticons’) are also quite limited in scope and contain mainly individual words that the author considered characteristic of the dialect in question or interesting for ethnographic or folkloristic reasons. The value of such sources for the study of grammatical variation, un- fortunately, tends to be negligible. Secondly, the sources generally provide little or no metadata which is now considered standard among descriptive linguists, such as information about data collection and elicitation, informants, and patterns of variation within the linguistic community. Many of the 19th-Century sources – as well as some more recent dialect dictionaries – were not written by professional linguists or philologists, but by enthusiasts who were often themselves native speakers of the dialect in question and relied on their own intuition and knowledge of it. If any information is given about the distribution or frequency of a pattern, it is generally too impressionistic to be of much use (‘this expression is often heard’, etc.). Accordingly, we are not in a position to make any claims about the usage frequency or pragmatics of the construction in the individual dialects. Our goal is simply to investigate whether the pattern is attested in the available sources for the various dialect areas. Finally, it needs to be stressed that the sources do not generally provide any negative evidence, and the truism about the absence of evidence should thus be kept in mind – if a pattern is not mentioned in a given source, this does not necessarily mean that it did or does not exist in the dialect in question. The author may have overlooked or ignored it, or it may not have occurred in the material used. In addition, despite our best efforts we may of course ourselves have overlooked the attestation of a form in our material. With these caveats in mind, we will now move on to the attestations of the new autonomous use with a first argument referring to a state of affairs in the various languages and dialects that we have investigated. 154 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

6.2 Findings

We begin our survey with the dialects of Dutch (6.2.1) before moving on to Frisian (6.2.2), Low German (6.2.3), and finally Afrikaans (6.2.4). For each example we indicate the location and the broader dialect area. Note that, as a matter of convenience, we also treat Low Saxon and Limburgish in section 6.2.1, although these Dutch dialects are in some respects linguistically closer to the Low German and Central Franconian dialects across the border.

6.2.1 Dutch dialects

We have found the construction recorded in dialect grammars and dictionaries from most of the Dutch-speaking area.47 Kunnen and mogen are both attested with a dynamic (‘is possible’) and a directive (‘is permitted’) meaning. (108a–c) give examples with kunnen, in (108b) with the older 3SG past-tense form kos (< Middle Dutch const(e), cf. Middelnederlandsch Woorden- boek, s.v. connen), in (108c) as part of a periphrastic perfect with the participial form ekennen. The examples are from Ter Laan (1929, s.v. kinn), Zegers (1999, s.v. kos), and Boekenoogen (1897, s.v. kunnen), respectively.

(108) a. Groningen (Low Saxon) Hou kin ‘t! ‘How is it possible!’

b. Land van Ravenstein (Brabantian) ’t kos mar nèt ‘It was just barely possible.’

c. Zaandam (Hollandic) O, dat had toch niet ‘ekennen ‘Oh, that was not possible after all.’

47 Namely the broader dialect areas Low Saxon (Nedersaksisch), Hollandic (Hollands), Brabantian (Brabants), and Limburgish (Limburgs). We have not found any examples in sources from the Zeelandic (Zeeuws) or Flemish (Vlaams) dialects. However, the material we have been able to access from these areas is quite limited, consisting mainly of a small number of dictionaries. We leave it open for future work whether the construction may be found in other materials from Zeeland and Oost- and West-Vlaanderen. RE-AUTONOMIZATION BEYOND (STANDARD) DUTCH 155

Two examples with mogen ‘is permitted’ are given in (109a–b) (from Van Schothorst 1904, s.v. magǝn, respectively Hoekschewaards Woordenboek 2006, s.v. magge):

(109) a. Barneveld (Low Saxon) da moes nie magen [damȯsnīmagǝn] ‘That should not be allowed’

b. Hoeksche Waard (Hollandic) ‘t Heb nooit gemagge; ‘t is al êêuwe verbôôje ‘It has never been allowed; it has been banned for centuries’

In addition, in (110) we see an example of mogen (from Pannekeet 1984, s.v. magge) with the older meaning ‘be possible’ rather than ‘be permitted’. This was the most frequent meaning of mogen in (written) Dutch until well into the early modern period (cf. Nuyts & Byloo 2015).

(110) West Friesland (Hollandic) Zuks mag graag [‘zoiets kan gemakkelijk (gebeuren)’] [lit:] ‘Such a thing may easily [happen]’

Uses of independent moeten are primarily recorded in the idiomatic ‘necessity’ expression wat moet, dat moet or the variant als het moet, dan moet het. This expression is attested not only in the Dutch dialects, but also in Frisian and Low German (cf. below). (111a) is from the north (from Molema 1887, s.v. mouten), (111b) from the south of The Netherlands (from De Bont 1958, s.v. mutǝ(n)).48

(111) a. Groningen (Low Saxon) wat mout dat mout [or:] as ‘t mout den mout ‘t ‘What has to be, has to be’

48 Further examples with kunnen, mogen and moeten can be found in sources from Maastricht and Thorn in Limburg (Endepols 1955, s.v. kunnen; Tonnaer & Sniekers 2012, s.v. mótte), Drente (Molema 1889, s.v. mag), Nuenen in Noord-Brabant (De Laat 2011, s.v. magge), and Arendonk in the province of Antwerp (Claessen & Van Deuren 2016: 22, 26). 156 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

b. Kempen (Brabantian) A’s ‘t moet dan moet et ‘If it has to be, then it has to be’

6.2.2 Frisian

As in Dutch, the construction with kunnen, mogen and moeten is attested in contemporary Frisian. Numerous examples with the cognate modal verbs kinne, meie and moatte can be found in the speech corpus KSF, cf. (112)–(114):49

(112) ik woe even wat wetter helje, kin dat? ‘I would just like to go and get some water, is that possible/permitted?’

(113) skriuwe ûnder pseudonym mei ‘writing under a pseudonym is permitted’

(114) fan Fryslân sizze wy ja, it moat wol ‘from Fryslân we say, yes, it is necessary [i.e. including Frisian in the curriculum]’

Because of the high level of Dutch influence on contemporary Frisian, one might suspect this to be a recent calque of the Dutch construction. However, if it were a calque, the basis for the loan can be traced much earlier in time: the available evidence, though limited, suggests that the construction is attested from various locations in Fryslân from the mid-nineteenth century onwards (comparable to the timing and situation in Dutch). In the YTF-corpus, we found a single example with meie, from one of the versions of the parable of the Prodigal Son in Winkler’s Dialecticon (see Winkler 1874: 458):50

49 The corpus consists of about 65 hours of recorded and transcribed spoken language. We have added punctuation, but otherwise the examples are reproduced as they appear in the corpus. 50 We have not systematically investigated any East or North Frisian material, i.e. from the Frisian languages spoken in northern Germany. However, the two East Frisian texts in Winkler (1874) suggest that the pattern may have been in use here as well, as they both contain the phrase det môt un det skel ôrs (Saterland Frisian), dait mut un dait sil ôrs (Wangeroog Frisian) ‘it must and it has to be different’ (Winkler 1874: 159, 172), cf. Dutch het moet anders. This, as well as the modal systems of the Frisian languages in general, deserves to be investigated in more detail. RE-AUTONOMIZATION BEYOND (STANDARD) DUTCH 157

(115) Schiermonnikoog Frisian (dated January 1871) In hi kriige so’n huenger dot hi wuuë wol graag siin liif fol ite mooi de swiine, mar dot mocht net. ‘And he got so hungry that he would have liked to stuff himself with the pig, but that was not allowed.’

It is worth noting that Winkler does not mention dat mocht net in his commentary to the text, suggesting that he did not consider this use out of the ordinary.51 The dictionary of Schier- monnikoog Frisian, mainly based on mid-twentieth century sources, mentions the construction for all three modal verbs, for example It is in hynjersmiddel, maar it mot ‘It is a rough remedy, but it has to be’ (Visser & Dyk 2002, s.v. motte; see also s.vv. kinne, meie). From other varieties of Frisian, we have found numerous attestations in the Nederlandse Volksverhalenbank. The two examples in (116), one with meie and one with kinne, are from the same anecdote, recorded in 1971 in eastern Fryslân. (117), from a 1965 recording, gives an example with moatte.

(116) Achtkarspelen (Wood Frisian) a. Dat gong oer in frou, de man hie oan ’t neidollen west op it boerene lân, dat mocht net. ‘It [i.e. the anecdote] was about a woman, her husband had been stealing pota- toes in the fields, that was not allowed.’

b. En nou seit mynhear ek noch, it kin net. Nou, at dat kin. ‘And now Sir is saying, that is not possible [i.e. resurrecting the dead]. Well, certainly that is possible.’

51 Note that while Johan Winkler (1840–1916) grew up in Leeuwarden and was thus not a native speaker of Frisian, he did learn to speak and write the language fluently (Meertens 1958). 158 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(117) Nijega (Wood Frisian) Tsjoensters en nachtmerjes komme troch ’t kaeisgat of troch it goatsgat. Dat moat wol, oars kin dat folkje ommers nergens troch yn ‘e hûs komme ‘Witches and mares enter through the keyhole or through the drain. And that has to be this way, you see, because those beings can’t get into the house in any other way.’

With moatte we have also found examples similar to the Dutch ‘necessity’ expression wat moet, dat moet mentioned above, as illustrated by (118) (from the Woordenboek der Friesche Taal (WFT), s.v. moatte; Frysl., 1916):

(118) Hwet moat, dat moat, sei de boer, en hy forkoft syn kou en kofte in prûk. ‘What has to be, has to be, said the farmer, and he sold his cow and bought a wig.’

The WFT (s.v. kinne; W. Dykstra, 1896) also records the idiomatic expression in (119), where both kinne and moatte have a first argument referring to a state of affairs:

(119) As ‘t net kin sa ‘t moat, den moat it sa ’t kin ‘If it cannot be the way it has to be, then it will have to be the way it can be.’ [= necessity is the mother of invention]

6.2.3 Low German

Despite the more than 8,000 bibliographical entries in VndS (2018), we have found almost no literature devoted specifically to the Low German modals. The only exception is Mortelmans (2007), which deals primarily with the inventory of modal verbs and their morphological characteristics compared with Dutch and High German. However, the available dictionaries and grammars make it abundantly clear that the new autonomous use with a first argument referring to a state of affairs is also recorded in traditional Low German dialects.52 We have found attestations from a number of different locations in northwestern Germany, the easternmost

52 For lack of a standard Low German orthography, we refer to the three modal verbs using the forms in Lindow et al. (1998): könen, mögen, and möten. RE-AUTONOMIZATION BEYOND (STANDARD) DUTCH 159 attestation being from Glückstadt on the Elbe (Bernhardt 1903: 20).53 For könen ‘can’, a West- phalian example is given in (120a) (from Piirainen & Elling 1992, s.v. können), a North Low Saxon one in (120b) (from Krüger 1843: 42).54

(120) a. Westmünsterland (Westphalian) Et konn ook, dat – ‘It was also possible that…’ or ‘It also occurred that…’

b. Emden (Ostfriesland, North Low Saxon) ‘t kann neet ‘It cannot be, is not possible’ (‘es kann nicht sein, ist nicht möglich’)

For mögen ‘may, be permitted’, an example (from Piirainen & Elling 1992, s.v. möggen) is given in (121).

(121) Westmünsterland (Westphalian) Dat magg nich! ‘That is not allowed!’

In addition, mögen is also found in a number of sources with the older meaning ‘be possible’, namely when it occurs in the collocation mag lichte/sachte. In the Westphalian example in (122) (from Rosemann gen. Klöntrup 1982 [1824], s.v. mügen), mag lichte/sachte seems to function as a complement-taking predicate (see also Schleef 1967 for an example from Dort- mund, and Tiling 1767–1771 for an example from Bremen); the Niedersächsisches Wörterbuch (1965–, s.v. mögen) appears to analyse the collocation as an adverb, as suggested by the gloss ‘vielleicht’ in (123).

53 “Nach können, müssen, sollen werden häufig Infinitive wie sein, werden u. ä. weggelassen: dat kann ni anners = das kann nicht anders gemacht werden; he sę, dat muss so er sagte, es müsse so sein.” (Bernhardt 1903: 20) 54 Uniquely among the works we consulted, the source of (120b) also speculates about the origins of the construction: ‘Die sogenannten Hülfsverben haben im Ostfriesischen eine eigene Bedeutung, die sich aus der ältesten Zeit herschreiben muß, wo diese Verben noch nicht zu Abstracten geworden waren, sondern einen vollen prägnanten Sinn hatten.’ [In the language of Ostfrisland, the so-called auxiliary verbs have a special meaning, which must go back to the ancient time when these verbs had not yet become abstract, but had their own full sense] (Krüger 1843: 42). 160 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(122) Osnabrück (Westphalian) et mag lichte, et mag sachte dat […] [lit:] ‘it may easily [be] that…’

(123) Helzendorf (Grafschaft Hoya, North Low Saxon) mach lichde bzw. mach sagde “kann sein, vielleicht” ‘may be, perhaps’

Examples with möten ‘must’ are recorded as well, cf. (124)–(125) (from Krüger 1843: 42, resp. Piirainen & Elling 1992, s.v. mütten). As (125) shows, the idiomatic ‘necessity’ expression wat moet, dat moet (cf. above for Dutch and Frisian) also has parallels in Low German:

(124) Emden (Ostfriesland, North Low Saxon) ‘t moot ‘it has to be’

(125) Westmünsterland (Westphalian) Et mutt man so! Wat mutt, dat mutt […] Wat mott denn nu? ‘It just has to be that way! What has to be, has to be […] What will have to happen now?’

In fact, the expression Wat mutt, dat mutt appears to be commonly known in Germany, where it is stereotypically associated with a certain (alleged) Northern German equanimity or resigna- tion. This is reflected in the description in (126), from a humorous radio vignette from NDR 1 Niedersachsen (see Klookschieter 2008):

(126) Es gibt im Norden gewisse Sachzwänge, an denen man nicht vorbei kommt. Wenn ein derartiger Sachzwang vorliegt, sagt der Nord-deutsche: “Wat mutt, dat mutt.” ‘In Northern Germany, there are certain practical necessities which one cannot avoid. If one such practical necessity presents itself, the Northern German will say, “Wat mutt, dat mutt”.’

RE-AUTONOMIZATION BEYOND (STANDARD) DUTCH 161

A search on the German version of Google.com (26 Jan 2019) returns more than 15,000 hits on the search string ‘wat mutt dat mutt’, including a Low German-themed board game and references to the phrase in various regional and national media (see e.g. Die Welt 2017). We have not encountered the new autonomous use with a first argument referring to a state of affairs in any East Low German (i.e. Mecklenburgish, Pomeranian, Markish) sources. Notably, however, in an East Pomeranian source we found a variation of wat mutt, dat mutt almost identical to the Frisian expression cited above, but in this version, given in (127) (from Haken 1994 [1806]: 179), the modal verb occurs with an overt infinitive, wesen ‘be’:

(127) Hinterpommern (East Pomeranian) Wat wesen mutt, datt mutt wesen! säd’ jenn’ gaud’ Buur; verköfft sin Ossen, unn köfft sick ‘n Prüück ‘“What has to be, has to be!” said the good farmer, and sold his ox and bought himself a wig.’

Whether this reflects an actual grammatical difference and the construction was indeed restricted to the western Low German dialects will have to be investigated further.

6.2.4 Afrikaans

Finally, we also find the construction in a West Germanic language spoken outside of Europe, namely Afrikaans. This language, which until the beginning of the 20th Century was regarded as a non-standard variety of Dutch, descends from the Dutch varieties spoken by settlers in South Africa from the 17th Century onwards. In contemporary Afrikaans, all three ‘central’ modals are used independently with a first argument referring to a state of affairs. (128)–(130) provide an example of each (respectively from the Handwoordeboek van die Afrikaanse Taal 2015, s.v. kan1; the Woordeboek van die Afrikaanse Taal 1996, s.v. 3mag, Volksbl., 1963; and the Woorde- boek van die Afrikaanse Taal 1996, s.v. 1moet, I.1.c):

(128) Wat nie kan nie, kan nie. ‘What is not possible, is not possible.’

162 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(129) Ons hoef ons nie te verbeel dat die laatste woord gespreek is oor wat op ’n Sondag mag en wat nie mag nie. ‘We need not imagine that the last word has been said about what is and is not allowed on a Sunday.’

(130) Ons sal help as dit moet. ‘We will help if it is necessary.’

The earliest example we have found from Afrikaans is from the second half of the 19th Century, cf. (131) (from the Genootskap van Regte Afrikaners 1882 [1876]):

(131) Di eerste boeki wat “Di Genootskap van Regte Afrikaanders” uitge is Di Eerste Beginsels van di Afrikaanse Taal. En dit kan oek ni anders ni, want di Genootskap is mos opgerig “om te staan ver ons Taal, ons Nasi en ons Land.” ‘The first booklet published by “The Society of True Afrikaaners” is The basic principles of the Afrikaans language. And it cannot be in any other way, because the Society was indeed founded “to stand for our language, our nation, and our country.”’

The history of Afrikaans – in particular how it developed out of Dutch vernacular varieties – has been the subject of much debate and several scholarly works (cf. Carstens & Raidt 2017, in particular chapter 11, for a recent overview and bibliography). However, we are not aware that the new autonomous use and its origins have received any attention so far.

6.3 Concluding remarks

As mentioned at the beginning of this chapter, our goal with this preliminary cross-linguistic study has been to document the areal distribution of the new autonomous use with a first argument referring to a state of affairs in the languages and dialects of continental West Germanic. As section 6.2 has shown, this type of use is not just a peculiarity of Dutch, but is in fact also found in Frisian, Afrikaans and several dialects of Low German. Figure 6 illustrates the distribution of the phenomenon on the German-Dutch continuum; the crosses represent the RE-AUTONOMIZATION BEYOND (STANDARD) DUTCH 163 locations (mentioned in section 6.2) in which the construction has been attested. Our interpretation of these facts is that the construction is an areal feature of Dutch, Frisian, and Low German at least west of the Weser. In addition, since it occurs in contemporary Afrikaans, it probably came over from the Dutch-speaking area at the moment the construction developed there, i.e. from the second half of the 19th Century. After all, it is until the first half of the 20th Century that South Africa remained an appealing destination for Dutch emigrants. While there would be nothing surprising about the existence of this kind of syntactic isogloss, the areal distribution of the new autonomous use with a first argument referring to a state of affairs appears to have been overlooked in the literature on West Germanic. To be sure, the pattern has clearly been noticed by lexicographers of Low German, but perhaps because of the relatively low visibility of this language, linguists outside of Low German studies do not appear to have been aware of it. It is worth stressing that we are not ruling out that it may be found outside of the area described above, as we have not (yet) systematically examined sources from High German dialects, eastern Low German, or the Frisian languages of Germany. We intend to investigate this question in further work.

Figure 6. Distribution of 'independent' modals on the German-Dutch continuum

CHAPTER 7

Conclusions and perspectives

The aim of this thesis has been to investigate the re-autonomization process of the modal verbs in Dutch in more detail. Specifically, we have tried to answer the following questions: When did the process start, and how did it develop? What are the grammatical and semantic properties of the autonomous uses? And what factors play a role in their emergence and development? The investigation has shown – in line with the earlier studies by Nuyts (2013) and Nuyts et al. (2018) – that kunnen, mogen, moeten and hoeven are re-autonomizing. They all show a strong increase in new autonomous uses, both in written and in spoken PDD, even if the intensity of the process differs between them. Yet, our present data indicate that the process set in earlier in hoeven (viz. already in ND, i.e. between 1750 and 1850) than in kunnen, mogen and moeten (viz. only after ND). Hence hoeven would seem to be leading the re-autonomization process, with the other modals following its example by analogy (but see Chapter 4 for a few provisos regarding this scenario). Moreover, the study has revealed that, from a grammatical point of view, there are two alternative pathways leading up to the ‘new main V’ uses, as the end stage of the re- autonomization process. On the one hand, there are instances with a first argument referring to a state of affairs (called the ‘new main V subtype 1’ uses). On the other hand, there are instances in which the first argument refers to an (abstract) entity (termed the ‘new main V subtype 2’ uses). These appear to behave differently in the process: the ‘subtype 2’ uses have emerged earlier in time (there are instances in the data from END onwards), but are marginal today, while the ‘subtype 1’ uses are the main responsible for the upsurge of the new autonomous uses in PDD (or, in hoeven, in ND).

166 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

We have argued that the re-autonomization process is a case of (collective) degrammaticalization. The available evidence indicates that the ‘new main V’ uses are not a continuation of the original main verbs, but have emerged from purely auxiliary uses with the same grammatical pattern (cf. type 1 vs type 2 uses), through a stage with an implied main verb. This analysis is further supported by the fact that the ‘new main V’ uses feature the same modal and related meanings as the auxiliary uses (see Byloo & Nuyts 2014, Nuyts & Byloo 2015, and Nuyts et al. 2018). They even focus, more than the purely auxiliary uses (in which the ‘dynamic- inherent’ and ‘dynamic-imposed’ modal meanings predominate), on the more subjective (‘dynamic-situational’ and ‘deontic’) and intersubjective (‘directive’, and in hoeven also ‘volitive’) meanings and uses. In a wider perspective our findings have implications for two specific linguistic concep- tions. On the one hand, they go against the assumption (cf. e.g. Heine et al. 1991) that the process of grammaticalization is fully unidirectional: not only the shift from a less to a more grammatical status of an item (i.e. from main verb to auxiliary verb), but also the opposite evolution (from auxiliary to main verb) is possible. On the other hand, our results imply that degrammaticalization does not have to involve a return to a ‘lexical’ meaning (pace Norde’s concept of ‘degrammation’). The ‘degrammaticalization’ of the modal verbs in Dutch does not involve de-(inter)subjectification (cf. the fact, pointed out by Traugott 2010, that grammaticalization and (inter)subjectification do not have to coincide, even if they often do). Another aim of this study was to find possible explanations for the re-autonomization process. Since autonomous uses in the presence of a directional in the clause are very common, not only in Dutch, but also in many other Germanic languages, Olbertz & Honselaar (2017), in a study into the diachrony of moeten in particular, have suggested that these might be the cause for the re-autonomization process in Dutch. Our investigation has shown that this assumption cannot be maintained, however, not even for moeten, even if it combines with a directional more often than the other modals. Although directionals do play a role, specifically in the ‘subtype 2’ instances of the four modal verbs, they cannot explain the developments along the pathway towards the ‘new main V subtype 1’ uses in Dutch. Another element we have investigated as a possible trigger for the re-autonomization process is the presence of a negative polarity context, since hoeven – as a negative polarity CONCLUSIONS AND PERSPECTIVES 167 item – appears to be leading the re-autonomization process (cf. above). In spite of major differences between the verbs in the share of negative polarity instances, in all four they are more frequent in the new autonomous uses, and especially in the ‘new main V’ uses (of subtypes 1 and 2), than in the purely auxiliary uses at the time of the upsurge of the new autonomous uses (i.e. in PDD for kunnen, mogen and moeten, and in ND for hoeven). Hence negative polarity does seem to play a role in the re-autonomization process. Why this is the case, and how it may have triggered or affected the process, remains an open issue, however. Apart from the latter, there are a few other questions the present study has left unanswered, which may thus be subject for further research. Firstly, one may wonder why directionals are a trigger for autonomous uses of the type involving a (pro)nominal first argument, not only in Dutch, but in the Germanic languages in general. And why are there such considerable differences in this matter between the different modals, and between the Germanic languages (cf. for instance the fact that in the Scandinavian languages autonomous uses with directionals are more or less confined to ‘strong’ modals, while in Dutch they also occur in ‘weak’ modals, such as kunnen and mogen). Yet another question left unanswered is why the developments in the uses with a first argument referring to a state of affairs did happen in Dutch, but not in other Germanic languages, such as Standard English, German and the Scandinavian languages. A preliminary cross- linguistic study has shown that, at least for the cognates of kunnen, mogen and moeten, the construction is not just a peculiarity of Dutch, but is also attested in a few other West Germanic varieties, such as Frisian, Low German and Afrikaans. It would be interesting to investigate how this pattern developed in these languages, not only for kunnen, mogen and moeten, but also for hoeven (which was not included in our preliminary study). Moreover, a more systematic investigation might also reveal whether the phenomenon is present in yet other West European varieties. A wider cross-linguistic study may further our understanding of the conditions in which a re-autonomization process may occur. Finally, it might be useful to try to determine the timing of the re-autonomization process even more precisely. Critical is the time frame from 1850 to 1950 (which was not examined in this study). A detailed analysis of the developments in that period (not only in Dutch, but also

168 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH in other languages/varieties) may possibly offer new insights into the factors that are responsible for the re-autonomization of the modal auxiliaries. Hopefully, this study will provide a starting point for future investigations, and may inspire other researchers to contribute to an even more complete picture of the re-autonomization process.

References

Aijmer, K. (1997). I think – an English modal particle. In T. Swan & O. Jansen Westwik (eds.), Modality in Germanic languages: Historical and comparative perspectives, 1–47. Berlin: Mouton de Gruyter. AWb = E.v. Steinmeyer, E. Karg-Gasterstädt & T. Frings (1952–). Althochdeutsches Wörterbuch. Leipzig. Andersen, H. (2008). Grammaticalization in a speaker-oriented theory of change. In Thórhallur Eythórsson (ed.), Grammatical change and linguistic theory: The Rosendal papers, 11– 44. Amsterdam: John Benjamins. Andersson, P. (2008). Swedish må and the (de)grammaticalization debate. In E. Seoane & M.J. López-Couso (eds.), Theoretical and empirical issues in grammaticalization, 15–32. Amsterdam: John Benjamins. Auwera, J. van der (2002). More thoughts on degrammaticalization. In I. Wischer & G. Diewald (eds.), New reflections on grammaticalization, 19–29. Amsterdam: John Benjamins. Auwera, J. van der & V.A. Plungian (1998). Modality’s semantic map. Linguistic Typology 2. 79– 124. Bech, G. (1951). Grundzüge der semantischen Entwicklungsgeschichte der hochdeutschen Modalverba (Det Kongelige Danske Videnskabernes Selskab: Historisk-filologiske Med- delelser 32, 6). Copenhagen: Munksgaard. Benveniste, É. (1971 [1958]). Subjectivity in language. In Problems in general linguistics, 223– 230. Coral Gables: University of Miami Press. Bernhardt, J. (1903). Zur Syntax der gesprochenen Sprache (Ein Versuch). Jahrbuch des Vereins für Niederdeutsche Sprachforschung 29. 1–25. Beths, F. (1999). The history of dare and the status of unidirectionality. Linguistics 37(6). 1069– 1110. Biber, D., S. Johansson, G. Leech, S. Conrad & E. Finegan (1999). Longman grammar of spoken and written English. Harlow: Longman. 170 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Bisang, W. (2008). Grammaticalization and the areal factor: The perspective of East and mainland Southeast Asian languages. In M.J. López-Couso & E. Seoane (eds.), Rethinking grammaticalization: New perspectives, 15–35. Amsterdam: John Benjamins. Bjerrum, A. (1966). Grammatik over Skånske Lov efter B74. Copenhagen: Gyldendal. Boekenoogen, G.J. (1897). De Zaansche volkstaal. Leiden: Sijthoff. Bont, A.P. de (1958). Dialekt van Kempenland, vol 2. Assen: Van Gorcum. Bordasch, O. (2018). Bibliographie digitaler niederdeutscher Literatur (15.-18. Jh.). Börjars, K. & N. Vincent (2011). Grammaticalization and directionality: data, analysis and explanation. In H. Narrog & B. Heine (eds.), The Oxford handbook of grammaticalization, 163–176. Oxford: Oxford University Press. Boye, K. & P. Harder (2012). A usage-based theory of grammatical status and grammaticalization. Language 88(1). 1–44. Brandt, S. (1999). Modal verbs in Danish (Travaux du Cercle linguistique de Copenhague 30). Copenhagen: Reitzel. Burridge, K. (1998). From modal auxiliary to lexical verb: The curious case of Pennsylvania Ger- man wotte. In R.M. Hogg & L. Bergen (eds.), Historical Linguistics 1995. Selected Papers from the 12th ICHL. Vol. 2: Germanic Linguistics, 19–31. Amsterdam: John Benjamins. Bybee, J. & W. Pagliuca (1987). The evolution of future meaning. In A.G. Ramat, O. Carruba & G. Bernini (red.), Papers from the 7th international conference on historical linguistics, 109–122. Amsterdam: Benjamins. Bybee, J., R. Perkins & W. Pagliuca (1991). Back to the future. In E.C. Traugott & B. Heine (red.), Approaches to grammaticalization, vol. 2, 17–58. Amsterdam: Benjamins. Bybee, J., R. Perkins & W. Pagliuca (1994). The evolution of grammar: Tense, aspect, and modality in the languages of the world. Chicago: University of Chicago Press. Byloo, P. & J. Nuyts (2011). The diachrony of Dutch mogen. Antwerp Papers in Linguistics 113. Byloo, P. & J. Nuyts (2014). Meaning change in the Dutch core modals: (Inter)subjectification in a grammatical paradigm. Acta Linguistica Hafniensia 46(1). 85–116. Caers, W. & S. Gregersen (2019). Wat mutt, dat mutt: ‘Independent’ modals in West Germanic vernaculars. Nederlandse Taalkunde 24(3). 399–417. REFERENCES 171

Caers, W. & J. Nuyts (submitted). Directionals and re-autonomization in the Dutch modals. Campbell, L. (2001a). What’s wrong with grammaticalization? Language Sciences 23(2–3). 113– 161. Campbell, L. (ed.) (2001b). Grammaticalization: A critical assessment. [Special issue]. Language Sciences 23(2–3). Carstens, W.A.M. & E.H. Raidt (2017). Die storie van Afrikaans, vol. 1. Pretoria: Protea. Claessen, R. & G. Van Deuren (2016). ’t Aorendoenks bukske: Arendonks dialectwoordenboek. Arendonk: Heemkundige Kring ‘Als Ice Can’. Coates, J. (1983). The semantics of modal auxiliaries. London: Croom Helm Coussé, E. (2010). Een digitaal compilatiecorpus historisch Nederlands [A digital compilation corpus historical Dutch]. Lexikos 20. 123–142. DBNL = Digitale Bibliotheek voor de Nederlandse Letteren (1999–). Den Haag. Diewald, G. (1997). Grammatikalisierung: Eine Einführung in Sein und Werden grammatischer Formen (Germanistische Arbeitshefte 36). Tübingen: Niemeyer. Diewald, G. (1999). Die Modalverben im Deutschen: Grammatikalisierung und Polyfunktionali- tät. Tübingen: Niemeyer. Diewald, G. (2010). On some problem areas in grammaticalization studies. In K. Stathi, E. Geh- weiler & E. König (eds.), Grammaticalization: current views and issues, 17–50. Amster- dam: John Benjamins. Die Welt (2017). Wat mutt, dat mutt. Die Welt 05/11/2017. Dixon, R.M.W. (1994). Ergativity. Cambridge: Cambridge University Press. Duinhoven, A.M. (1997). Middelnederlandse syntaxis synchroon en diachroon, vol. 2. Groningen: Nijhoff. Eide, K. (2005). Norwegian modals. Berlin: Mouton de Gruyter. Endepols, H.J.E. (1955). Woordenboek of diksjenaer van ‘t Mestreechs. Maastricht: Boosten & Stols. eWND = Nicoline van der Sijs (samensteller, Meertens Instituut) (2015–). Elektronische Woor- denbank van de Nederlandse Dialecten. Amsterdam.

172 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Faarlund, J.T., S. Lie & K.I. Vannebo (1997). Norsk referansegrammatikk. Oslo: Universitets- forlaget. Fischer, O. (2007). Morphosyntactic change: Functional and formal perspectives. Oxford: Ox- ford University Press. Fischer, O. (2010). Some problem areas in grammaticalization: Lehmann’s parameters and the issue of scope. In A. Van linden, J.-C. Verstraete & K. Davidse (eds.), Formal evidence in grammaticalization research, 17–42. Amsterdam: John Benjamins. Fischer, O., A. Rosenbach & D. Stein (eds.) (2000). Pathways of change: Grammaticalization in English. Amsterdam: John Benjamins. Fritz, G. (1997a). Deutsche Modalverben 1609: Nicht-epistemische Verwendungsweisen. In G. Fritz & T. Gloning (eds.), Untersuchungen zur semantischen Entwicklungsgeschichte der Modalverben im Deutschen (Reihe Germanistische Linguistik 187), 249–305. Tübingen: Niemeyer. Fritz, G. (1997b). Historische Semantik der Modalverben: Problemskizze – exemplarische Analysen – Forschungsüberblick. In G. Fritz & T. Gloning (eds.), Untersuchungen zur semantischen Entwicklungsgeschichte der Modalverben im Deutschen (Reihe Germanis- tische Linguistik 187), 1–158. Tübingen: Niemeyer. Fritz, G. & T. Gloning (eds.) (1997). Untersuchungen zur semantischen Entwicklungsgeschichte der Modalverben im Deutschen (Reihe Germanistische Linguistik 187). Tübingen: Nie- meyer. Furmaniak, G. (2011). On the emergence of epistemic use of must. SKY Journal of Linguistics 24. 41–73. Gabelentz, G. von der (1891). Die Sprachwissenschaft, ihre Aufgaben, Methoden und bisherigen Ergebnisse. Leipzig: T. O. Weigel Nachfolger (Tauchnitz). Gamon, D. (1993). On the development of epistemicity in the German modal verbs mögen and müssen. Folia Linguistica Historica 14(1–2). 125–176. Genootskap van Regte Afrikaners (1882 [1876]). Eerste beginsels van di Afrikaanse taal, 2nd edn. Paarl: Du Toit & Co. Givón, T. (1971). Historical syntax and synchronic morphology: an archeologist’s field trip. Chicago Linguistic Society (CLS) 7. 394–415. REFERENCES 173

Goossens, L. (1982). On the development of the modals and of the epistemic function in English. In A. Ahlqvist (ed.), Papers from the Fifth International Conference on Historical Linguis- tics, Galway, April 6–10 1981, 74–84. Amsterdam: John Benjamins. Goossens, L. (1987). Modal tracks: The case of magan and motan. In A.-M. Simon- Vandenbergen (ed.), Studies in honour of René Derolez, 216–236. Gent: Vitgeuer. Goossens, L. (1992). Cunnan, conne(n), can: The development of a radial category. In G. Kellermann & M.D. Morrissey (eds.), Diachrony within synchrony: Language history and cognition, 377–394. Frankfurt: Peter Lang. Goossens, L. (1999). Metonymic bridges in modal shifts. In K.-U. Panther & G. Radden (eds.), Metonymy in language and thought, 193–210. Amsterdam: John Benjamins. Grondelaers, S., K. Deygers, H. Van Aken, V. Van den Heede & D. Speelman (2000). Het ConDiv- corpus geschreven Nederlands. Nederlandse Taalkunde 5. 356–363. Haeseryn, W., K. Romijn, G. Geerts, J. de Rooij & M.C. van den Toorn (1997). Algemene Nederlandse spraakkunst. 2de herziene uitgave. Groningen: Nijhoff. Haken, J.C.L. (1994 [1806]). Hinweisung auf einige Idiotismen und Sprüchwörter der plattdeutschen Mundart in Hinter-Pommern. In W. Haas (ed.), Provinzialwörter: Deutsche Idiotismen-sammlungen des 18. Jahrhunderts (= Historische Wortforschung 3), 172– 180. Berlin: De Gruyter. Hansen, E. (1972). Modal interessens. Danske Studier 1972. 5–36. Hansen, E. & L. Heltoft (2011). Grammatik over det danske sprog. Copenhagen: Det Danske Sprog- og Litteraturselskab. Hansen, E. & L. Heltoft (2019 [2011]). Grammatik over det Danske Sprog, 2nd edn. Copenhagen: Det Danske Sprog- og Litteraturselskab. Hanson, K. (1987). On subjectivity and the history of epistemic expressions in English. Chicago Linguistic Society 23. 133–147. Harris, A.C. & L. Campbell (1995). Historical syntax in cross-linguistic perspective. Cambridge: Cambridge University Press. (Cambridge studies in linguistics 74) Haspelmath, M. (2004). On directionality in language change with particular reference to grammaticalization. In O. Fischer, M. Norde & H. Perridon (eds.), Up and down the cline: The nature of grammaticalization, 17–44. Amsterdam: John Benjamins. 174 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

HAT = Handwoordeboek van die Afrikaanse Taal (2015). 6th edn. Cape Town: Pearson. Heine, B. & M. Reh (1984). Grammaticalization and reanalysis in African languages. Hamburg: Buske. Heine, B., U. Claudi & F. Hünnemeyer (1991). Grammaticalization: A conceptual framework. Chicago: The University of Chicago Press. Heltoft, L. & M.A. Nielsen (2019). Modalverbernes udvikling. Online appendix to Ebba Hjorth et al. (eds.), Dansk Sproghistorie 3: Bøjning og bygning. Aarhus: Aarhus Universitetsforlag. Herreweghe, M. van (2000). *Motan in the Anglo-Saxon poetic records. Belgian Journal of Linguistics 14. 207–239. Honselaar, W. & H. Olbertz (2016). The use of Dutch moeten without an infinitive: A case of degrammaticalization? In A. Bannink & W. Honselaar (eds.), From variation to iconicity: Festschrift for Olga Fischer, 185–201. Amsterdam: Pegasus. Hopper, P.J. (1991). On some principles of grammaticization. In E.C. Traugott & B. Heine (eds.), Approaches to grammaticalization, vol. I, 17–36. Amsterdam: John Benjamins. Hopper, P.J. & E.C. Traugott (2003 [1993]). Grammaticalization, 2nd edn. Cambridge: Cam- bridge University Press. Horst, J. van der (1985). Verkenning van onpersoonlijke constructies. Tijdschrift voor Neder- landse Taal- en Letterkunde 101. 34–63. Humboldt, W. v. (1825). Über das Entstehen der grammatischen Formen und ihren Einfluß auf die Ideenentwicklung. Abhandlungen der königlichen Academie der Wissenschaften zu Berlin 1822–23. 401–430. HW = Hoekschewaards Woordenboek (2006). Klaaswal: Deboektant. IA = Internet Archive (1996). San Francisco: Presidio. Jacobsson, B. (1994). Recessive and emergent uses of modal auxiliaries in English. English Stu- dies 75(2). 166–182. Jakobson, R. (1971 [1959]). Boas’ view of grammatical meaning. In Selected writings, vol. 2, 489–496. The Hague: Mouton. [First published in W. Goldschmidt (ed.), The anthropo- logy of Franz Boas: Essays on the centennial of his birth. San Francisco: Howard Chand- ler]. REFERENCES 175

Janda, R.D. (2001). Beyond “pathways” and “unidirectionality”: on the discontinuity of language transmission and the counterability of grammaticalization. Language Sciences 23(2–3). 265–340. Jespersen, O. (1931). A Modern English grammar on historical principles, vol. 4. Heidelberg: Winther. Joseph, B.D. (2001). Is there such a thing as ‘grammaticalization?’. Language Sciences 23(2–3). 163–186. Joseph, B.D. (2004). Rescuing traditional (historical) linguistics from grammaticalization “theory”. In O. Fischer, M. Norde, & H. Perridon (eds.), Up and down the cline – the nature of grammaticalization, 44–71. Amsterdam: John Benjamins. Joseph, B.D. (2011). Grammaticalization: A general critique. In H. Narrog & B. Heine (eds.), The Oxford handbook of grammaticalization, 193–205. Oxford: Oxford University Press. Klarén, G.A. (1913). Die Bedeutungsentwicklung von können, mögen und müssen im Hoch- deutschen. PhD dissertation, Lund University. Klookschieter (2008). Wat mutt, dat mutt. Published 07 Dec 2008. Krug, M.G. (2000). Emerging English modals: A corpus-based study of grammaticalization. Berlin: Mouton de Gruyter. Krüger, E. (1843). Uebersicht der heutigen plattdeutschen Sprache (besonders in Emden). Emden: Woortman. KSF = Fryske Akademy (2001). Korpus Sprutsen Frysk. Leeuwarden. Kuryłowicz, J. (1965). The evolution of grammatical categories. Diogenes 51. 55–71. Kytö, M. (1987). On the use of the modal auxiliaries indicating ‘possibility’ in early American English. In M. Harris & P. Ramat (eds.), Historical development of auxiliaries (Trends in Linguistics. Studies and Monographs 35), 145–170. Berlin: Mouton de Gruyter. Kytö, M. (1988). Can or may? Choice of the variant form in Early Modern English, British and American. Georgetown University Round Table on Languages and Linguistics (GURT) 1988. 163–178.

176 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Kytö, M. (1991). Can (could) vs. may (might): Regional variation in Early Modern English? In D. Kastovsky (ed.), Historical English syntax (Topics in English linguistics 2), 233–289. Berlin: Mouton de Gruyter. Laan, K. ter (1929). Nieuw Groninger woordenboek. Groningen: Wolters. Laat, G. de (2011). Zoo prôte wèij in Nuejne mi mekaâr. Nuenen: Heemkundekring De Drije- hornick. Langacker, R.W. (1977). Syntactic reanalysis. In C.N. Li (ed.), Mechanisms of syntactic change, 57–139. Austin: University of Texas Press. Langacker, R.W. (1985). Observations and speculations on subjectivity. In J. Haiman (ed.), Iconicity in syntax, 109–150. Amsterdam: John Benjamins. Lehmann, C. (1982). Thoughts on grammaticalization: A programmatic sketch, vol. I (Arbeiten des Kölner Universalienprojekts 48). Köln: Institut für Sprachwissenschaft. Lehmann, C. (1995). Thoughts on grammaticalization, revised and expanded version (LINCOM Studies in Theoretical Linguistics 1). München: LINCOM. Lehmann, C. (2002). Thoughts on grammaticalization, 2nd, revised edition (ASSidUE 9). Erfurt: Seminar für Sprachwissenscha der Universität. Lehmann, C. (2004). Theory and method in grammaticalization. Zeitschrift für germanistische Linguistik 32(2). 152–187. Lehmann, C. (2015). Thoughts on grammaticalization, 3rd edition (Classics in Linguistics 1). Berlin: Language Science Press. Lightfoot, D. (1979). Principles of diachronic syntax. Cambridge: Cambridge University Press. Lindow, W., D. Möhn, H. Niebaum, D. Stellmacher, H. Taubken & J. Wirrer (1998). Nieder- deutsche Grammatik. Leer: Schuster. Lühr, R. (1987). Zu Veränderungen im System der Modalverben. In R. Bergmann, H. Tiefenbach & L. Voetz (eds.), Althochdeutsch. Teil I: Grammatik, Glossen, Texte, 262–289. Heidel- berg: Carl Winter. Lyons, J. (1977). Semantics. Cambridge: Cambridge University Press. Lyons, J. (1982). Deixis and subjectivity: Loquor, ergo sum? In R.J. Jarvella & W. Klein (eds.), Speech, place and action: Studies on deixis and related topics, 101–124. Chichester: Wiley. REFERENCES 177

MDZ = Münchener DigitalisierungsZentrum (1997–). München. MED = Middle English Dictionary (1952–). A. Arbor: University of Michigan. Meertens, P.J. (1958). Johan Winkler en zijn Dialecticon. Taal en Tongval 10. 4–15. Meillet, A. (1912). L’évolution des formes grammaticales. Scientia: rivista internazionale di sintesi scientifica 12. 384–400. Mitchell, B. (1985). Old English Syntax, vol. 1. Oxford: Oxford University Press. MNW = E. Verwijs & J. Verdam (1885–1929). Middelnederlandsch woordenboek. Den Haag: Nijhoff. Molema, H. (1887). Woordenboek der Groningsche volkstaal in de 19e eeuw. Winsum: Mekel. Molema, H. (1889). Proeve van een woordenboek der Drentsche volkstaal in de 19e eeuw. Un- published ms. Mortelmans, T. (2007). Modalverben im Niederdeutschen: Ansatz zu einem Vergleich mit dem Modalverbbestand im Deutschen und im Niederländischen. Niederdeutsches Wort 47/48. 135–148. Mortelmans, T., K. Boye & J. van der Auwera (2009). Modals in the Germanic languages. In B. Hansen & F. de Haan (eds.), Modals in the languages of Europe: a reference work, 11– 69. Berlin: Mouton de Gruyter. Nagle, S.J. (1989). Quasi-modals, marginal modals, and the diachrony of the English modal auxiliaries. Folia Linguistica Historica 9(2). 93–104. Nederlandse Taalunie (2004). Corpus Gesproken Nederlands, versie 2.0. Leiden: TST-Centrale INL. Newmeyer, F.J. (2001). Deconstructing grammaticalization. Language Sciences 23(2–3). 187– 229. Norde, M. (2009). Degrammaticalization. Oxford: Oxford University Press. Nuyts, J. (2008). Qualificational meanings, illocutionary signals, and the cognitive planning of language use. Annual Review of Cognitive Linguistics 6(1). 185–207. Nuyts, J. (2011). Degrammaticalisatie in de Nederlandse modale hulpwerkwoorden. Verslagen & Mededelingen van de Koninklijke Academie voor Nederlandse Taal- en Letterkunde 121(2). 155–182. 178 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Nuyts, J. (2013). De-auxilarization without de-modalization in the Dutch core modals: a case of collective degrammaticalization? Language Sciences 36. 124–133. Nuyts, J. (2014). Zelfstandig gebruikte modalen: Een functioneel perspectief. Nederlandse Taal- kunde 119(3). 351–373. Nuyts, J. (2016). Analyses of the modal meanings. In J. Nuyts & J. van der Auwera (eds.), The Oxford handbook of modality and mood, 31–49. Oxford: Oxford University Press. Nuyts, J. (2017). Evidentiality reconsidered. In J.I. Marin-Arrese (ed.), Evidentiality revisited: cognitive grammar, functional and discourse-pragmatic perspectives, 57–83. Amster- dam: Benjamins. Nuyts, J. & P. Byloo (2015). Competing modals: Beyond (inter)subjectivity. Diachronica 32(1). 34–68. Nuyts, J., P. Byloo & J. Diepeveen (2010). On deontic modality, directivity, and mood: The case of Dutch mogen and moeten. Journal of Pragmatics 42. 16–34. Nuyts, J., H.-J. Goelen & W. Caers (2018). ‘Hoeven’ diachroon. Taal en Tongval 70(1). 17–46. Nuyts, J., H.-J. Goelen & W. Caers (2019). Hoeveel systeem zit er in het systeem van de Neder- landse modale werkwoorden? Leuvense Bijdragen. In press. NVb = Meertens Instituut (2004–). Nederlandse Volksverhalenbank. Amsterdam. NW = Niedersächsisches Wörterbuch (1965–). Göttingen: Institut für Historische Landes- forschung. Obe, R. (2013). Modalverbernes semantiske system i gammeldansk. Roskilde: Roskilde Univer- sity PhD thesis. Öhlschläger, G. (1989). Zur Syntax und Semantik der Modalverben des Deutschen. Tübingen: Niemeyer. Olbertz, H. & W. Honselaar (2017). The grammaticalization of Dutch moeten. In K. Hengeveld, H. Narrog & H. Olbertz (eds.), The grammaticalization of tense, aspect, modality and evidentiality, 273–300. Berlin: Mouton de Gruyter. ONW = Oudnederlands woordenboek (2012). Leiden: INL. Ostaeyen, G. van & J. Nuyts (2004). De diachronie van ‘kunnen’. Antwerp Papers in Linguistics 109. REFERENCES 179

Palmer, F.R. (1990 [1979]). Modality and the English modals, 2nd edn. London: Routledge. Palmer, F.R. (2001 [1986]). Mood and modality, 2nd edn. Cambridge: Cambridge University Press. Pannekeet, J. (1984). Westfries woordenboek. Wormerveer. Paul, H. (2002 [1897]). Deutsches Wörterbuch, 9th edn. Tübingen: Niemeyer. Peilicke, R. (1997). Zur Verwendung der Modal verben können und mögen im frühneuzeitlichen Deutsch (1500 bis 1730). In G. Fritz & T. Gloning (eds.), Untersuchungen zur semantischen Entwicklungsgeschichte der Modalverben im Deutschen (Reihe Germanistische Linguistik 187), 209–247. Tübingen: Niemeyer. Piirainen, E. & W. Elling (1992). Wörterbuch der westmünsterländischen Mundart (= Beiträge des Heimatvereins Vreden zur Landes- und Volkskunde 40). Vreden: Heimatverein Vreden. Plank, F. (1984). The modals story retold. Studies in Language 8(3). 305–364. Reed, J.O. (1981). The modal verbs will/would, shall/should, can/could, may/might, need, must, ought and dare/durst in the English of the time of Shakespeare. Unpublished dissertation, University of Newcastle upon Tyne. Rosemann gen. Klöntrup, J.G. (1982 [1824]). Niederdeutsch-westphälisches Wörterbuch, vol. 1. Hildesheim: Lax. Schleef, W. (1967). Dortmunder Wörterbuch (= Niederdeutsche Studien 15). Cologne: Böhlau. Schothorst, W. van (1904). Het dialect der Noord-West-Veluwe. Utrecht: Kemink. Schlüter, J. (2010). To dare to or not to. Is grammaticalization irreversible? In A. Van linden, J.- C. Verstraete & K. Davidse (eds.), Formal evidence in grammaticalization research, 289– 325. Amsterdam: John Benjamins. Shepherd, S. (1982). From deontic to epistemic: an analysis of modals in the history of English, creoles, and language acquisition. In A. Ahlqvist (ed.), Papers from the Fifth International Conference on Historical Linguistics, Galway, April 6–10 1981, 316–323. Amsterdam: John Benjamins. Solo, H.J. (1977). The meaning of ‘*motan’: A secondary denotation of necessity in Old English? Neuphilologische Mitteilungen 78. 215–232. 180 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Sommerer, L. (2015). The influence of constructions in grammaticalization: Revisiting category emergence and the development of the definite article in English. In J. Barðdal, E. Smirnova, L. Sommerer & S. Gildea (eds.), Diachronic Construction Grammar (Construc- tional Approaches to Language 18), 107–138. Amsterdam: John Benjamins. Standop, E. (1957). Syntax und Semantik der modalen Hilfsverben im Altenglischen magan, motan, sculan, willan. Bochum-Langendreer: Pöppinghaus. Svenska Akademiens grammatik (1999). Stockholm: Svenska Akademien. Sweetser, E. (1990). From etymology to pragmatics: Metaphorical and cultural aspects of semantic structure. Cambridge: Cambridge University Press. Tabor, W. & E.C. Traugott (1998). Structural scope expansion and grammaticalization. In A.G. Ramat & P.J. Hopper (eds.), The limits of grammaticalization, 229–272. Amsterdam: John Benjamins. Taeymans, M. (2004). An investigation into the marginal modals DARE and NEED in British present-day English: A corpus-based approach. In O. Fischer, M. Norde & H. Perridon (eds.), Up and down the cline: The nature of grammaticalization, 97–114. Amsterdam: John Benjamins. Tiling, E. (1767–1771). Versuch eines bremisch-niedersächsischen Wörterbuchs, vols. 1–5. Bremen: Förster. Tonnaer, M. & H. Sniekers (eds.) (2012). Thoears Woeardebook. Thorn: Werkgroep Thoears Woeardebook. Traugott, E.C. (1986). From polysemy to internal semantic reconstruction. Berkeley Linguistics Society 12. 539–550. Traugott, E.C. (1989). On the rise of epistemic meaning in English: An example of subjectification in semantic change. Language 65. 31–55. Traugott, E.C. (2010). (Inter)subjectivity and (inter)subjectification: a reassessment. In K. Davidse, L. Vandelanotte & H. Cuyckens (eds.), Subjectification, Intersubjectification and Grammaticalization, 29–71. Berlin: Mouton de Gruyter. Traugott, E.C. & B. Heine (1991). Approaches to grammaticalization. Amsterdam: John Ben- jamins.

REFERENCES 181

Traugott, E.C. & R.B. Dasher (2002). Regularity in semantic change. Cambridge: Cambridge University Press. UGent (2007–). Google books project. Ghent: Ghent University. Vandeweghe, W. (2013). Grammatica van de Nederlandse zin. 8ste herziene druk. Antwerpen/ Apeldoorn: Garant. Visser, F.T. (1963). An historical syntax of the English language, vol. 1. Leiden: Brill. Visser, W. & S. Dyk (2002). Eilander Wezzenbúek: Woordenboek van het Schiermonnikoogs (= Rige Fryske Dialektstúdzjes 10). Leeuwarden: Fryske Akademy. VMNW = Pijnenburg, W.J.J, K.H. van Dalen-Oskam, K.A.C. Depuydt & T.H. Schoonheim (red.) (2000). Vroegmiddelnederlands woordenboek. Leiden: INL. VndS (2018). Bibliographie des Vereins für niederdeutsche Sprachforschung. Vries, J. de (1971). Nederlands etymologisch woordenboek. Leiden: Brill. Warner, A.R. (1993). English auxiliaries: Structure and history. Cambridge: Cambridge Univer- sity Press. WAT = Woordeboek van die Afrikaanse Taal, vol. 10 (1996). Stellenbosch: Buro van die WAT. WFT = Woordenboek der Friesche Taal (2010). Leiden: INL. Wijk, N. van (1912). Franck’s etymologisch woordenboek der Nederlandsche taal, 2nd edn. Den Haag: Nijhoff. Winkler, J. (1874). Algemeen Nederduitsch en Friesch dialecticon, vol. 1. Den Haag: Nijhoff. WNT = Vries, M. De & J.A. Te Winkel (1864–1998). Woordenboek der Nederlandsche taal. ’s- Gravenhage/Leiden: Nijhoff/Sijthoff. Wouden, T. van der (1996). Hoeven. Tabu 26. 164–182. Yanovich, I. (2016). Old English *motan, variable-force modality, and the presupposition of inevitable actualization. Language 92(3). 489–521. YTF = Fryske Akademy (2011). Yntegrearre Taaldatabank Frysk. Leeuwarden. Zegers, A. (1999). Het dialect van het land van Ravenstein. Uden: n.p. Ziegeler, D. (2003). Redefining unidirectionality: Insights from demodalisation. Folia Linguistica Historica 24(1–2). 225–266.

Appendix 1

This appendix renders an overview of all text material we have used from, respectively:

- The Digitale Bibliotheek voor de Nederlandse Letteren: p. 184–205

* For END: p. 184–188 (for North) and p. 189–190 (for South) * For ND: p. 191–199 (for North) and p. 200–201 (for South) * For PDDW: p. 202–203 (for North) and p. 204–205 (for South)

- The Google Books project of the University of Ghent: p. 206–219

* For END: p. 206 (for North) and p. 207 (for South) * For ND: p. 208–215 (for North) and p. 216–219 (for South)

- The Google (Scholar) websites: p. 220–221

* For PDDW: p. 220 (for North) and p. 220–221 (for South)

184 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Digitale Bibliotheek voor de Nederlandse Letteren

Table 58 END sources – Northern Dutch (part 1)

Author name Text name Text type Samuel Ampzing 1. Het Hoog-lied van den heyligen ende wijsen koning ende propheet rhymed/lyrical Salomon 2. Naszousche lauren-kranze rhymed/lyrical 3. Rym-catechismus dat is de Christelijcke Catechismus rhymed/lyrical 4. Westindische triumphbazuin op de verovering van de zilveren vloot rhymed/lyrical Philips Angel Lof der schilder-konst secular (argumentation) Jan van Arp Singhende klucht van droncke Goosen rhymed/lyrical Jan J. van Asten & Isaak A. Haarlemsche mei-bloempjes. Derde offer aen de vreughd-lievende rhymed/lyrical van Vaerlen nymphjes Willem Bartjens De cijfferinghe secular (science) Willem Baudartius 1. Afbeeldinge ende beschryvinghe van alle de veldslagen, secular (history) belegeringen ende and're notabele geschiedenissen ghevallen in de Nederlanden 2. Morghen-wecker der vrye Nederlantsche Provintien statutory (politics) Isaac Beeckman Journal tenu par Isaac Beeckman de 1604 à 1634 personal (diary) Nicolaes II Biestkens Claes Kloet rhymed/lyrical Adriaan Boelens Oneenige-trouw rhymed/lyrical Willem IJ. Bontekoe Journael ofte gedenckwaerdige beschrijvinghe van de Oost-Indische personal (travel story) reijse Pieter C. Bor Den oorspronck, begin ende aenvanck der Nederlandtscher oorlogen rhymed/lyrical Lambert van den Bosch Konst kabinet van Marten Kretzer rhymed/lyrical Marijn de Brauwer Schalmeye, inhoudende veel geestelijcke liedekens rhymed/lyrical Salomon de Bray Architectura moderna ofte bouwinge van onsen tyt secular (science) Gerbrand A. Bredero 1. Het daget uyt den oosten rhymed/lyrical 2. Rodd’rick ende Alphonsus rhymed/lyrical Gerrit H. van Breughel 1. Cupido’s lusthof rhymed/lyrical 2. De tweede vijftigh lustige historien ofte nieuwigheden Johannis narrative Boccatii

Johan de Brune (de Jonge) Wetsteen der vernuften secular (argumentation) Johan de Brune (de Oude) Emblemata of Zinne-werck narrative Dirk R. Camphuysen 1. Stichtelycke rymen rhymed/lyrical 2. Uytbreyding over De Psalmen des propheten Davids rhymed/lyrical Roeland de Carpentier Verhael-boecken van den cardinael Bentivoglio personal (autobiography) Jacob Cats 1. Aenmerckinghe op de tegenwoordige steert-sterre en drie rhymed/lyrical lofdichten op Philips van Lansbergen 2. Houwelick rhymed/lyrical 3. Klagende maeghden en raet voor de selve rhymed/lyrical 4. Maechden-plicht ofte ampt der ionkvrouwen, in eerbaer liefde, rhymed/lyrical aenghewesen door sinne-beelden 5. Silenus Alcibiadis, sive Proteus rhymed/lyrical 6. Trouringh rhymed/lyrical Leenaert Clock 1. Het groote liede-boeck rhymed/lyrical 2. Veelderhande schriftuerlicke nieuwe liedekens, vermaninghen, rhymed/lyrical leeringhen, ghebeden ende lofsanghen Jacob J. Colevelt, e.a. 1. De Amsterdamsche Pegasus rhymed/lyrical 2. Nieu liedt-boeck genaemt den Amsterdamsche pegasus rhymed/lyrical

APPENDIX 1 185

Table 58 END sources – Northern Dutch (part 2)

Author name Text name Text type Dirck V. Coornhert 1. t Bedrogh des werelts secular (philosophy) 2. Boëthius: Van de vertróósting der wysheyd narrative 3. Consistorie, handelende van 't niet hanteren des nachtmaels secular (argumentation) 4. Corte berispinge. Vande leere Calvini vande voorsienigheyt Godes secular (argumentation) 5. Dat des Duyvels wet swaar is ende lastigh secular (philosophy) 6. Dat Godts gheboden licht zijn ende leerlijck secular (argumentation) 7. Dat onverstandigh blijven des menschen eenighe zonde ende secular (philosophy) oorsake van alle doolinghen zy 8. Dattet afsterven vant quade licht valt secular (philosophy) 9. Disputatie over den catechismus van Heydelbergh, openbaerlijck secular (argumentation) voor den volcke gehouden op't Hof van 'sGraven-Haghe in Hollandt, anno 1583 10. Gesprake van liefhebbers des ghemeynen nuts secular (philosophy) 11. Hemel-werck, ofte quay-toe-verlaet secular (philosophy) 12. Het kruyt-hofken secular (argumentation) 13. Kleyn-munster, des groot-roemigen David Jorisens roemrijcke ende secular (argumentation) wonderbare schriften elckerlijck tot een proeve gestelt 14. Ladder Iacobs, of trappe der deughden secular (philosophy) 15. Levende kalck secular (argumentation) 16. Lied-boeck rhymed/lyrical 17. Lieffelijcke tsamenspreecking van de droefheydt secular (philosophy) 18. Lijdens troost religious (theology) 19. Nadencken opt sevende Capittel totten Romeynen secular (argumentation) 20. Ofde siele, dan of de wille zondight, ondersoeck secular (philosophy) 21. Oogh-water secular (argumentation) 22. Oorsaken ende middelen vander menschen saligheyt ende secular (philosophy) verdoemenisse 23. Op 't beclagh van 't mesbruyck des avontmaels secular (argumentation) 24. Paradoxa dialogue 25. Proeve vande Heydelberghsche Catechismo secular (philosophy) 26. Recht ghebruyck ende misbruyck van tydlycke have rhymed/lyrical 27. Schijndeugd der secten religious (theology) 28. Schole der deughden narrative 29. Suyveringe opten titule religious (theology) 30. Toetzsteen der ware leeraren secular (philosophy) 31. Tsamensprake of de quade willich quaat zijn of onwilligh: tusschen secular (philosophy) Peter ende Jan 32. Tsamensprekinghe, waar in bewesen wort dat hy niet goet en wil secular (philosophy) worden die quaat blijft 33. Twee-spraeck, of waerheydt vry maeckt? secular (philosophy) 34. Tweede Verantwoordinge Eens eenigen Sendbriefs, Eerst secular (philosophy) aengetast zijnde by Doctor Lambertus Danevs Ende nu andermael by den Predicanten tot Delft 35. Uut-roedinge van des verderfs plantinghe, dragende die secular (argumentation) verderffelijcke vrucht 36. Van 't kerck-bouwen der dooperen opten waenscherm secular (philosophy) 37. Van den aflaet Iesu Christi secular (argumentation) 38. Van den bejaerden doope secular (argumentation) 39. Van den boom des levens rhymed/lyrical 40. Van den onderscheyt tusschen die ware ende valsche leere secular (philosophy) 41. Van de onwetenheyt der menschen, die daer is onschuldigh of secular (philosophy) schuldigh

186 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 58 END sources – Northern Dutch (part 3)

Author name Text name Text type Dirck V. Coornhert 42. Vande predestinatie, verkiesinghe en de verwerpinghe Godes religious (theology) 43. Vande sendinghe secular (philosophy) 44. Van des menschen natuerlijcke vleesch wondersproock secular (philosophy) 45. Van de toelatinge ende decrete Godts secular (argumentation) 46. Van de vreemde sonde, schulde, straffe nasporinghe secular (argumentation) 47. Vande zendinghe der Lutheranen, Swinglianen, ende Mennonisten, secular (philosophy) kort, klaer ende vast bewijs, dat sy niet en is wettelijck 48. Vant oordelen, gesprake tusschen vermetel oordeel ende secular (philosophy) bescheyden ondersoeck 49. Vereeninghe van sommighe strijdich-schijnende sproken der H. secular (argumentation) Schrifturen 50. Vijftigh lustighe historien oft nieuwigheden Joannis Boccatij narrative 51. Vre-reden, of onderwijs tot eendracht, vrede ende liefde, in dese secular (philosophy) tyden hoogh-noodigh 52. Waarachtighe aflaat van zonden secular (argumentation) 53. Wagen-spraeck secular (argumentation) 54. Zedekunst dat is wellevenskunste secular (argumentation) Jan J. Deutel t Kleyn Hoorns-liet-boeck, inhoudende eenige psalmen Davids, lof- rhymed/lyrical sanghen, en geestelijcke liedekens Pieter Dubbels Helikon. Bestaande in zangen, kusjes en mengel-rijm rhymed/lyrical Desiderius Erasmus Lof der Zotheid narrative Thomas Fonteyn Arions vingertuig rhymed/lyrical Simon J. Fortuyn Geestelyck lietboeck genaemt de Basuyn rhymed/lyrical Pieter van Gelre Het triumph-hofjen Iesu Christi ende Marie rhymed/lyrical Soetken Gerijts Een nieu gheestelijck lietboecxken, twelck noch noyt in druck rhymed/lyrical gheweest en is wt den Ouden ende Nieuwen Testament ghemaeckt Pieter L. van der Goes Een nieu liedt-boeck genaemt den druyven-tros der amoureusheyt rhymed/lyrical Leonardus Gouwerak Erato rhymed/lyrical Hugo de Groot 1. Christelicke gesanghen, ghetrocken uyt het Oude ende Nieuwe rhymed/lyrical Testament 2. Tractaet vande oudtheyt vande Batavische nu Hollandsche secular (history) republique Jan J. Harlingen Eenighe gheestelijcke liedekens rhymed/lyrical Johan van Heemskerck 1. Inleijdinghe tot het ontwerp van een Batavische Arcadia rhymed/lyrical 2. Pvb. Ovidii Nasonis Minne-kunst, gepast op d'Amsterdamsche rhymed/lyrical vryagien: met noch andere minne-dichten ende mengel-dichten, alle nieu ende te voren niet gesien Johan van Heemskerck & Minne-plicht ende kuysheyts-kamp rhymed/lyrical Joost van den Vondel Frans E. den Heussen Byvoeghsel tot den Christelijcken jongelingh rhymed/lyrical Zacharias Heyns 1. Const-thoonende juweel rhymed/lyrical 2. Emblemata moralia rhymed/lyrical 3. Emblemata. Vol-sinnighe uytbeelsels rhymed/lyrical Pieter C. Hooft 1. Baeto, of oorsprong der Hollanderen rhymed/lyrical 2. Nederlandsche Historien secular (history) Pieter C. Hooft, e.a. Emblemata amatoria rhymed/lyrical Willem D. Hooft, e.a. Nieu liedt-boeck ghenaemt Der minnaers harten jacht ofte de Groote rhymed/lyrical Aemstelredamsche rommelzoo Jan van Hout 1. Liedt op de Standtvasticheyt' rhymed/lyrical 2. Loff-sang op 't verlossen vande Burch Graven' rhymed/lyrical 3. Onrijmich vreuchden-liedt rhymed/lyrical 4. Openingsgedicht Album Amicorum' rhymed/lyrical

APPENDIX 1 187

Table 58 END sources – Northern Dutch (part 4)

Author name Text name Text type Jan van Hout 5. Opt ontset van Leyden. Lofsang' rhymed/lyrical 6. Rederijkerskaart 1577 rhymed/lyrical 7. Rederijkerskaart 1578 rhymed/lyrical Anthonis de Hubert De Psalmen des Propheeten Davids, in't Hebreeuz genaamd het lof- rhymed/lyrical bouk: bij den propheet verdeeld in vijf bouken Anthony Jansen Christelijck vermaeck rhymed/lyrical Nicolaes Janssens Een nieu devoot geestelijck lietboeck rhymed/lyrical Daniel Joncktys 1. Hedensdaegse Venus en Minerva of twist-gesprek tusschen die rhymed/lyrical zelfde 2. Roseliins oochies, ontleedt rhymed/lyrical Alhardt L. Kok Ont-werp der Neder-duitsche letter-konst secular (linguistics) Jan S. Kolm Levenders reden-feest, oft Amsteldams Helicon rhymed/lyrical Jan A. Leeghwater Haerlemmermeerboeck secular (science) Cornelis de Leeuw Christelijcke plicht-rymen om te singen of te leesen rhymed/lyrical Jacob C. Mayvogel 1. Row klacht ver-toont. in tyt van droefheydt rhymed/lyrical 2. Schadt-kist der Liefde rhymed/lyrical Willem Meerman Comoedia vetus secular (argumentation) Petrus Montanus De spreeckonst secular (linguistics) Pieter C. van der Morsch Cort verhael van tPrincipael, in Leyden bedreven, by Sotten meest, die rhymed/lyrical op Vrou Lors Feest waren verschreven Nicolaus Mulerius Hemelsche trompet morgenwecker secular (science) Pieter Nootmans Jeugdige minne-spiegel rhymed/lyrical Jillis Noozeman Lichte Klaartje rhymed/lyrical Guillaume A. Ooijevaer Haerlemsche Somer-Bloempjes tweede offer, Aen de Vreught-lievende rhymed/lyrical Nymphjes Willem van Oranje Apologie, ofte Verantwoordinghe statutory (politics) Jan J. Orlers 1. Beschrijvinge der stad Leyden (fragment) secular (history) 2. Verscheyden bruyloft dichten ende liedekens ghedicht ende rhymed/lyrical gecomponeert by verscheyden gheesten E. Pels (de Jonge) 1. t Amsteldams flvytertie fluytende op zyn fluytje sijn Amsteldamsche rhymed/lyrical boertighe ghesangen op boertighe voysjes 2. t Lof van Cupido rhymed/lyrical Dirck P. Pers 1. Bellerophon of Lust tot wiisheit, Gesangh der zeeden, Urania of rhymed/lyrical Hemel-sangh 2. Suyp-stad of Dronckaarts leven rhymed/lyrical Lucas van de Poll Bruylofs-gedicht, ter eeren van den E. Gosuinus de Wit, ende joffr. rhymed/lyrical Elizabeth de l'Homell Hendrick J. Prins Medenblicker scharre-zoodtje rhymed/lyrical Gillis J. Quintijn De Hollandsche Liis met de Brabandsche Bely rhymed/lyrical Jacobus Revius 1. De CL Psalmen Davids, eerst in Nederlantschen dichte gebracht door rhymed/lyrical Petrum Dathenum, ende nu in sin ende rijmen gebetert 2. Over-Ysselsche sangen en dichten rhymed/lyrical Jan Six Medea rhymed/lyrical J. Smeerbol Bruylofts-kost rhymed/lyrical Hendrik J. Soetenboom t Kleyn lust-hofje vol van bruyloft-zangen, echt-liedekens, houwlijcx- rhymed/lyrical baken en feest-gedichten Haerlem Soetendal Liedekens ende Refereynen Ghemaeckt by Haerlem Soetendal, van rhymed/lyrical zijn Avontueren Hendrik L. Spiegel 1. Twe-spraack; Ruygh-bewerp; Kort begrip; Rederijck-kunst secular (argumentation) 2. Hert-spiegel rhymed/lyrical Jacob Steendam Den distelvink rhymed/lyrical Mattheus G. Tengnagel 1. Frik in 't veur-huys rhymed/lyrical

188 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 58 END sources – Northern Dutch (part 5)

Author name Text name Text type Mattheus G. Tengnagel 2. Het leven van Konstance: waer af volgt het toneelspel De Spaensche rhymed/lyrical heidin Augustinus van Teylingen Devote oeffeninghe op de vijf lettere van de soete ende rhymed/lyrical alderheylichste namen Iesus ende Maria Pieter J. Twisck Kleyn liedtboecxken rhymed/lyrical Dirck A. Valcooch Den reghel der Duytsche schoolmeesters rhymed/lyrical Jan van der Veen 1. Over-zeesche zege en bruylofts-zangen rhymed/lyrical 2. Zinne-beelden, oft Adams appel rhymed/lyrical Gerrit de Veer Waerachtighe beschryvinghe van drie seylagien, ter werelt noyt soo personal (diary) vreemt ghehoort Matthijs van Velden De roemster van den Aemstel off poëtische beschrijvinghe van de rhymed/lyrical riviere Aemstel Adolphus Venator Een claer ende doorluchtich vertooch van d'Alckmaersche kerckelicke rhymed/lyrical gheschillen Christianus Vermeulen t Ronde jaer, of den schat der geestelijcke lofsangen rhymed/lyrical Andries Vierlingh Tractaet van dyckagie secular (argumentation) Roemer Visscher Sinnepoppen narrative Michiel Vlack Den nieuwen lust-hof rhymed/lyrical Joost van den Vondel 1. Aenleidinge ter Nederduitsche Dichtkunde secular (linguistics) 2. Dichtwerken en oorspronklijke prozaschriften. Deel 1: 1605-1620 rhymed/lyrical 3. Elektra rhymed/lyrical 4. Gebroeders rhymed/lyrical 5. Het Pascha ofte de Verlossinge Israels wt Egypten rhymed/lyrical 6. Hierusalem verwoest rhymed/lyrical 7. Huigh de Groots Josef of Sofompaneas rhymed/lyrical 8. Joseph in Dothan rhymed/lyrical 9. Joseph in Egypten rhymed/lyrical 10. Maeghden rhymed/lyrical 11. Maria Stuart of Gemartelde majesteit rhymed/lyrical 12. Nasoos Heldinnebrieven personal (letter) 13. Leeuwendalers rhymed/lyrical 14. Palamedes oft Vermoorde onnooselheyd rhymed/lyrical 15. Peter en Pauwels rhymed/lyrical Isaak Vos Klucht van de Moffin rhymed/lyrical Lucas J. Waghenaer Spiegel der zeevaert secular (science) Boudewijn J. Wellens t'Vermaeck der jeught rhymed/lyrical Thonis van Wervershoef Een suyverlick boecxken begrypende alle de gheestelicke liedekens rhymed/lyrical Joannes S. van der Wiele 1. Extractum Katholicum, tegen alle gebreken van Verwarde Harsenen rhymed/lyrical 2. Gulde-jaers feest-daghen of den schat der geestelycke lof-sangen rhymed/lyrical gemaeckt op elcken feest dagh van 't geheele Iaer Claes J. Wits Stichtelijcke bedenckinge, onledige ledigheyt, stichtelijcke tijt-kortinge rhymed/lyrical Pieter van Zeerijp D'eersuchtige wraak rhymed/lyrical

APPENDIX 1 189

Table 59 END sources – Southern Dutch (part 1)

Author name Text name Text type Philips van Marnix van 1. De bijencorf der H. Roomsche Kercke narrative Sint Aldegonde 2. Het boeck der Psalmen. Wt de Hebreische sprake in Nederduitschen rhymed/lyrical dichte Cornelis de Bie Den groeyenden Lierschen blom-hof rhymed/lyrical Guilielmus Bolognino Den Gheestelycken leeuwercker vol godtvruchtighe liedekens ende rhymed/lyrical leyssenen Willem van der Borcht Brusselschen Blom-hof van Cupido rhymed/lyrical Hendrik de Coster Het oudt Huysken van Bethleem rhymed/lyrical Petrus Datheen De Psalmen Davids, ende ander lofsanghen rhymed/lyrical Jan David Christeliicken waerseggher religious (sacral) Eduard de Dene 1. De warachtighe fabulen der dieren rhymed/lyrical 2. Testament rhetoricael rhymed/lyrical Johan Fruytiers 1. Den Sendtbrief Pauli tot den Romeynen rhymed/lyrical 2. Ecclesiasticus of de wijse sproken Iesu des soons Syrach. Nu rhymed/lyrical eerstmael eurdeelt ende ghestelt in Liedekens, op bequame en ghemeyne voisen Philip Fruytiers, e.a. Typus Mundi rhymed/lyrical Cornelis van Ghistele Der Griecxser princerssen, ende jonckvrouwen clachtige sendtbrieven, rhymed/lyrical Heroidum epistolae ghenaemt Lodovico Guicciardini Beschrijvinghe van alle de Neder-landen secular (science) Godevaert van Haecht De kroniek van Godevaert van Haecht over de troebelen van 1565 tot secular (history) 1574 te Antwerpen en elders Willem van Haecht De Psalmen Dauids rhymed/lyrical Justus de Harduwijn 1. Den val ende op-stand vanden coninck ende prophete David rhymed/lyrical 2. De weerliicke liefden tot Roose-mond rhymed/lyrical 3. Goddelicke lof-sanghen rhymed/lyrical 4. Goddelycke wenschen narrative Lucas de Heere 1. Den hof en boomgaerd der poësien rhymed/lyrical 2. Psalmen Davids na d'Ebreeusche waerheit rhymed/lyrical Abraham de Koning Het tweede dochters-speeltjen rhymed/lyrical Justus Lipsius Twee boecken vande stantvasticheyt secular (argumentation) Joannes de Lixbona Hemelsch nachtegaelken oft gheestelijcke liedekens, om deught met rhymed/lyrical vreught te beoeffenen Lodewijk Makeblijde 1. Den berch der geestelijcker vreughden narrative 2. Den Lusthof der gheestelicke oeffeninghen religious (sacral) 3. Den schat der christelicker leeringhe tot verklaringhe van den religious (sacral) catechismus 4. Troost der siecken ende overleden secular (philosophy) Karel van Mander 1. Bethlehem dat is het broodhuys rhymed/lyrical 2. Bucolica en Georgica, dat is, Ossen-stal en Landt-werck rhymed/lyrical 3. Dat hooghe liedt Salomo, met noch andere gheestelycke liedekens rhymed/lyrical 4. De gulden harpe, inhoudende al de geestelijcke liedekens rhymed/lyrical 5. De gulden harpe, inhoudende al de liedekens, die voor desen by rhymed/lyrical K.V.M. gemaeckt, ende in verscheyden Boecxkens uyt-ghegaen zijn 6. De harpe, oft des herten snarenspel rhymed/lyrical 7. Het Herder Pijpken rhymed/lyrical 8. Het schilder-boeck secular (information) Lucas van Mechelen 1. Cloosterken der gheestelijcke verryssenisse ofte der rhymed/lyrical ontwordentheyt 2. Den boeck der gheestelijcke sanghen rhymed/lyrical Jan Mommaert Het Brabandts nachtegaelken, met zijn driederley gesangh, te weten rhymed/lyrical minne-liedekens, herders-sanghen, ende boertigheden

190 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 59 END sources – Southern Dutch (part 2)

Author name Text name Text type Adriaen Poirters 1. Het masker van de wereldt afgetrocken narrative 2. Ydelheyt des werelts rhymed/lyrical G. de Preter Gheestelijck paradiisken der wel-lusticheden. (Delen 1-6) rhymed/lyrical Hans de Ries Lietboeck inhoudende schriftuerlijcke vermaen liederen, claech rhymed/lyrical liederen, gebeden, danck liederen, lofsanghen, psalmen, ende ander stichtelijcke liederen Anthonis de Roovere Rethoricale wercken rhymed/lyrical Jacob van der Schuere Nederduytsche spellinge secular (linguistics) Anthoni Smyters Het versierde woord. De Epitheta of woordcombinaties van Anthoni secular (information) Smyters uit 1620 Simon Stevin 1. De Thiende secular (science) 2. Het burgherlyck leven secular (lifestyle) 3. Wisconstighe gedachtenissen secular (science) Antonius Sucquet Den wech des eeuwich levens narrative Jan Thieullier De schadt-kiste der philosophen ende poeten rhymed/lyrical Jan Utenhove 1. LXIIII, Psalmen end ander ghesangen, diemen in de Duytsche rhymed/lyrical Ghemeynte te Londen was ghebruyckende 2. 11 ander Psalmen rhymed/lyrical 3. 25 Psalmen rhymed/lyrical 4. 26 Psalmen rhymed/lyrical 5. Psalmen Davidis rhymed/lyrical Marcus van Vaernewyck 1. De historie van Belgis secular (history) 2. Van die beroerlicke tijden in die Nederlanden en voornamelick in secular (history) Ghendt 1566-1568 Jacobus Viverius Den lvst-hof van de christelicke zielen rhymed/lyrical Joris Wybo Gheestelijcke liedekens rhymed/lyrical Joan Ysermans Triumphus Cupidinis. Encomium matrimonii. Nederlantsche poëmata rhymed/lyrical Ambrosius Zeebout Tvoyage van Mher Joos van Ghistele secular (information) Jacob van Zevecote Belech van Leyden rhymed/lyrical

APPENDIX 1 191

Table 60 ND sources – Northern Dutch (part 1)

Author name Text name Text type Christianus van der Aa 1. Lenteloveren rhymed/lyrical 2. Veldviooltjes rhymed/lyrical Christianus van der Aa & Bekroonde volksliederen, uitgegeven door de Maatschappij tot Nut rhymed/lyrical Carel G. Withuys van ’t Algemeen Joseph Alberdingk Thijm 1. Beatrijs' rhymed/lyrical 2. De klok van Delft rhymed/lyrical 3. Drie gedichten rhymed/lyrical A.C. Alebers Klagt op het verlies mijner beide beenen rhymed/lyrical Algemene konsten Algemene konsten letterbode: Jaargang 1 secular (information) letterbode (tijdschrift) Hieronymus van Alphen 1. Dichtwerken (3 delen) rhymed/lyrical 2. Digtkundige verhandelingen secular (argumentation) 3. Kleine gedigten voor kinderen rhymed/lyrical 4. Proeve van liederen en gezangen voor den openbaaren godsdienst rhymed/lyrical 5. Proeve van stigtelijke mengel-poëzij (3 delen) rhymed/lyrical 6. Theorie der schoone kunsten en wetenschappen. Eerste deel secular (science) Nicolaas Anslijn Nz. 1. De brave hendrik narrative 2. De brave Maria narrative Archief voor Neder- Archief voor Nederlandsche taalkunde secular (information) landsche taalkunde Robert H. Arntzenius 1. Alexander, keizer aller Russen rhymed/lyrical 2. De Hollander zoo als 'er wel meer zijn rhymed/lyrical Dirk Bax Hulde aan de Nederlandsche zeemagt op de Schelde rhymed/lyrical Adriaan Beeloo 1. 's Gravenhage rhymed/lyrical 2. Nederland in gebed tot God; op den algemeenen bededag, den rhymed/lyrical tweeden december 1832 Nicolaas Beets 1. Jose, een Spaansch verhaal rhymed/lyrical 2. Kuser rhymed/lyrical 3. Twaalf preeken religious (theology) Jacobus Bellamy 1. Gedichten rhymed/lyrical 2. Gezangen mijner jeugd rhymed/lyrical 3. Twee nagelaaten leerredenen secular (argumentation) 4. Proeven voor het verstand, den smaak en het hart secular (argumentation) Katharina W. Bilderdijk- 1. Feestzang rhymed/lyrical Schweickhardt 2. Gedichten voor kinderen rhymed/lyrical Willem Bilderdijk 1. Avondschemering rhymed/lyrical 2. Beginsels der woordvorsching secular (argumentation) 3. Brieven (5 delen) personal (letter) 4. De kunst der poëzy rhymed/lyrical 5. De ziekte der geleerden rhymed/lyrical 6. Dichterlijke uitspanning rhymed/lyrical 7. Galante dichtluimen rhymed/lyrical 8. Geschiedenis des vaderlands. (Deel 1-13) secular (history) 9. Het nicotiaansche kruid en Uitzicht op mijn dood rhymed/lyrical 10. Korte aanmerkingen op Huydecopers proeve van taal- en secular (argumentation) dichtkunde 11. Nederlandse spraakleer secular (linguistics) 12. Nieuwe taal- en dichtkundige verscheidenheden (4 delen) secular (linguistics) 13. Taal- en dichtkundige verscheidenheden (4 delen) secular (linguistics) 14. Winterbloemen rhymed/lyrical Willem Bilderdijk, e.a. Nieuw liederenboekje, op aangename en bekende wijzen rhymed/lyrical Pieter Boddaert jr. Levensgeschiedenis van den vermaarden dichter personal (autobiography)

192 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 60 ND sources – Northern Dutch (part 2)

Author name Text name Text type Lodewijk N. Bonaparte Geschiedkundige gedenkstukken en aanmerkingen over het bestuur personal (autobiography) van Holland (3 dln) Elias A. Borger Op de bevalling der Prinses van Oranje rhymed/lyrical Anna L.G. Bosboom- 1. De vrouwen van het Leycestersche tijdvak (2 delen) narrative Toussaint 2. Don Abbondio II narrative 3. Engelschen te Rome. Romantische épisode uit de regering van paus narrative Sixtus V 4. Het huis Lauernesse narrative 5. Verspreide verhalen narrative Bernardus Bosch & Liederen voor het vaderland rhymed/lyrical Petronella Moens Bernardus de Bosch Toejuiching aan den welëdelen grootachtbaaren heer mr. Pieter rhymed/lyrical Clifford Johannes van den 1. Nederlandsche bezittingen in Azia, Amerika en Afrika secular (argumentation) Bosch 2. Verhandeling over de mogelijkheid, de beste wijze van invoering, en secular (argumentation) de belangrijke voordeelen eender algemeene armen-inrigting in het Rijk der Nederlanden, door het vestigen eener Landbouwende Kolonie in deszelfs Noordelijk gedeelte Joannes N. de 1. Ambiorix rhymed/lyrical Brauwere van Steeland 2. Dichtluimen rhymed/lyrical 3. Een reisje in het noorde narrative 4. Ernst en boert rhymed/lyrical Gerrit Brender à 1. De gouden bruiloft van Kloris en Roosje dialogue Brandis 2. De schaaking uit het Serail rhymed/lyrical J.D.B. Brouwer Zielsgestalten des christens. Acht predikatien over het Hooglied van religious (theology) Salomo Abraham J. de Bull & De val van Jeruzalem rhymed/lyrical Jacob van Lennep Anne van Buren Schele Magdalena Moons, of het Beleg van Leyden narrative Johannes II Bussingh Gezangen voor de gereformeerde kerk van Nederland rhymed/lyrical Petrus Camper Berigt van den zaaklyken inhoud van twee lessen, gegeeven aan de secular (argumentation) leden van de Teken-Akademie te Amsterdam Joan D. van der Capellen Aan het volk van Nederland secular (argumentation) tot den Pol Willem de Clercq 1. Gedenkzuil voor W. Bilderdijk secular (information) 2. Hulde aan den heer J.M. Kemper rhymed/lyrical 3. Tafelzang rhymed/lyrical 4. Verhandeling ter beandwoording der vraag welken invloed heeft secular (argumentation) vreemde, inzonderheid de Italiaansche, Spaansche, Fransche en Duitsche, gehad op de Nederlandsche taal- en letterkunde sints het begin der vijftiende eeuw tot op onze dagen? 5. Waarheid. Eene behoefte ook in de letterkunde secular (argumentation) 6. Welkomst-groete aan Z.M. Koning Willem II rhymed/lyrical Maarten Corver Tooneel-aantekeningen personal (autobiography) Isaäc da Costa Bezwaren tegen den geest der eeuw secular (argumentation) Bruno Daalberg 1. Jan Perfect of De weg der volmaking, vertoond in het leven en de narrative zonderlinge lotgevallen van een voornaam wijsgeer 2. Willem Hups. Eene anecdote uit de XVII eeuw; ongelooflijk zelfs in narrative de onze Johannes Daane Haarlems schutterslied rhymed/lyrical Herman W. Daendels Aan zyne Geldersche en Overysselsche landgenooten secular (argumentation) De Denker (tijdschrift) De Denker secular (information)

APPENDIX 1 193

Table 60 ND sources – Northern Dutch (part 3)

Author name Text name Text type Aagje Deken & Betje 1. Economische liedjes rhymed/lyrical Wolff 2. Historie van den heer Willem Leevend (8 delen) narrative 3. Historie van mejuffrouw Cornelia Wildschut (6 delen) narrative 4. Historie van mejuffrouw Sara Burgerhart narrative Bartholomeus Verhalen uit het nieuwe testament narrative Doorenweerd Aernout Drost Schetsen en verhalen narrative Lukas Egeling De weg der zaligheid, naar het beloop des Bijbels religious (theology) Jan van Eijk Liederen voor den landman. Deel 1 rhymed/lyrical Kornelis Elzevier & Maendelyks musikaels tydverdryf rhymed/lyrical Antoine Mahaut Cornelius van Engelen Eene wysgeerige verhandeling over den schouwburg in 't algemeen' secular (argumentation) Jan van Esveldt Holtrop Gezangboek voor vrijmetselaaren rhymed/lyrical Don Experientia Het Surinaamsche leeven, toneelschwyse verbeeld rhymed/lyrical Rhijnvis Feith 1. Brieven aan Sophie rhymed/lyrical 2. Brieven over verscheide onderwerpen (6 delen) personal (letter) 3. De eenzaamheid en de wereld rhymed/lyrical 4. De ouderdom rhymed/lyrical 5. De patriotten dialogue 6. Het ideaal in de kunst secular (philosophy) 7. Het graf rhymed/lyrical 8. Julia narrative 9. Lady Johanna Gray rhymed/lyrical Philip Fermin Nieuwe algemeene beschryving van de colonie van Suriname secular (information) Quirijn de Flines 1. Aan Neerlands heldenstoet, trekkende ten heiligen stryd, voor rhymed/lyrical vryheid en onafhanglykheid 2. Een lied voor Nederlanders rhymed/lyrical H.J. Foppe Oproer in Brussel rhymed/lyrical Johannes le Francq van 1. Het verheerlijkt Leyden rhymed/lyrical Berkhey 2. Lijkzang in heldendicht, ter nagedagtenis van den beroemden rhymed/lyrical Leydschen schaatsrijder Cornelis Fleur 3. Natuurlyke historie van Holland. Deel 1-9 secular (science) Albertus Frese & De inënting rhymed/lyrical Christiaan Schaaf Cornelis Gébel De dertigste november 1813 rhymed/lyrical Jacob Geel 1. Mededeeling aan alle recenserende geleerden in ons vaderland secular (argumentation) 2. Onderzoek en phantasie secular (argumentation) 3. Over den gang van het onderzoek in Oude Geschiedenis en Letteren secular (argumentation) en deszelfs grenzen Gerrit Gerrits & Pieter G. Schoonheden en merkwaardige tafereelen uit de Nederlandsche secular (history) Witsen Geysbeek geschiedenis De Gids (tijdschrift) De Gids secular (information) Rijklof M. van Goens Vrymoedige bedenkingen over de vergelyking der oude dichteren met secular (argumentation) de hedendaegschen Leendert Grendel Nederlandsche gezangen rhymed/lyrical Hendrik v. Griethuysen De zekere verwachting des Christens religious (theology) Bernard ter Haar 1. De Sint-Paulusrots rhymed/lyrical 2. Huibert en Klaartje rhymed/lyrical 3. Joannes en Theagenes. Eene legende uit de apostolische eeuw rhymed/lyrical J.G. Haffner 1. Lotgevallen op eene reize van Madras over Tranquebaar naar het narrative eiland Ceilon

194 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 60 ND sources – Northern Dutch (part 4)

Author name Text name Text type J.G. Haffner 2. Reize in eenen palanquin, of Lotgevallen en merkwaardige aan- narrative teekeningen op eene reize langs de kusten Orixa en Choromandel 3. Reize naar Bengalen en terugreize naar Europa narrative 4. Reize te voet door het eiland Ceilon narrative Maurits C. van Hall 1. Bij den dood van Richeus van Ommeren, rector der Latijnsche rhymed/lyrical schoole te Amsteldam 2. Gedenkschriften van en door Frank Floriszoon van Arkel narrative Onno Zwier van Haren 1. Pietje en Agnietje, of de doos van Pandora dialogue 2. Van Japan secular (science) Betsy Hasebroek 1. Te laat narrative 2. Twee vrouwen narrative Jan F. Helmers 1. De Hollandsche natie rhymed/lyrical 2. Gedichten rhymed/lyrical A. Gijsberti Hodenpijl De feestdag van Nederland rhymed/lyrical Willem van Hogendorp Kraspoekol, of de droevige gevolgen van eene te verre gaande secular (argumentation) strengheid, jegens de slaaven Johan van Hoogstraten 1. Op den vijftigjaarigen trouwdag mijner hoogstgeëerde ouderen: rhymed/lyrical Hendrik van Hoogstraten en Anna Catharina Koningh 2. Op het smertelyk overlyden van den hoog eerwaarden en zeer rhymed/lyrical geleerden heer Herman Gideon Clemens Willem van den Hoonaard Bij het vertrek van het Koopvaardijschip De Wilhelmina, naar Batavia, rhymed/lyrical den november 1822, waarmede drie mijner gewezene kostleerlingen die reis ondernamen Adriaan van der Hoop 1. De renegaat rhymed/lyrical 2. Aan het kasteel van Antwerpen rhymed/lyrical Adriaan van der Hoop 1. De horoskoop rhymed/lyrical jr. 2. Han van Ijsland dialogue 3. Leyden ontzet, in 1574 rhymed/lyrical 4. Vier Neêrland feest rhymed/lyrical Arie de Jager Proeve over de invloed van Bilderdijk's dichtwerken op onze taal, ten secular (linguistics) aanzien van het vormen van oude woorden en spreekwijzen Gerrit van der Jagt Het vaderland, met geheel nieuw geteekende en gekleurde kaarten secular (history) van iedere provincie, de generaliteits-landen en de geheele republiek, en zeer gewigtige vaderlandsche afbeeldingen Jakob Jansen Verhaal der Merkwaardige Reize van den Kommandeur Jakob Jansen personal (travel story) Marten Jansen Kort, doch echt-verhaal van commandeur Marten Jansen (Uitgever G. secular (science) Tresling, Leeuwarden) Johannes Jelgerhuis Theoretische lessen over de gesticulatie en mimiek secular (information) Willem J.A. Jonckbloet Boekbeoordeelingen. [Recensie van: B.H. Lulofs, Handboek van den secular (argumentation) vroegsten bloei der Nederlandsche letterkunde […]. Groningen, 1845' Jan Jordens Nachtgepeinzen, bij de belegering der stad Naarden rhymed/lyrical Jan W. Kals Neerlands hooft- en wortelsonde religious (theology) Hidde D. Kat Dagboek eener reize ter walvisch- en robbenvangst, gedaan in de jaren personal (travel story) 1777 en 1778 door den kommandeur Hidde Dirks Kat, met eene kaart van Groenland Roeland de Kater Vlaardings vissers lied-boek rhymed/lyrical Frederik A. van der Kemp Lierzang aen de hoog-ed. gestrenge heeren F.J.J. Eisinga, S.H. Roorda rhymed/lyrical van Eisinga, H. Buma, F.B. Aebinga van Humalda, Regn. Liv. Andringa de Kempenaer, Johannes Wielinga, C.L. van Beyma Franciscus L. Kersteman Hollandsche pleit-zaal, of magazyn van singuliere processen; dienende statutory (administration) tot een vervolg van de academie der jonge practizyns Jan P. de Keyser Nederland. Proza en poëzij van Nederlandsche auteurs narrative

APPENDIX 1 195

Table 60 ND sources – Northern Dutch (part 5)

Author name Text name Text type Johannes Kinker 1. De menschheid in het Lazarushuis rhymed/lyrical 2. De Post van den Helicon secular (information) 3. Gedichten rhymed/lyrical 4. Proeve eener Hollandsche prosodia secular (linguistics) Willem Kist Het leven, gevoelens en zonderlinge reize van den landjonker Govert narrative Hendrik Godefroi van Blankenheim tot den Stronk (2 delen) Pieter H. Klaarenbeek Jan de Lapper. Een verhaal uit de geschiedenis des vaderlands rhymed/lyrical Cornelis van Koetsveld Schetsen uit de pastorij te Mastland narrative Jan Lagendaal Zegengroet aan den welëerwaardigen, zeergeleerden heere, den rhymed/lyrical heere Joachimus Mobachius Pieter Langendijk 1. De gedichten. Deel 3 rhymed/lyrical 2. De gedichten. Deel 4 rhymed/lyrical Juliana C. de Lannoy Aan mynen geest rhymed/lyrical Abraham J. Lastdrager Mnemosyne narrative Jacob van Lennep 1. De lotgevallen van Ferdinand Huyck narrative 2. De pleegzoon (2 delen) narrative 3. Hulde aan de nagedachtenis van Hollands zeeheld J.C.J. van Speyk rhymed/lyrical 4. Nederland in den goeden ouden tijd personal (diary) 5. Onze voorouders in verschillende taferelen geschetst (5 delen) narrative Thomas van Limburg De watervloed van 1809 openbaar voorgeleezen in de maatschappij rhymed/lyrical Diligentia, ten behoeve van onze ongelukkige landgenooten, op donderdag, den 23 februari 1809 Petrus van Limburg 1. Handboek der Grieksche mythologie, ten dienste van Latijnsche secular (information) Brouwer scholen en gymnasiën 2. Het leesgezelschap te Diepenbeek narrative Joannes van der Linden Verhandeling van het notaris-ambt in Frankrijk secular (argumentation) Izaak J. Lion Proeven van een humoristisch-satyriek woordenboek der zamenleving narrative Adriaan Loosjes 1. Frank van Borselen en Jacoba van Beijeren dialogue 2. Het leven van Johannes Wouter Blommesteyn narrative 3. Het leven van Maurits Lijnslager narrative 4. Historie van Mejuffrouw Susanna Bronkhorst (6 delen) narrative 5. Volks lieden-boek rhymed/lyrical 6. Zangen voor het feest des Bataafschen volks rhymed/lyrical Adriaan Loosjes & Jan van Democritische Tafelliedjes rhymed/lyrical Walré Petrus Loosjes & Jan Vaderlandsche historie secular (history) Wagenaar Barthold H. Lulofs 1. De kunst der mondelijke voordracht of uiterlijke welsprekendheid secular (lifestyle) 2. Gronden der Nederlandsche woordafleidkunde, voor zoo ver secular (linguistics) dezelve eenigzins zeker is 3. Kakographie, of opstellen met deze en gene feilen in taal en stijl, secular (information) benevens eenige taalkundige aanteekeningen 4. Vlugtige woorden over Nederlandsche taalzuivering en taalverrijking secular (linguistics) Hendrik Lussing, Gezangen voor de oude mannen en vrouwen, in 't Diaconie huis rhymed/lyrical Matthijsz. Jacob W. Lustig Vervolg van het musikaels tydverdryf rhymed/lyrical Aeneas Mackay Drie liederen aan mijn medeburgers rhymed/lyrical Adrianus Mandt Boet-zangen ter gelegenheid en gedachtenis van den geduchten rhymed/lyrical water-vloed des jaars MDCCXCIX Gerrit Manheer 1. Het zangeresje aan de Maas of Vervolg op het kransje van rhymed/lyrical letterbloempjes 2. Kransje van letterbloempjes rhymed/lyrical

196 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 60 ND sources – Northern Dutch (part 6)

Author name Text name Text type Jan F. Martinet 1. Katechismus der natuur (4 delen) secular (science) 2. Kleine katechismus der natuur voor kinderen secular (science) Martinus van Marum Schets der elektriciteit-kunde secular (science) Fenna Mastenbroek Wilhelmina Noordkerk. Eene geschiedenis ter aanprijzing van oud narrative vaderlandsche zeden, voor Nederlandsche vrouwen en meisjes Jan J. Mauricius 1. Besluit der dichtlievende uitspanningen, met verscheidene rhymed/lyrical byvoegzelen 2. Dichtlievende uitspanningen rhymed/lyrical 3. Vervolg der dichtlievende uitspanningen rhymed/lyrical Johan Meerman Eenige berichten omtrent het noorden en noord-oosten van Europa: personal (travel story) Deel 1-6. Anna B. van Meerten- 1. De kermiswandeling narrative Schilperoort 2. Gids voor jonge lieden van beschaafden stand secular (philosophy) Hendrik A. Meijer De boekanier rhymed/lyrical Lucretia van Merken Germanicus rhymed/lyrical Marie Michon & Thérèse Natuurlijk kookboek van beproefde en ondervonden echte recepten secular (lifestyle) de Milly voor een zindelijk huijshouden Petronella Moens 1. De dankbare Willem, of Het huisgezin van den heer Lausbach narrative 2. De jonge Sofia narrative 3. Gebeden voor kinderen rhymed/lyrical 4. Herfstbloempje, voor de lieve jeugd rhymed/lyrical 5. Hugo de Groot in zeven zangen rhymed/lyrical 6. Legaat aan mijne vrouwelijke landgenooten secular (argumentation) 7. Letter-kransje voor lieve en brave kinderen rhymed/lyrical 8. 's Menschen begin, midden en einde rhymed/lyrical 9. Mevrouw Veltman en hare voedsterlingen narrative De muzen (tijdschrift) De muzen secular (information) Margaretha J. de 1. Acht oorspronkelijk Nederduitsche verhalen voor de jeugd narrative Neufville 2. De kleine pligten narrative 3. De schildknaap narrative 4. Elisabeth Basmooth, of eene Engelsche plant op Hollandschen narrative bodem Joannes Nomsz 1. Iemant en Niemant rhymed/lyrical 2. Michiel Adriaansz. de Ruyter rhymed/lyrical Willem A. Ockerse Ontwerp tot eene algemeene characterkunde (3 delen) secular (psychology) Jan F. Oltmans De schaapherder narrative Joachim Oudaen Bruiloftzang voor den heere Pieter Claris en jufvrouwe Maria de rhymed/lyrical Planke Gerrit Paape 1. Het land der willekeurigen (onder pseudoniem J.A. Schasz) narrative 2. Mijne vrolijke wijsgeerte in mijne ballingschap personal (autobiography) 3. Reize door het Aapenland (onder pseudoniem J.A. Schasz) narrative Johannes van der Palm 1. Salomo religious (theology) 2. Tiental leerredenen. religious (sacral) Willem de Perponcher 1. Onderwijs voor kinderen secular (pedagogy) 2. Willem Essensteyn narrative Freerk Pieters 1. Aan-teekening, gehouden op het schip de Vrouw Maria secular (science) 2. Omstandig journaal of reys-beschryving personal (travel story) Olivier Porjeere Zanglievende uitspanningen rhymed/lyrical Elisabeth M. Post 1. Gezangen der liefde rhymed/lyrical 2. Het land, in brieven narrative 3. Reinhart, of natuur en godsdienst narrative

APPENDIX 1 197

Table 60 ND sources – Northern Dutch (part 7)

Author name Text name Text type Guillaume G. van 1. Aan G. Graaf Schimmelpenninck, minister van staat, enz. over de statutory (politics) Prinsterer vrijheid van onderwijs 2. Adviezen in de Tweede Kamer der Staten-Generaal. Zitting van 1849 statutory (politics) 3. Handboek der geschiedenis van het vaderland secular (history) 4. Ongeloof en revolutie secular (argumentation) 5. Verscheidenheden over staatsregt en politiek (6 delen) secular (argumentation) Simon Rivier 1. De jeugdige Michiel Adriaansz. de Ruiter in de touwbaan rhymed/lyrical 2. De verovering van Den Briel, of Grondlegging der Batavische vryheid rhymed/lyrical Bartholomeus Ruloffs Inwyding van den Amsteldamschen Schouwburg rhymed/lyrical Gijsbertus van Sandwijk De kleine mimiek, of De vrolijk zingende knaap rhymed/lyrical Willem Schellenbach De zangminnende kindervriend, of School-liederen voor drie stemmen rhymed/lyrical Jan Schenkman St. Nikolaas en zijn knecht rhymed/lyrical Jan E. Schut Galama of de bevrijding van Vriesland narrative Rutger Schutte Stichtelijke gezangen (4 delen) rhymed/lyrical Matthijs Siegenbeek 1. Beknopte geschiedenis der Nederlandsche letterkunde secular (linguistics) 2. Redevoering over het openbaar onderwijs in de Nederduitsche secular (argumentation) welsprekendheid Arend F. Simonsz 1. De moderne Helicon narrative 2. Dorus of het wonderkind secular (argumentation) 3. Het leven is een droom secular (argumentation) 4. Het onscheidbaar drietal redenwezens verlichting, deugd en tijd secular (argumentation) 5. Verhandelingen secular (argumentation) Rijer H. van Someren Zegelied voor den prins van Oranje en hoogstdeszelfs broeder rhymed/lyrical Abraham van Spall Evangelische feest-zangen rhymed/lyrical Hajo A. Spandaw Poëzij rhymed/lyrical Anthony C.W. Staring Gedichten rhymed/lyrical Frans van Steenwijk 1. Ada, gravin van Holland en Zeeland rhymed/lyrical 2. Beon, koning van Egipte rhymed/lyrical Simon Stijl 1. De torenbouw van het vlek Brikkekiks in het landschap Batrachia dialogue 2. De opkomst en bloei der Vereenigde Nederlanden secular (history) 3. Het leven van Jan Punt secular (information) 4. Krispyn filozoof rhymed/lyrical 5. Vier zamenspraaken nopens de aanmerkingen van den rustenden secular (information) tooneelspeeler M. Corver, op Het leven van Jan Punt Naatje van Streek- 1. Charakters en lotgevallen van Adelson, Héloïse en Elius narrative Brinkman 2. Op de terugkomst van Neerlands jongelingen, uit de Fransche rhymed/lyrical slavernij, eene voorspelling in den geest van Jesaia Johan H. Swildens 1. Deugden-boekje narrative 2. Vaderlandsch A-B boek voor de Nederlandsche jeugd secular (information) Marten D. Teenstra 1. Bijzonderheden betrekkelijk den brand te Paramaribo, in den nacht secular (science) van den 3den op den 4den september 1832 2. De negerslaven in de kolonie Suriname secular (information) Hendrik Tollens 1. Bij de geboorte van den jongen prins rhymed/lyrical 2. De algemeene bededag in Nederland (2 December 1832) rhymed/lyrical 3. s Konings verjaardag rhymed/lyrical 4. Lierzang bij de verheffing van Zijne Koninklijke Hoogheid Willem rhymed/lyrical Frederik, Prins van Oranje en Nassau, op den troon der Nederlanden 5. Vaderlandsch krijgslied rhymed/lyrical 6. Vaderlandsche wapenkreet in maart, 1815 rhymed/lyrical 7. Zangen, ter gelegenheid van de algemeene vergadering der rhymed/lyrical Maatschappij: tot Nut van 't Algemeen, op dinsdag, den 11den augustus, 1835.

198 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 60 ND sources – Northern Dutch (part 8)

Author name Text name Text type Vaderlandsche letter- Vaderlandsche letteroefeningen secular (information) oefeningen (tijdschrift) Nicolaas J. Veerman Vreugde-galm op de verkiesingh van den wel edelen gestrengen heer rhymed/lyrical mr. Wigbold Slicher, tot praesident van den Hove van Holland, Zeeland en Vriesland W.J. van der Vegt Alexis en Iwan, of de overwintering op Spitsbergen narrative Abdias Velingius CL geestlyke gezangen rhymed/lyrical Joos Verschuere Gezangen der vrije-metzelaaren rhymed/lyrical Reynvaan Jan Verveer Lijkzang voor den weledelen, gestrengen en grootachtbaren heere mr. rhymed/lyrical Adriaen Reepmaker J.G. Volckhardt Zang, tot uitgeleide, aan mijnen braaven, geleerden en hooggeachten rhymed/lyrical vriend J. Scharp Jacob Vosmaer Nagelaten en verspreide letter-arbeid secular (science) Pieter Vreede Vaderlandsche liederen, voor het genootschap van wapenhandel te rhymed/lyrical Leyden (4 delen) Jeronimo de Vries Proeve eener geschiedenis der Nederduitsche dichtkunde secular (information) Matthias de Vries De heerschappij over de taal, het beginsel der welsprekendheid secular (argumentation) Jan van Walré XII Volks-Liedekens, op bekende wijzen, Ter vervrolijking van Laurens rhymed/lyrical Jansz. Kosters Jan J.F. Wap Nieskruid voor den heer J.L. Nierstrasz, Jr secular (argumentation) Willem H. Warnsinck 1. Feestzangen, ter gelegenheid van het vijftigjarig bestaan der rhymed/lyrical maatschappij Felix Meritis 2. Op den heldendood van Justus Klinkhamer, luitenant ter zee rhymed/lyrical 3. Wachtlied, voor de vrijwillige schutters te Amsterdam rhymed/lyrical 4. Zangen, ter gelegenheid van de algemeene vergadering der rhymed/lyrical Maatschappij tot Nut van 't Algemeen, op dingsdag, den 8sten augustus, 1826 5. Zangen, ter gelegenheid van de algemeene vergadering der rhymed/lyrical Maatschappij: tot Nut van 't Algemeen, op dingsdag, den 9den augustus, 1831 6. Zangen, ter gelegenheid van de algemeene vergadering der rhymed/lyrical Maatschappij: tot Nut van 't Algemeen, op dingsdag, den 13den augustus, 1833 Petrus Weiland Nederduitsche spraakkunst Nederduitsche spraakkunst secular (linguistics) Marten Westerman De landstorm rhymed/lyrical Hendrik van Wijn Historische en letterkundige avondstonden secular (linguistics) Jacob C. Willem (le Jeune) Letterkundig overzigt en proeven van de Nederlandsche volkszangen secular (linguistics) sedert de XVde eeuw R. Willequet Vlaemsche kunst-schat, ten gebruike der lagere scholen secular (information) Nicolaas S. van Winter 1. De Amstelstroom rhymed/lyrical 2. De jaargetyden rhymed/lyrical Samuel I. Wiselius De dood van Karel rhymed/lyrical Carel G. Withuys Cantate, ter gelegenheid der viering van het vijftigjarig bestaan van rhymed/lyrical het Amsterdamsche eerste departement der Maatschappij tot Nut van 't Algemeen Pieter G. Witsen Alleenspraak van Gysbrecht van Aemstel, boertig berymd rhymed/lyrical Geysbeek Pieter van Woensel Aanteekeningen, gehouden op eene reize door Turkeyen, Natoliën, de personal (travel story) Krim en Rusland in de jaren 1784-89 Maria P. Woesthoven Nederland en Frankrijk in maart, 1815 rhymed/lyrical Betje Wolff Brieven van Constantia Paulina Dortsma personal (letter)

APPENDIX 1 199

Table 60 ND sources – Northern Dutch (part 9)

Author name Text name Text type Betje Wolff & Aagje Geschrift eener bejaarde vrouw narrative Deken Roelof Wouters Amstelskerk verheugd en dankbaar, over de herstelling van haaren rhymed/lyrical heilgezant, den weleerwaardigen Heere Johannes Henricus Westerhoof, van eene gevaarlijke krankheid Annaeus Ypeij Beknopte geschiedenis der Nederlandsche tale secular (linguistics) Willem J. van Zeggelen In de speeluren narrative Daniël J. ten Zeldam 1. Handleiding tot de kennis van het staatsbestuur in het Koninkrijk statutory (politics) Ganswijk der Nederlanden. Deel 1, eerste stuk 2. Handleiding tot de kennis van het staatsbestuur in het Koninkrijk statutory (politics) der Nederlanden. Deel 1, tweede stuk

200 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 61 ND sources – Southern Dutch (part 1)

Author name Text name Text type Francis J. Blieck Mengelpoezy rhymed/lyrical Philippe Blommaert 1. Aenmerkingen over de verwaerloozing der Nederduitsche tael' secular (linguistics) 2. Iwein van Aelst (1128); Het blijspel van Mevrouw de Caumartin narrative 3. Oudvlaemsche gedichten der XIIe, XIIIe en XIVe eeuwen (3 delen) rhymed/lyrical Caspar H. van Boekel Arnold van Rummen narrative Adrianus Bogaers Verhandeling over het wezen der uiterlijke welsprekendheid, hare secular (argumentation) voordeelen, en de meest geschikte middelen ter bevordering van hare beoefening hier te lande Frans Bon Vlaemsche spraekkonst volgens het Belgisch taelstelstel secular (information) Pieter J. de Borchgrave, Verzameling der dichtwerken over de Belgen rhymed/lyrical e.a. Jacob L. de Clercq Baudewyn en Avezoeta, of Vryheydsliefde en godsdienst narrative Hendrik Conscience 1. Avondstonden narrative 2. Baes Gansendonck narrative 3. De leeuw van Vlaenderen narrative 4. Geschiedenis van België secular (history) 5. In 't Wonderjaer narrative 6. Phantazy narrative Jan Baptist David 1. Nederduytsche spraekkunst (2 delen) secular (information) 2. Vaderlandsche historie. Deel 2 secular (history) Maria Doolaeghe 1. De avondlamp rhymed/lyrical 2. Palfyn rhymed/lyrical Prudens van Duyse 1. De spellingsoorlog rhymed/lyrical 2. De wanorde en omwenteling op den Vlaemschen zangberg rhymed/lyrical 3. Gedichtjes voor kinderen rhymed/lyrical 4. Godfried of De godsdienst op 't veld rhymed/lyrical 5. Natalia rhymed/lyrical 6. Prijsverzen op de dood van Egmont rhymed/lyrical 7. Prys-gedichten, ter beantwoording der vraeg: 'Wat de schouwburg rhymed/lyrical vermag op de beschaving der volkeren?' 8. Vaderlandsche poëzy (3 delen) rhymed/lyrical Prudens van Duyse & Jan Willem Tell rhymed/lyrical V. Wouters Pieter Ecrevisse 1. De bokkenryders in het land van Valkenberg narrative 2. De Drossaert Clercx, eene omwerking van De Teuten in de narrative Limburger Kempen Pieter Ecrevisse & Eugeen Simon Cokkermoes; Egmont's einde narrative Zetternam André van Hasselt Het dorp der goudmakers narrative Augustijn d’Huygelaere Nationale poëzy rhymed/lyrical Jacob Kats Het aerdsch paradys, of den zegeprael der broederliefde dialogue Pieter van Kerckhoven Ziel en lichaem narrative Karel L. Ledeganck 1. Bloemen mijner lente rhymed/lyrical 2. De drie zustersteden rhymed/lyrical 3. De vrede rhymed/lyrical 4. De zinnelooze rhymed/lyrical 5. Het burgslot van Zomergem rhymed/lyrical 6. Zegeprael van 's lands onafhankelykheid. Lotsbestemming des rhymed/lyrical vaderlands Hippoliet van Peene 1. Brigitta dialogue 2. Het likteeken dialogue 3. Jacob van Artevelde dialogue

APPENDIX 1 201

Table 61 ND sources – Southern Dutch (part 2)

Author name Text name Text type Hippoliet van Peene 4. Klaes Kapoen dialogue 5. Roosje zonder doornen dialogue 6. Thyl Uilenspiegel, of De gefopte bruidegoms dialogue 7. Wit en zwart dialogue Joseph Ronsse 1. Arnold van Schoorisse, episode uit den opstand der Gentenaers narrative (1382-1385) (4 delen) 2. Pedro en Blondina narrative Emmanuel Rosseels De verfranschte landmeisjes dialogue Joseph Sadones 1. Klugtigen voorval rhymed/lyrical 2. Lof-gezang der boeren rhymed/lyrical Ferdinand A. Snellaert 1. Schets eener geschiedenis der Nederlandsche letterkunde secular (information) 2. Verhandeling over de Nederlandsche dichtkunst in België, sedert secular (linguistics) hare eerste opkomst tot aen de dood van Albert en Isabella August Snieders Myne eerste zangen rhymed/lyrical Jan B.C. Verlooy Verhandeling op d'onacht der moederlyke tael in de Nederlanden secular (argumentation) Jozef van Walleghem Merckenweerdigste voorvallen en daegelijcksche gevallen. 13 delen. secular (history) Jan F. Willems 1. Aen de Belgen. Aux Belges rhymed/lyrical 2. Belgisch museum voor de Nederduitsche tael- en letterkunde en de secular (information) geschiedenis des vaderlands 3. Oude Vlaemsche liederen rhymed/lyrical 4. Verhandeling over de Nederduytsche tael- en letterkunde, secular (argumentation) opzigtelyk de Zuydelyke provintien der Nederlanden 5. Voorbericht' secular (linguistics) Eugeen Zetternam 1. Bernhart de Laet narrative 2. Margaretha van Constantinopelen dialogue 3. Mijnheer Luchtervelde narrative 4. Rowna. Eene fantastische legende narrative

202 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 62 PDDW sources – Northern Dutch (part 1)

Author name Text name Text type Peter Andriesse, e.a. Manifest voor de jaren zeventig secular (argumentation) Konrad Boehmer & Van Provo naar Oranje Vrijstaat secular (argumentation) Ton Regtien Jan H. van den Berg Metabletica of leer der veranderingen secular (psychology) Wouter Blok Verhaal en lezer secular (linguistics) Petrus J. Buijnsters Hieronymus van Alphen (1746-1803) secular (information) Frederik J.J. Buytendijk De vrouw secular (lifestyle) Anton L. Constandse Anarchisme van de daad statutory (politics) Alphons Diepenbrock Verzamelde geschriften secular (argumentation) Eduard J. Dijksterhuis De mechanisering van het wereldbeeld secular (science) Jacobus A.A. van Doorn Gevangen in de tijd statutory (politics) Jacobus A.A. van Doorn & De stagnerende verzorgingsstaat secular (argumentation) Cornelis J.M. Schuyt Willem Drees Drees aan het woord secular (argumentation) Redbad L.K. Fokkema Het komplot der Vijftigers secular (linguistics) Pim Fortuyn De puinhopen van acht jaar Paars statutory (politics) Pieter A.M. Geurts De Nederlandse Opstand in de pamfletten 1566-1584 secular (argumentation) Jan Greshoff Afscheid van Europa personal (autobiography) Gerrit Groenhuis De predikanten religious (theology) Jan P. Guépin De beschaving secular (argumentation) Hella S. Haasse Zelfportret als legkaart personal (autobiography) Maarten 't Hart De stekelbaars secular (science) Arnold Heertje De kern van de economie secular (science) Jérôme L. Heldring Heel ons fundament kraakt en andere kanttekeningen secular (argumentation) Willem F. Hermans Erosie secular (science) Johan Huizinga Verzamelde werken deel 7. Geschiedwetenschap Hedendaagsche secular (history) cultuur Willem M.H Hummelen De sinnekens in het rederijkersdrama secular (linguistics) Loe de Jong De Duitse Vijfde Colonne in de Tweede Wereldoorlog secular (history) Gerrit Komrij De stankbel van de Nieuwezijds secular (argumentation) Ernst H. Kossmann 1. Familiearchief personal (autobiography) 2. Politieke theorie en geschiedenis statutory (politics) Arthur Lehning H. Marsman, de vriend van mijn jeugd personal (autobiography) Ton Lemaire Filosofie van het landschap secular (philosophy) Arend Lijphart Verzuiling, pacificatie en kentering in de Nederlandse politiek statutory (politics) Johannes Linschoten Idolen van de psycholoog secular (psychology) Aleida Loosjes-Terpstra Moderne kunst in Nederland 1900-1914 secular (science) Mathias Matthijssen Klasse-onderwijs secular (pedagogy) Philip Mechanicus In dépôt personal (diary) Merlyn (Tijdschrift) Merlyn secular (information) Ileen Montijn Leven op stand 1890-1940 secular (lifestyle) Rob Nieuwenhuys Oost-Indische spiegel secular (information) Hans U.J. d'Oliveira Vondsten en bevindingen. Essays over Nederlandse poëzie secular (argumentation) Jacob J. Oversteegen Vorm of vent secular (linguistics) Michel van der Plas Uit het rijke Roomsche leven secular (history) Judith Pollmann Een andere weg naar God religious (theology) Jacques Presser Ondergang secular (history) Paul Rodenko 1. De sprong van Münchhausen secular (argumentation) 2. Tussen de regels secular (argumentation) Bernard V.A. Röling Polemologie. Een inleiding tot de wetenschap van oorlog en vrede statutory (politics) Piet de Rooy Republiek van rivaliteiten. Nederland sinds 1813 secular (history)

APPENDIX 1 203

Table 62 PDDW sources – Northern Dutch (part 2)

Author name Text name Text type Piet de Rooy & Henk te Met Kok statutory (politics) Velde Ferdinand Sassen Geschiedenis van de wijsbegeerte in Nederland tot het einde der secular (philosophy) negentiende eeuw Ard Schenk Ard apart. Ard schrijft over schaatsen personal (autobiography) Ivo Schöffer Het nationaal-socialistische beeld van de geschiedenis der secular (history) Nederlanden August L. Sötemann De structuur van Max Havelaar secular (linguistics) Ricky W. Stutgard De eerste Surinaamse sportencyclopedie (1893-1988) secular (information) Ed van Thijn Dagboek van een onderhandelaar personal (diary) Jan Tinbergen Een leefbare aarde secular (science) Maarten van den Toorn Nederlandse grammatica secular (linguistics) Henk te Velde Stijlen van leiderschap. Persoon en politiek van Thorbecke tot Den Uyl statutory (politics) Isaac van der Velde De tragedie der werkwoordsvormen secular (linguistics) Victor E. van Vriesland Onderzoek en vertoog secular (argumentation) Leo Vroman Bloed secular (science) Henri van de Waal Drie eeuwen vaderlandsche geschied-uitbeelding, 1500-1800 secular (history) Wilhelmina, prinses der Eenzaam maar niet alleen personal (autobiography) Nederlanden Cor H.E. de Wit De Nederlandse revolutie van de achttiende eeuw 1780-1787 secular (history) Jan J. Woltjer Tussen vrijheidsstrijd en burgeroorlog secular (history)

204 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 63 PDDW sources – Southern Dutch (part 1)

Author name Text name Text type Johan Anthierens 1. GAL: De overspannen jaren secular (history) 2. Het Belgische domdenken secular (argumentation) Geert Bekaert Verzamelde opstellen (2 delen) secular (argumentation) Rob Belemans & Ronny Belgisch-Limburgs secular (information) Keulen Amand Berteloot Bijdrage tot een klankatlas van het dertiende-eeuwse secular (linguistics) Middelnederlands Marnix Beyen Held voor alle werk. De vele gedaanten van Tijl Uilenspiegel secular (history) Willem P. Blockmans & Prinsen en poorters secular (history) Walter Prevenier Willy L. Braekman Middeleeuwse witte en zwarte magie in het Nederlands taalgebied secular (argumentation) Renaat Braem 1. Het lelijkste land ter wereld secular (argumentation) 2. Het schoonste land ter wereld personal (autobiography) Hugo Brems De dichter is een koe secular (argumentation) Frans van Cauwelaert Gedenkschriften over Vlaamse Beweging en Belgische politiek personal (autobiography) Manu Claeys Het Vlaams Blok in elk van ons secular (history) Dirk Coigneau Bedongen creativiteit. Over retoricale productieregeling secular (linguistics) Rik Van Daele, e.a. Het land van Reynaert secular (science)

Andrée Despy-Meyer, 1. Geschiedenis van de wetenschappen in België van de Oudheid tot secular (science) 1815

e.a. 2. Geschiedenis van de wetenschappen in België. 1815 - 2000. secular (science) Magda Devos & Rein-hild West-Vlaams secular (information) Vandekerckhove Steven Dhondt & Onverwerkt verleden secular (history) Luc Huyse Hendrik J. Elias Vijfentwintig jaar Vlaamse Beweging statutory (politics) Antonin van Elslander Het refrein in de Nederlanden tot 1600 secular (linguistics) Gaston Eyskens De memoires personal (autobiography) Joos Florquin Ten huize van… (18 delen) secular (interview) Rita Ghesquiere Waar komen die engelen vandaan? Filosofie en kinderliteratuur. secular (philosophy) Paul Hadermann Het vuur in de verte secular (linguistics) Theo D'haen Europa buitengaats (2 delen) secular (argumentation) Luc Huyse 1. De gewapende vrede. Politiek in België na 1945 statutory (politics) 2. Passiviteit, pacificatie en verzuiling in de Belgische politiek secular (science) Karel van Isacker Mijn land in de kering 1830-1980. Deel 1: Een ouderwetse wereld secular (history) 1830-1914 Ria Jansen-Sieben De Pseudo-Hippokratische Iatromathematika in vier secular (science) Middelnederlandse versies Rudi Laermans Ruimten van cultuur secular (lifestyle) Paul Louyet België in de Tweede Wereldoorlog (deel 1-6, 8-10) secular (history) Angèle Manteau Ja, maar mevrouw, deze schrijven Nederlands personal (autobiography) Guido Marnef Antwerpen in de tijd van de Reformatie secular (history) Patricia de Martelaere Een verlangen naar ontroostbaarheid secular (argumentation) Wilfried Martens De memoires personal (autobiography) Lut Missinne Kunst en leven, een wankel evenwicht secular (linguistics) Jan Nuyts & Georges De Stadsantwerps secular (information) Schutter Paul van Ostaijen Verzameld werk. Deel 4: proza personal (contemplative) Joris van Parys Masereel secular (information) Piet Piryns Er is nog zoveel ongezegd. Vraaggesprekken met schrijvers. secular (interview) Karel Porteman De mystieke lyriek van Lucas van Mechelen (1595/96 - 1652) secular (information) Willem M. Roggeman Beroepsgeheim secular (interview)

APPENDIX 1 205

Table 63 PDDW sources – Southern Dutch (part 2)

Author name Text name Text type Mathieu Rutten & Jean Van Arm Vlaanderen tot De Voorstad groeit secular (linguistics) Weisgerber Manu Ruys De Vlamingen statutory (politics) Edward Schillebeeckx Jezus, het verhaal van een levende religious (theology) Pieter De Somer Een visie op de universiteit secular (argumentation) Leo Tindemans De memoires personal (autobiography) Jo Tollebeek De toga van Fruin secular (history) Frank vande Veire De geplooide voorstelling secular (argumentation) Adriaan Verhulst Zoon van een 'foute' Vlaming personal (autobiography) Bart Verschaffel Figuren / Essays secular (argumentation) Raymond Vervliet De literaire manifesten van het fin de siècle in de Zuidnederlandse secular (linguistics) periodieken 1878-1914 Freddy de Vree Rita Renoir, enz. secular (argumentation) Jean Weisgerber Aspecten van de Vlaamse roman, 1927-1960 secular (linguistics) Albert Westerlinck Gesprekken met Walschap (2 delen) secular (interview) Bruno de Wever Greep naar de macht secular (history) Lode Wils Flamenpolitik en aktivisme statutory (politics) Paul de Wispelaere Met kritisch oog secular (argumentation)

206 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Google books project of the University of Ghent

Table 64 END sources – Northern Dutch

Author name Text name Text type Marcus Z. van Boxhorn Bedenckingen aengaende de successie, ende het recht der oudst- statutory (administration) geborene in het aenveerden van een erfrijck Petrus Cupus Ondersoeck over D'Onschult ende Afwijsinge Adriani Smoutii over 't secular (information) ghene hem van weghen de t' onrecht-verdachte predikanten toe- ghewesen is, in hun Aenwijsinge van de onbehoorlijcke wijse van doen, [...] onder den name P. Cupi ende de letteren C. S. a. G. binnens maendts wtghegheven Vincent Van Drielenburch Een helder licht, daer in klaerlick ghesien wort, dat d'Arminianen niet religious (theology) alleen de oorsaecke zijn van de hedendaechse twisten der kercken ende landen, […] Maer datse oock arbeyden (door liegen, lasteren, ende valsch beschuldigen) om de oude ware gereformeerde religie, ende alle liefhebbers derselver uyt te roeyen, door Inquisitie, ofte Vervolginge [...] Paschier de Fijne Vermeerdert Kamper-steurtje van harde Eyeren, stercke Boter, ende secular (lifestyle) bittere Mostaert. Toegemaeckt van eenen Hollandtschen Koopman, ende eenen Camperschen Schipper [...] Gedruckt voor die Campers Adrianus Heereboord Brief van Adrianus Heereboord ghesonden aen sijn Vrouw Johanna de personal (letter) la Cour Christiaen van Heule De Nederduytsche spraec-konst, ofte Tael-beschryvinghe secular (linguistics) Nicolaes Tulp Drie boecken der medicijnsche aenmerkingen secular (science) Gisbertus Voetius Grondige ende pertinente Verklaringe over de Vrage Wien de Kerke- religious (theology) lijcke Macht toekomt, In drie onderscheydene Disputatien voor- ghestelt van […] D. Gysbertus Voetius

APPENDIX 1 207

Table 65 END sources – Southern Dutch

Author name Text name Text type Antonius Anselmo Codex Belgicus seu Ius edictale a principibus Belgarum sancitum ofte statutory (administration) De Nederlandtsche, nieuwe soo gheestelycke, als wereltlycke rechten, ghetrocken uyt de vier placcaet-boecken, tot Gendt, ende Antwerpen, uytgegeven. Midtsgaders uyt noch eenighe andere placcaeten, edicten, en ordonnantien, die inde voorschreven vier placcaet- boecken niet bekent, ende over-sulcx hier achter aenghevoeght sijn Judocus Van Assche Den schat des heylighen scapuliers inhoudende den oorspronck, religious (theology) voordeelen, ende aflaten van het broederschap, der suyvere maget ende moeder Gods Maria des berghs Carmeli Carolus Borromaeus & Corte verclaringhe van t'groot Jubilee. Met een salich onderwijs om personal (letter) Michael Ophovius t'selve profytelijck te ghenieten. Ghenomen vvt eenen sendtbrief vanden E. cardinael Borromevs. Waer by ghevoecht zijn sommige ghebedekens ende vermaninge, dienende ten selven eynde. Door den Eervv. P. Michael Ophoven... Hendrik Van Brederode Propositie ende Requeste opt stuck vande Inquisitie, ghedaen ende statutory (politics) overgegeven aen Mijn Vrouwe dHertoginne van Parme ende Plaisance, &c. Bijden ... Heer Hendrick, heer tot Brederoede ... opten V.sten april M.D.Lxvi Franciscus Costerus Een cort tractaet vande af-laeten van nieuws ghedruckt ter secular (argumentation) oorsaecken vanden jubilee ocerghesonden in dese Neder-landen van [...] den paus Urbanus den VIII. Joost de Damhoudere Practycke ende handbouck in criminele zaeken, verchiert met statutory (administration) zommeghe schoone figuren en beilde ter materie dienende ... Arnout van Geluwe Kort verhael van een achthien-jarighe Hollandtsche reyse: ghewandelt personal (travel story) van eenen vlaemschen boer ... met naeme Arnout van Geluwe Joannes Hauchinus Clare belijdinghe des Christen Gheloofs, twelck alle oprechte religious (theology) Catholijcke menschen tot allen tijden daertoe versocht zijnde, schuldich zijn te belyden Jacob de Keysere Sekere remedien voor de ellenden van desen tijt. Gherepresenteert in religious (theology) een voorbeeldinge vande waerachtige bekeeringhe ende penitentie van die van Niniue Michael F. Van Langren Profytelycken middel om met in-dyckinghe van landt, de zee-haven secular (science) van Oostende te verbeteren, met een klare demonstratie, denende om te bewysen, datter water in de zee is, dat sich niet en beweeght door ebbe oofte vloedt Justus Lipsius, e.a. Die heylighe maghet van Halle, door Iustus Lipsius: hare weldaden secular (information) ende mirakelen oordententlijck ende ghetrouwelijck beschreven Ghileyn Manilius & Jan Ordinantie ghemaect by de[n] Raedt van Vlaendre[n] vpde maniere statutory (politics) Van den Steene van procederen, voor hemlieden, ende de continuatien van[den]… Gualterus Manilius continuatie vanden cours vander munte ... secular (science) Jan Van der Moten Cort ondervvys vande letter-const secular (pedagogy) Jan Pelsers Van de Peste. Een generale methodus, om te cureren die contagieuse secular (science) zieckte der pestilentiaele cortse met haer sympthomata. Ghecopuleert duer Ian Peisers ... Adriana Teypins Declaratie ... ende causen om welcker wille de scepenen ende statutory (politics) wethauders der stede van Gent ... Cornelius C. Vrancx Van eenen zeer machtighe[n] coopma[n], die alle iaere inden vastene secular (information) voor Paesschen is ouer commende, ende zijn goed voort doende openbaerlick voor een yeghelick [...] in die stadt van Ghendt Philips Wielant Tractaet van den Leenrechten nae de Hoven van Vlaanderen, secular (argumentation) metsgaders de diensten daertoe staende Olivier de Wree De seghelen der graven van Vlaendren ende voor-schriften van hunne secular (history) brieven

208 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 66 ND sources – Northern Dutch (part 1)

Author name Text name Text type Abraham J. Van der Aa Nederland: handboekje voor reizigers door ons vaderland secular (lifestyle) Henricus Aeneae Antwoorden op de rekenkundige vraagen secular (science) Hiëronymus van 1. De gronden mijner geloofs-belijdenis, opengelegd voor mijne religious (theology) kinderen Alphen 2. Verhandeling over de kenmerken van waar en valsch vernuft, als secular (argumentation) ook over de behoedmiddelen tegen het laatste Albertus P. van Amstel Verhandeling over het recht van commercie tusschen onzydige en secular (argumentation) oorlogvoerende volken Cornelis P. van Amstel De bouworde van 't stadhuis van Amsterdam, en de smaak der secular (science) Nederlanderen: ten opzigte der konsten en weetenschappen, verdedigd, tegen de ongegronde berispingen van den heer C. Nicolaas Anslijn Nz. Kruidkundig leerboek, inzonderheid voor hen, die tot de secular (science) artsenijmengkunde worden opgeleid Obbe S. Bangma Inleiding tot de algebra ten dienste der scholen: behelzende eene secular (science) korte verklaring van de eerste beginselen dezer wetenschap… Abraham L. Barbaz 1. De logenaar: blyspel rhymed/lyrical 2. Johanna, Koninginne van Napels: treurspel rhymed/lyrical 3. Makin of De ontdekking van Madera: tooneelspel dialogue Johannes Barueth Historie van het stadhouderschap der heeren princen van Orange, secular (history) hoognodig tot bewaring van de vryheid in den kerk- en burgerstaat Martinus N. Beets Volks-scheikunde, of onderwijzingen en raadgevingen tot nuttig secular (science) gebruik Nicolaas Beets Des Christens schuld aan den heiden religious (theology) Jacobus Bellamy Proeven voor het verstand, den smaak en het hart narrative Johannes In de Betouw Bijvoegsel tot de Annales en Chronyk van Nijmegen uit de secular (history) rekenboeken en Guedesdagboeken Abraham Biben Zondagsboek voor de huisgezinnen der christenen religious (theology) Katharina W. Bilderdijk Wapenkreet voor Neêrlands volk rhymed/lyrical Willem Bilderdijk 1. Aan Leyden, op den twaalfden van loumaand rhymed/lyrical 2. Brieven van Mr. Willem Bilderdijk, aan Joannes Franciscus Willems personal (letter) 3. Elius: romance rhymed/lyrical 4. Geslachtlijst der Nederduitsche naamwoorden, op stellige secular (linguistics) taalgronden gevestigd 5. Mijn verlustiging rhymed/lyrical 6. Nederlandsche spraakleer secular (linguistics) 7. Opstellen van godsgeleerden en zedekundigen inhoud religious (theology) 8. Taal- en dichtkundige verscheidenheden secular (linguistics) 9. Van het letterschrift secular (linguistics) 10. Verhandelingen, ziel-, zede-, en rechtsleer betreffende secular (argumentation) 11. Verhandeling over het verband van de dichtkunst en secular (argumentation) welsprekendheid met wijsbegeerte Steven Blaupot T. Cate Geschiedenis van Nederlands zeevaart en handel secular (history) Pieter Boddaert Levensgeschiedenis van den vermaarden dichter P. Boddaert, personal (autobiography) benevens zijne poëtische en prozaïsche portefeuille Lambert van Bolhuis & Beknopte aanleiding tot de kennis der spelling, spraakdelen en secular (linguistics) Klaas Stijl zintekenen; van de nederduitsche taal… Andreas Bonn Ontleed- en heelkundig onderzoek der schouderontwrichting, in drie secular (science) lessen: gehouden in het theatrum anatomicum der stad Amsterdam Gisbert Bonnet Twee redevoeringen van Gisb. Bonnet secular (argumentation) Herman Bosscha Geschiedenis der Nederlandsche staats-omwenteling in secular (history) achttienhondert dertien

APPENDIX 1 209

Table 66 ND sources – Northern Dutch (part 2)

Author name Text name Text type Gerrit Brender à Brandis Taal-, dichten letterkundig magazijn: of Verzameling van secular (linguistics) verhandelingen, de taal-, dichten letterkunde betreffende; benevens eenige Dichtstukken… Willem G. Brill Hollandsche spraakleer secular (linguistics) Berend Brugsma Opmerkingen en wenken betrekkelijk onderwijs en tucht, secular (pedagogy) inzonderheid in de lagere scholen Willem F. Büchner Geneeskundig handboek voor beginnende kunstoefenaren secular (science) Johannes Buijs 1. Natuurkundig schoolboek secular (science) 2. Volks-natuurkunde, of onderwijs in de natuurkunde voor secular (science) mingeoefenden, tot wering van wanbegrippen, vooroordeel en bijgeloof Caspar Burmannus Utrechtsche jaarboeken van de vyftiende eeuw, vervattende het secular (information) merkwaardige in het gesticht, en voornamentlyk in de stadt Utrecht Petrus Camper Gerechtelyke en ontleedkundige verhandeling over de teekenen van secular (argumentation) leven, en dood in nieuwgeborene kinderen Godert v. d. Capellen Rapport aan Zijne Majesteit den Koning van Holland statutory (politics) Jan B. Christemeijer Oorkonden uit de gedenkschriften van het strafregt; en uit die der statutory (administration) menschelijke misstappen: tot een vervolg op de belangrijke tafereelen uit de geschiedenis der lijfstraffelijke regtspleging, enz Isaäc Da Costa Rouw en trouw rhymed/lyrical Evert Jan Diest Lorgion Geschiedenis van de Kerkhervorming in Friesland secular (history) Jacob Dirks Geschiedkundig onderzoek van den koophandel der Friezen van de secular (history) vroegste tijden tot aan den dood van Karel den Grooten Petrus Driessen Scheikundige verhandeling over de magnesia alba, en het nuttig secular (science) gebruik van de moederloogen van zeezout in ons vaderland Frederik Duim Alexander en Artemize: Treurspel rhymed/lyrical Daniël M. van Dussel- Geschiedkundige verhandelingen secular (history) dorp de Superville Wopke Eekhoff Geschiedkundige beschrijving van Leeuwarden, de hoofdstad van secular (history) Fries-land: vermeldende den oorsprong, den aanwas en de uitbreiding van deze stad en van hare openbare gebouwen, gestichten, inrigtingen enz. Henrik Van Elvervelt Het eiland van Verwarring: Blyspel rhymed/lyrical Hendrik C. d'Escury Holland's roem in kunsten en wetenschappen secular (science) Rhijnvis Feith 1. De eenzaamheid en de wereld rhymed/lyrical 2. De opwekking van Lazarus rhymed/lyrical 3. Het graf: in vier zangen rhymed/lyrical 4. Oden en gedichten rhymed/lyrical Daam Fockema Iets betrekkelijk de afscheiding van het Zuiden en Noorden van de secular (history) Nederlanden Joannes Le Francq van 1. De zeetriumph der Bataafsche vryheid op Doggersbank: bevochten rhymed/lyrical Berkhey den 5den van oogstmaand 1781 2. Gedichten van Joannes Le Francq van Berkhey rhymed/lyrical 3. Het huwelyk van Telemachus en Antiope, in Ithaka: in drie bedryven rhymed/lyrical Jacob de Gelder 1. Allereerste gronden der cijferkunst: bevattende de verklaring van secular (science) het tientallige stelsel van tellen, het betoog der vier grondregels, de behandeling der gewone en tiendeelige breuken… 2. Allereerste gronden der stelkunst, ten gebruike der Latijnsche secular (linguistics) scholen en andere kollegien David van Gesscher Proeve over de voornaamste langduurige gezwellen secular (science) Matthias van Geuns De heerschende persloop (dysenteria epidemica), die in de laatste secular (science) jaaren, vooral in 1783 de provincie van Gelderland fel getroffen heeft

210 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 66 ND sources – Northern Dutch (part 3)

Author name Text name Text type Paulus K. Görlitz Geschiedkundig overzigt van het lager onderwijs in Nederland secular (pedagogy) Johannes de Gorter Gezuiverde geneeskonst, of kort onderwys der meeste inwendige secular (science) ziekten; ten nutte van chirurgyns…

William Grant & Evert J. Waarnemingen omtrent de natuur en genezing der koortsen secular (science) Thomassen à Thuessink Petrus H. de Groot Is bezuiniging op het Onderwijs, vooral door opheffing eener secular (pedagogy) Hoogeschool aan te raden? Gerard ten Haaff Korte verhandeling door voorbeelden gesterkt, nopens de nieuwe secular (argumentation) wyze om de cataracta, met de daar door veroorzaakte blindheid te genezen, door middel van het Kristallyne Vocht uit het Oog te nemen Maurits C. van Hall Redevoering, ten betooge: dat de geest van orde, moed, secular (argumentation) zelfopoffering- en andere volksdeugden, die de ingezetenen van Oud- Nederland thans zoo loffelijk onderscheiden…

Stephanus Hanewinckel Zedekundig handboek voor den militairen stand secular (lifestyle) Ottho G. Heldring Opmerkingen op eene reis langs den Rijn secular (lifestyle) Wesselius van Hengel Over de godgeleerdheid in het algemeen en hare betrekking tot het secular (pedagogy) onderwijs op 's lands Hoogescholen in het bijzonder Henricus Heppener De oorsprong en het nadeel van het vooroordeel religious (theology) Johan H. van Heurn Historie der stad en meyerye van 's Hertogenbosch alsmede van de secular (history) voornaamste daaden der hertogen van Brabant Nicolaas Hoefnagel Brief van Niekolaas Hoefnagel aan de voorname Noort-Hollandsche personal (letter) heer, of de ontmaskerde Willem Ockers. Jacob H. Hoeufft 1. Taalkundige aanmerkingen op eenige Oud-Friesche spreekwoorden secular (linguistics) 2. Taalkundige bijdragen tot de naams-uitgangen van eenige, meest secular (linguistics) Nederlandsche, plaatsen Petrus Hofstede De Belisarius van den heer Marmontel beoordeeld en de kwade zeden religious (theology) der vermaardste heidenen aangetoond ten bewyze hoe onbedagtsaam men dezelve om hunne deugdsaamheid verhemeld heeft Gijsbrecht K. van 1. Bijdragen tot de huishouding van staat in het Koningrijk der statutory (politics) Hogendorp Nederlanden, verzameld ten dienste der Staten Generaal 2. Verhandelingen over den Oost-Indischen handel secular (argumentation) Arnold Hoogvliet Abraham, de aartsvader, in XII boeken religious (theology) Willem van den Hoonaard Geschiedkundige en topographische beschrijving van de dorpen secular (science) Hillegersberg en Bergschenhoek Martinus Houttuyn Natuurlyke historie of uitvoerige beschryving der dieren, planten en secular (science) mineraalen, volgens het samenstel van den heer Linnaeus Carel S. Hurau Huisselijke godsdienst-oefening, nuttig en noodzakelijk religious (theology) Balthasar Huydecoper Gedichten van Balthazar Huydecoper rhymed/lyrical Daniel H. Jacobszoon Verzameling van advijzen over het Wetboek van Koophandel voor het secular (science) Koningrijk der Nederlanden Arie de Jager Proeve over de werkwoorden van herhaling en during in de secular (linguistics) Nederduitsche taal Leonhardt J.F. Janssen De grieksche, romeinsche en etrurische monumenten van het secular (history) Museum van oudheden te Leyden Johannes C. de Jonge Geschiedenis van het Nederlandsche zeewezen secular (history) Marinus W. de Jonge Pleitredenen in de zaak van den officier van justitie by de regtbank van statutory (administration) eersten aanleg te Haarlem als prokureur des konings,… Nicolaas G. van Kampen 1. Beknopte geschiedenis der letteren en wetenschappen in de Neder- secular (linguistics) landen, van de vroegste tijden af, tot op het begin der negentiende eeuw

APPENDIX 1 211

Table 66 ND sources – Northern Dutch (part 4)

Author name Text name Text type Nicolaas G. van Kampen 2. Gedenkboek van Nederlands moed en trouw, gedurende den secular (history) Belgischen opstand 3. Geschiedenis der Nederlanders buiten Europa, of Verhaal van de secular (history) togten, ontdekkingen, oorlogen, veroveringen en inrigtingen der Nederlanders in Aziën, Afrika, Amerika en Australië. 4. Proeve eener geschiedenis der Kruistogten naar het Oosten tot de secular (history) herovering van Ptolemais in het jaar 1291 5. Redevoering over den geest der Nederlandsche letterkunde, secular (argumentation) vergeleken met die van andere volken 6. Staat- en aardrijkskundige beschrijving van het koningrijk der secular (science) Nederlanden, of der 17 Nederlandsche provinciën, benevens het Groot-Hertogdom Luxemburg 7. Verkorte geschiedenis der Nederlanden, of der XVII. Nederlandsche secular (history) gewesten van de vroegste tijden, tot op den vrede te Parijs in 1815 Johan de Kanter Philz. 1. Lofrede op Jacob Hendrik Schorer, uitgesproken in eene secular (argumentation) buitengewone vergadering van directeuren en leden van het zeeuwsch genootschap der wetenschappen 2. Natuur- en geschiedkundige beschrijving van den watervloed secular (science) tusschen den 14 en 15 Januarij 1808 Johan de Kanter Philz. & De provincie Zeeland secular (science) Johannes Dresselhuis Petrus J. Kasteleijn De tooneelsluiting, met choorzangen en dansen rhymed/lyrical Johannes Kinker Eerstelingen rhymed/lyrical Johannes Kinker & Willem De Post van den Helicon secular (philosophy) Bilderdijk Christiaan Klaarbout Mozes aloude schaduwleer, vertonende de voortreflykheid van Vorst religious (theology) Messias: in vier boeken Jacob Klinkhamer Het leven van Hugo de Groot: getrokken uit de voornaamste secular (information) historieschryvers en dichters, doormengd met onpartydige aanmerkingen, en versierd met twee juiste afbeeldingen van het Koffer, waarin De Groot zyne gevangenis ontkomen is Adriaan Kluit Inwijdingsrede over 't recht, 't welk de Nederlanders gehad hebben, statutory (administration) om hunnen wettigen vorst en heer Philips, koning van Spanje… Hendrik P. Kluit 1. De hervorming der policie in Nederland statutory (administration) 2. De zelfstandigheid der policie verdedigd secular (argumentation) Jacobus Kok Amsteldamsche jaarboeken, behelzende de merkwaardigste secular (information) geschiedenissen; welken, zo binnen de stad als daaromtrent, zijn voorgevallen, sedert haare eerste beginselen, tot op den tegenwoordigen tijd. Jacobus de Bruyn Kops Beginselen van staathuishoudkunde statutory (politics) Jan Kops Staat van den landbouw in de Vereenigde Nederlanden gedurende secular (science) den jaare 1814 Cornelis J. von Krayenhoff De aloude metzelwerken, vergeleken tegen de hedendaagsche of secular (science) Vertoog waarom de zwaare muuren van dezen tyd, krachteloos en bouwvallig zyn, in tegenoverstelling van die der ouden. Cornelis R.T. Krayenhoff Proeve van een ontwerp tot sluiting van de rivier den Neder-Rhijn en secular (science) Leck en het storten van derzelver water op den ijssel Pieter Langendijk 1. De gedichten van Pieter Langendyk rhymed/lyrical 2. Het wederzyds huwelyks bedrog: blyspel rhymed/lyrical Juliana C. de Lannoy Dichtkundige werken, van Juliana Cornelia baronesse de Lannoy rhymed/lyrical Ulrich G. Lauts Invloed van de Fransche staatsomwenteling van het jaar 1789, op de secular (history) lotgevallen van Nederland

212 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 66 ND sources – Northern Dutch (part 5)

Author name Text name Text type Conradus Leemans Romeinsche oudheden te Maastricht secular (history) Jacob Van Lennep De voornaamste geschiedenissen van Noord-Nederland secular (history) Gheraert van Lienhout & Sterre- en natuurkundig onderwijs, gemeenlijk genoemd: natuurkunde secular (science) Johannes Clarisse van het geheel-al, en gehouden voor het werk van zekeren broeder Gheraert Barthold H. Lulofs 1. De declamatie of de kunst van declaméren of recitéren en van de secular (lifestyle) mondelijke voordragt of uiterlijke welsprekendheid in het algemeen 2. Kort overzigt van de geschiedenis der Nederlanden, met name der secular (history) Noord-Nederlanden 3. Schets van een overzigt der Duitsche taal of der Germaansche secular (linguistics) taaltakken, in derzelfder oorsprong en tegenwoordige verdeeling in het hoogduitsch, nederlandsch, deensch, zweedsch, engelsch, en andere soortgelijke verwantschapte talen en tongvallen 4. Over Nederlandsche spraakkunst, stijl, en letterkennis, als secular (linguistics) voorbereiding voor de redekunst, of welsprekendheidsleer in hare hoogere beteekenis. Jean S. Magnin De voormalige kloosters in Drenthe: geschiedkundig beschouwd secular (history) Louis S. Mercier & De deserteur: tooneelspel rhymed/lyrical Bartholomeus Ruloffs Lucretia W. van Merken 1. Het beleg der stad Leyden: treurspel rhymed/lyrical 2. Maria van Bourgondiën, gravinne van Holland: treurspel rhymed/lyrical Frans van Mieris Verhandeling over het saamenstellen der historien of het beschryven secular (argumentation) der geschiedenissen, inzonderheid die van Holland Petronella Moens 1. De geschiedenis der menschheid rhymed/lyrical 2. Johan van Oldenbarneveld secular (information) 3. Uitboezeming aan mijn geliefd vaderland en de Noord-Nederlanders rhymed/lyrical Gerardus J. Mulder 1. De weg der wetenschap, zinen leerlingen op nieuw aanbevolen secular (science) 2. Wetenschap en Volksgeluk: een woord voor Nederland geschreven secular (lifestyle) Samuel I. Mulder Kort overzigt van de geschiedenis der nederlandsche letterkunde, naar secular (linguistics) aanleiding van het grootere werk over dit ontwerp van den hoogleeraar Matthijs Siegenbeek Martinus Nieuwenhuyzen Leeslesjens behoorende by de verhandeling over het kunstmatig secular (argumentation) leezen Pieter Nieuwland Gedichten van Pieter Nieuwland rhymed/lyrical Pieter van Noemer Stichtelyke gedichten rhymed/lyrical Johannes Nomsz 1. Bartholomeus Las Casas: treurspel dialogue 2. De Nederlandsche dichtkundige schouwburg rhymed/lyrical Anthonius G. Van De operatieve heelkunde secular (science) Onsenoort Cornelis W. Opzoomer 1. Aanteekening op de wet, houdende algemeene bepalingen der statutory (administration) wetgeving van het koningrijk 2. De belangen van het hooger onderwijs: memorie aan zijne majesteit secular (pedagogy) den koning, naar aanleiding van het ministeriëele rapport van 13 november 1848 3. De hervorming onzer hoogescholen. Rapport, wetsontwerp, en secular (pedagogy) memorie van toelichting 4. De twijfel des tijds, de wegwijzer der toekomst. Eene voorlezing, ter secular (pedagogy) opening der akademische lessen uitgesproken Jan G. Ottema Kores, historisch chronologisch onderzoek naar den tijd der secular (history) Babylonische ballingschap Gerrit Paape 1. Blijspelen van Gerrit Paape rhymed/lyrical

APPENDIX 1 213

Table 66 ND sources – Northern Dutch (part 6)

Author name Text name Text type Gerrit Paape 2. De vlugt van Willem den Vyfden of De zegepraal der Bataafsche rhymed/lyrical vryheid: klugtige opera in twee bedryven Zacharius Paspoort Beschrijving van Zeeland vervolgd: zijnde het negende en tiende deel secular (science) van den tegenwoordigen staat der Vereenigde Nederlanden Jan J. Pennink Gezondheidsleer voor het volk secular (lifestyle) Willem E. de Perponcher Eerste brief van den heer W.E. de Perponcher aan den heer H. van personal (letter) Alphen Caspar Philips Wis-, meet- en doorzicht-kundige handleiding, volgens welke men ten secular (science) allen tyde en plaatse, den stand der zonne en maane; de verlichting der voorwerpen door dezelven; de strekking en lengten van derzelver slagschaduwen, gemaklyk vinden, en in perspectiefsche regelen overbrengen en bepaalen kan... Everhardus J. Potgieter Het noorden in omtrekken en tafereelen secular (science) Pieter-J. Prinsen Pestalozzi's leerwijze in de kennis der getallen secular (science) Pieter Pypers De karavaan van groot Kairo: zangspel, met balletten rhymed/lyrical Willem L.F.C. van Rappard Gedachten over eenige onderwerpen, betrekking hebbende tot de statutory (administration) aanstaande Nederlandsche wetgeving Joachim Rendorp Verhandeling over het recht van de jagt secular (argumentation) Arnoldus Rotterdam Gods weg met Nederland of Vervolg op Blomherts geschiedenissen religious (theology) van het Vereenigde Nederland Joachim Le Sage Ten Brief aan W. Da Costa personal (letter) Broek Eduard Sandifort Beschrijving en afbeelding van eene aangebooren liesbreuk secular (science) Jacobus H. Van der Proeve van vergelijking tusschen Nederland als gemeenebest in 1743 statutory (politics) Schaaff en Nederland als koningrijk in 1843 Jacobus Scheltema 1. De laatste veldtogt van Napoleon Buonaparte secular (history) 2. De uitrusting en ondergang der onoverwinnelijke vloot van Philips II, secular (history) koning van Spanje, in 1588 3. Levensschets van Simon Styl secular (information) 4. Peter de Groote, keizer van Rusland, in Holland en te Zaandam in secular (history) 1697 en 1717 5. Redevoering over de brieven van Pieter Corneliszoon Hooft secular (argumentation) 6. Volksgebruiken der Nederlanders bij het vrijen en trouwen secular (lifestyle) Johannes M. Schrant 1. Gezondheids-lessen en regelen voor den kinderlijken leeftijd secular (lifestyle) 2. Leven van Jezus Christus, een geschenk aan de jeugd religious (theology) 3. Verhandelingen over den Bijbel secular (argumentation) Willem Sewel Nederduytsche spraakkonst, waarin de gronden der Hollandsche taale secular (linguistics) naauwkeuriglyk opgedolven, en zelfs voor geringe verstanden, zo ten aanzien der Spellinge als Bewoordinge duidelyk aangeweezen zyn Matthijs Siegenbeek 1. Kort begrip der Verhandeling over de Nederduitsche spelling: secular (linguistics) uitgegeven in naam en op last van het staatsbewind der Bataafsche Republiek 2. Lofrede op den raadpensionaris Simon van Slingelandt secular (argumentation) 3. Geschiedenis der Leidsche hoogeschool, van hare oprigting in den secular (history) jare 1575 tot het jaar 1825 Arend F. Simonsz. Handboek der vaderlandsche historie, vervattende in eene zaakelyke secular (history) en tevens beknopte orde, alle de voornaamste gebeurtenissen, die, van den aanbeginne des lands, tot heden toe in ons vaderland zyn voorgevallen Martinus Slabber Natuurkundige verlustingen, behelzende microscopise waarnemingen secular (science) van in en uit-landse water en land-dieren Wilhelmus Smits Handleiding en schoolboek voor het volks-zangonderwijs: derde deel secular (pedagogy)

214 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 66 ND sources – Northern Dutch (part 7)

Author name Text name Text type Willem A. van Spaen Historie van Gelderland secular (history) Pibo Steenstra Grondbeginsels der meetkunst of kort begrip der ses eerste boeken secular (science) met het elfde en twaalfde van Euclides… Simon Stijl De Mityleners: treurspel rhymed/lyrical Simon Stijl & Johannes Levensbeschryving van eenige voornaame meest Nederlandsche secular (information) Stinstra mannen en vrouwen Leonardus Stocke De kwynende ziektens der Nederlanders, spruitende uit hunne lands- secular (science) gesteltheid en levenswyzen: beschreven tot dienst der landgenooten in 't algemeen, en tot heil der lyders in 't byzonder Arnoldus B. Strabbe Eerste beginselen van de Arithmetica of Rekenkunst, ten gebruike der secular (science) schoolen Martinus Stuart 1. Romeinsche geschiedenissen secular (history) 2. Vaderlandsche historie, vervattende de geschiedenissen der secular (history) Vereenigde Nederlanden, van de vestiging van het erfstadhouderschap in de mannelijke en vrouwelijke linie tot aan 's lands verlossing uit de inlijving in het Fransche Keizerrijk Willem H. Suringar Sluikerij, een valstrik voor de zedelijkheid; en het bederf der secular (lifestyle) maatschappelijke welvaart Justus G. Swaving Roomsche feest- en heilige dagen, of Verbijstering van het menschelijk religious (theology) verstand Johan H. Swildens Grondwettige herstelling van nederlands staatswezen, zo voor het statutory (administration) algemeen bondgenootschap als voor het bestuur van elke byzondere provincie… Johannes L. Terwen Etymologisch handwoordenboek der nederduitsche taal: of proeve secular (linguistics) van een geregeld overzigt van de afstamming der nederduitsche woorden Evert Jan Thomassen à 1. Algemeen overzigt der epidemische ziekte, welke in het jaar 1826 te secular (science) Thuessink Groningen geheerscht heeft 2. Geneeskundige waarnemingen secular (science) 3. Geneeskundige waarnemingen door E. J. Thomassen à Thuessink secular (science) 4. Nader onderzoek omtrent de besmettelijkheid of niet secular (science) besmettelijkheid der gele koorts Johann Thorbecke Aanteekening op de grondwet statutory (politics) Hendrik Tollens Feestzang bij het huwelijk van zijne koninklijke hoogheid den prins van rhymed/lyrical Oranje en hare keizerlijke hoogheid Anna Paulowna Hendrik W. Tydeman Verhandeling over de wetenschappelijke beoefening van het regt in statutory (administration) Nederland, na het uitvaardigen van nieuwe wetboeken in de taal des lands Jacobus A. Uilkens Technologische handboek, of beschrijving van het gebruik, hetwelk de secular (science) mensch van de voortbrengselen der natuur maakt: derde stukje Pieter J. Uylenbroek Kleine dichterlyke handschriften rhymed/lyrical Lodewijk G. Visscher 1. Bijdragen tot de oude letterkunde der Nederlanden secular (linguistics) 2. Bloemlezing uit de beste schriften der Nederlandsche dichters van rhymed/lyrical de 13de tot en met de 18de eeuw 3. Ferguut rhymed/lyrical 4. Iets over Jacob de Coster van Maerlant secular (information) Hans Corn. A. Visser & Archief voor vaderlandsche, en inzonderheid vriesche geschiedenis secular (history) Henricus Amersfoordt Johannes Van Vloten Enkele wenschen omtrent de wijziging van het hooger onderwijs secular (pedagogy) Cornelius J. Vos Genees- en heelkundige verhandeling over het been en spek-gezwel secular (science) Jeronimo De Vries Proeve eener geschiedenis der Nederduitsche dichtkunde secular (linguistics)

APPENDIX 1 215

Table 66 ND sources – Northern Dutch (part 8)

Author name Text name Text type Reinier Vryaarts Reinier Vryaarts openhartige brieven: om te dienen tot opheldering en personal (letter) regte kennis van de vaderlandsche historie Jan Wagenaar 1. Amsterdam, in zyne opkomst, aanwas, geschiedenissen, voorregten, secular (history) koophandel, gebouwen, kerkenstaat, schoolen, schutterye, gilden en regeeringe 2. Schets van het leeven, den aart en het gedrag van den Heere Mr. secular (information) Hermannus Noordkerk, advokaat te Amsterdam 3. Verzameling van historische en politieke tractaaten secular (argumentation) Jan Wagenaar & Petrus A. Vaderlandsche historie, vervattende de geschiedenissen der nu secular (history) Loosjes Vereenigde Nederlanden, inzonderheid die van Holland, van de vroegste tyden af Jan van Walré Willem de Eerste, prins van Oranje: treurspel rhymed/lyrical Willem Te Water Historie der hervormde kerke te Gent, van haeren aenvang tot secular (history) derzelver einde; Mitsgaders een kort verhael der gereformeerde doorluchtige schoole te Gent Antoine Warin Bedenkingen over het muntwezen in het koningrijk der Nederlanden secular (science) Willem H. Warnsinck De heer Roze en zijne vrienden in gesprekken en raadgevingen personal (conversation) Pieter Weiland 1. Kunstwoordenboek, of verklaring van allerhande vreemde woorden, secular (linguistics) benamingen, gezegden en spreekwijzen, die, uit verscheidene talen ontleend, in de zamenleving en in geschriften, betreffende alle vakken van kunsten, wetenschappen en geleerdheid voorkomen. 2. Nederduitsche spraakkunst secular (linguistics) 3. Nederduitsche spraakkunst ten dienste der scholen secular (linguistics) Hendrik Wester Schoolboek over de geschiedenissen van ons vaderland secular (information) Willem Westreenen van Korte schets van den voortgang der boekdrukkunst in Nederland, in de secular (lifestyle) Tiellandt XVde, en haare verdere volmaaking in de XVIde en XVIIde eeuw Adriaan van der Willigen Aanteekeningen op een togtje door een gedeelte van Engeland, in het personal (travel story) jaer 1823 Hendrik de Winter Beredeneerde catalogus van alle de prenten van Nicolaas Berchem… secular (information) Nicolaas S. van Winter Monzongo of De koningklyke slaaf: treurspel rhymed/lyrical Samuel I. Wiselius De tooneelspeelkunst, inzonderheid met betrekking tot het treurspel, secular (linguistics) alsmede het nut en de zedelijke strekking des regelmatigen en beschaafden schouwtooneels. Pieter G. Witsen Apollineum: bijdragen, de nederduitsche taal, dichtkunst, secular (linguistics) Geysbeek welsprekendheid en fraaije letteren betreffende. Thomas White & Evert Verhandeling over de ziekte en ontaarding der watervaten en klieren, secular (science) Jan Thomassen à gewoonlijk struma of scrophula genoemd Thuessink Betje Wolff 1. De natuur is mijne zanggodin: Dichtstuk rhymed/lyrical 2. Walcheren, in vier gezangen rhymed/lyrical Ernst Zeydelaar Néderduitsche spraakkonst: ten dienste der Néderlandsche secular (linguistics) taalbeminnaars

216 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 67 ND sources – Southern Dutch (part 1)

Author name Text name Text type Jan L. van Aelbroeck Waarheid-zoekende redeneringen over den twist, opzigtelijk den secular (science) vrijen graan-handel in het Koninkrijk der Nederlanden, tusschen eenen grond-eigenaar, eenen boer en eenen koopman in vreemde granen Jan L. van Aelbroeck & Werkdadige landbouw-konst der Vlamingen, verhandeld in zes secular (science) Franciscus Kierdorff zamenspraken, tusschen eenen grond-eigenaar en zijnen pachter Louis Anseaume & Den soldaet tooveraer: boertig blyspel in een deel, gemengt met rhymed/lyrical Joannes F. Cammaert sangen Jacob J. Antheunis 1. Constitutionneele acte ofte Grond-wetten der Fransche republiek: statutory (administration) voorgegaen door de verklaering van de regten des mensch ende des burgers 2. Omstandig en nauwkeurig begryp van het crimineel proces... statutory (administration) opgesteld tegens Anna Moyet en haare medeverdagten 3. De onverwagte bruiloft, blyspel in twee bedryven rhymed/lyrical Martin J. De Bast 1. Mag en moet men de misse hooren der priesters, die den eed religious (theology) hebben gedaen volgens de wet van 19. fructidor? 2. Ontdekking van het graf des graven van Egmond: voorafgegaan van secular (history) een kort verslag wegens den aanvang der Nederlandsche onlusten in de zestiende eeuw, onder het bestuur van den hertog Van Alba 3. Redenvoeringe uytgesproken door den eerweirden heer De Bast… secular (argumentation) den 25. november 1789… Patrice A. de Beaucourt 1. Beschryving van den opgank, voortgank en ondergank der statutory (administration) de Noortvelde Brugschen koophandel… 2. Jaer-boeken van den lande van den vryen, zydert zyn eerste secular (information) beginzelen, tot en met den jaere 1784 Pieter Behaegel 1. Grond-beginsels der Latynsche tael secular (linguistics) 2. Nederduytsche spraakkunst secular (linguistics) 3. Verhandeling over de vlaemsche spelkunst: byzonderlyk ingerigt ter secular (linguistics) beslissing van de geschilpunten der taelgeleerde omtrent de spelling Philip M. Blommaert Geschiedenis: Julius Caesar in België (56-54 v.C.) secular (history) Ludovicus Blosius Zoete verzugtingen tot Jesus vol van goddelyke liefde, voor iederen religious (theology) dag van de week Félix Bogaerts De goede oude tyd in België: schets van zeden, gebruiken, levenswyze, secular (lifestyle) stichtingen enz., onzer vaderen voor het Fransch gebied… Pieter-Joost De Vriende-traenen gestort op het graf van Augustinus Baude, lid der rhymed/lyrical Borchgrave eerweerde vaders Augustynen, binnen Brugge, overleden den 28 van wey-maend 1816 François Joseph Van den Verhandeling over de vlaemsche tael, in vergelyking met de secular (argumentation) Bossche hollandsche Petrus Van den Bossche Den Katholyken pedagoge, ofte Christelyken onderwyzer in den secular (pedagogy) catechismus Jacobus Bowens Nauwkeurige beschryving der oude en beroemde zee-stad Oostende, secular (science) gelegen in Oostenryksch Vlaenderen […] tot het jaer 1787 Joannes F. Cammaert 1. Den deserteur, bly-spel, in dry deelen: gemengt met sangen dialogue 2. Den hervonden man: kluchtspel rhymed/lyrical Joseph B. Cannaert 1. Bydragen tot de kennis van het oude strafrecht in Vlaenderen: statutory (administration) verrykt met vele tot dusverre onuitgegevene stukken 2. Iets over het oude strafregt in België, gevolgd door eenige notabele statutory (administration) decisien van vroegere tijden Ludovicus-Albertus 1. Bewegelyk gebed van dankbaerheyd en erkentenisse over de religious (theology) Caytan godelyke weldaeden ten opzichte van ons vaderland, en om van God te verzoeken den Vrede en Eendracht onder de Inwoonders van 't Nederland

APPENDIX 1 217

Table 67 ND sources – Southern Dutch (part 2)

Author name Text name Text type Ludovicus-Albertus 2. Gebeden, die den H. Kerke tegenwoordig stort om den hemelschen religious (theology) Caytan bystand te verwerven, tot het ganschelyk verpletteren der vyanden van de Nederlandsche Religie en Vryheyd Hendrik Conscience 1. Blinde Rosa narrative 2. De loteling narrative Jean-Baptiste Coomans & Geschiedenis van Belgien secular (history) Josse Coomans Jan Baptist Courtmans Handleiding voor het onderwys in het lezen: bepaeldelyk ingerigt om secular (pedagogy) aen kinderen in korten tyd het lezen te leeren, ten gebruike der onderwyzers Charles-François Custis Jaer-boecken der stad Brugge, behelsende de gedenckweerdigste secular (information) geschiedenissen… Emmanuel V. Driessche Natuerkundig onderwys (phisiek) secular (science) Alexis M. Eenens Ontginning der Kempensche heiden door het leger secular (lifestyle) Govard G. van Eersel 1. Den waeragtigen Gods-yver: Ofte de lof-verkondende zang-goddin religious (theology) 2. Herderlyken brief van Syne Doorlugtigste Hoogweerdigheyd G. G. personal (letter) Van Eersel Bisschop van Gend, heere van St-Baefs, Grave Van Everghem ... tot uyt-roeyinge der Ledig-gangers ende lastige bedelaers Edward Van Even Geschiedenis der stad Diest secular (history) Karel Gobinet Onderwys der jeugt in de christelyke godvrugtigheyt, getrokken uyt de secular (pedagogy) H. Schrifture en d'HH. Vaders Fulgentius Hellynckx 1. Christelyke onderwysingen voor de lands en ambagts-lieden ... religious (theology) 2. De onlichamelykheyt ende onsterffelykheyt der redelyke ziele… religious (theology) 3. Sorge der saligheyt religious (theology) Jacques F.J. Heremans Beknopte Nederduitsche spraekleer, ten gebruike der scholen van secular (linguistics) middelbaer onderwys Adrianus Heylen Antwoord van… A. Heylen… op het vraeg-stuk… secular (argumentation) Jan-Baptist J. Hofman Justina, of De onderwerping van Namen aan de gehoorzaamheid van rhymed/lyrical zyne Majesteit den keizer en koning Leopoldus II: burger treurspel Jan-Baptist J. Hofman & Barbaersheyd der Afrikaenen, of Laetsten dicht-stryd van het kunst- rhymed/lyrical Thomas van Loo verdrag, tusschen de maetschapyen van rhetorica, der steden Brugge, Ieper, Kortryk en Oostende François-Dominique Redevoering gehouden op de groote zaele van het stadhuys der stad secular (argumentation) d'Hoop Gend den 9 Juny 1778 ... Jan B. Jacobs & Jan Meyer Kort onderwys hoe dat men de breuken ofte scheursels alsmede den secular (science) voorval der lyfmoeder ende arsdarm kan voorkomen ende genezen Jacob Kats De vyanden van het licht of de tegenwerkingen van den Maetschappy dialogue der Verbroedering: historisch bly- en tooneelspel in 2 bedryven Pieter Van Kerckhoven Ziel en lichaam secular (lifestyle) Jacob-Lodewijk Kesteloot De koepok-inënting, getoetst aan het gezond verstand, in dorps- secular (dialogue) gesprekken Jean-Jacques Lambin Zegeprael behaeld door Maurits van Nassau, prins van Oranje, op het rhymed/lyrical leger der Spanjaerden, onder het bevel van den aertshertog Albert Karel L. Ledeganck Het Burgelyk wetboek statutory (administration) Jean-François Lemaire De meetkunst op de kunsten en ambachten toegepast secular (science) Philippe de Lens Regering der stad Gend. De burgemeesteren en schepenen der stad statutory (politics) Gend. Albert-Louis De Antwoord van Syne Excellentie den Bisschop van Namen, op het religious (theology) Lichtervelde depeche aen haer geaddresseert in date 24 Februarii, door het Gouvernement Ludovicus Meyere Meditatien, op het geheel lyden, ende doodt van Christus religious (theology) Amand Neut Levens-beschryving van Leopoldus, eersten koning der Belgen secular (information)

218 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Table 67 ND sources – Southern Dutch (part 3)

Author name Text name Text type Pierre d'Oudegherst Kronyke of jaer-boeken van Vlaenderen secular (history) Louis J.A. De Potter Grondregelen van verdraagzaamheid, ten gebruike der Belgische secular (philosophy) roomsch-katholijken Joseph-Ignace van Praet Jaer-boek der keyzerlyke ende koninglyke hoofd-gilde van den edelen secular (information) ridder Sint-Joris in den Oudenhove binnen de stad Brugge Adolphe Quetelet & Volks-sterrekunde secular (science) Charles Meerts Jean-Joseph Raepsaet Reden-voeringe op de middelen om binnen Vlaenderen te erstellen de secular (argumentation) schilder en bouw-kund', uytgesproken in de stad Audenaerde… Henricus Schynckele 1. Christelyke zedelessen met godvruchtige oeffeningen voor ider religious (theology) week des jaers 2. Uitvindingen der liefde Godts ten opzigt van den mensch religious (theology) Constant P. Serrure Beoefening der Moedertaal te Leuven: Kamers van Rhetorica secular (linguistics) François-Joseph de Smet Voor de regtbank van eersten aanleg, zitting houdende te statutory (administrative) Dendermonde… Ferdinand A. Snellaert Ferdinand Augustijn Snellaert secular (linguistics) Cornelius M. Spanoghe Het Verlost Nederland, vereerlykt door de lang gewenschte aenkomst statutory (politics) huner Koninglyke Hoogheden, de arts-hertogin Maria Christina, en den Koninglyken Prins Albertus Casimirus ... Jean Stecher De eerste Fransche revolutie secular (history) Johannes F. van Den Nederlandschen verstandigen hovenier, over de XII. maenden van secular (science) Sterbeeck het jaer, beschryvende hoe men op de beste en bequaemste maniere hoven, thuynen, lusthoven en boomgaerden verordineren, bereyden, beplanten en bezaeyen ... Als ook vermeerdert met eenen noodigen en gerieffelyken Medecyn-winkel ... vermeerdert met den Neirstigen biën-houder ... Corneille Stevens Bemerkingen over het besluyt der nieuwe opregting van de personal (letter) Napoleonsche gasthuysnonnen, gevolgd door eenen brief van… Cornelius… Jean-François Thys Memorie of vertoog door Isfridus Thys, canonik van Tongerloo secular (argumentation) Petrus F. Valcke Sermoenen op de sondagen en feestdagen… religious (theology) Jan F. Van de Velde Onderwyzingen om den jubilé wel te verdienen, met eene zeer religious (theology) voordeelige uytlegging en aenwakkering tot eene opregte bekeering, en met schoone gebeden Joannes B.R. van Velde de Kort begryp des levens, van zyne doorlugtige hoogheyd Joannes- personal (autobiography) Melroy Baptista-Robertus baron van Velde, heer van Melroy… Willem F.G. Verhoeven 1. Grond-wet ofte constitutie van Mechelen met de nederduytsche statutory (administration) overzettinge en aen-merkingen door G.F. Verhoeven 2. Historische tyd-en oordeelkundige aenteekeningen, ... ; dienende secular (science) tot antwoord op de vraege, Hoedaenig was den staet van de hand- werken, en van den koophandel in de Nederlanden, ten tyde van de derthienste en veerthienste eeuwe? 3. Proeve van dicht-kunde op de oudtheydt, eer, achtbaerheydt, en rhymed/lyrical voort-gangh der vrye konsten Jan B.C. Verlooy Zyn geloof, vryheid en eygendommen in gevaer? secular (philosophy) Karel A. Vervier 1. Dichtregelen gelezen ter gelegenheid van de eerste bijeenkomst der rhymed/lyrical Maatschappij van Nederlandsche taal en letterkunde te Gend 2. Letteroefening rhymed/lyrical Jean Jacques Philippe Verhael ende reflexien op de verbreydinge der besmettelyke siekte in secular (science) Vilain het hoorn-vee Jan Wagenaar Wagenaar's beschryving van Amsterdam, gevolgd, in eene geregelde secular (science) aanwyzing van de sieraaden der publieke gebouwen dier stad,…

APPENDIX 1 219

Table 67 ND sources – Southern Dutch (part 4)

Author name Text name Text type Jacobus T.J. Wellens 1. Herderlyken brief om aen de inwoonders van Antwerpen te personal (letter) verzoeken aelmoessen en dienstbaerheyd voor de nieuwe bestieringe van den algemeynen armen deze stad 2. Herderlyken brief van zyne doorlustigste hoogweerdigheyd den personal (letter) heere bisschop van Antwerpen nopende de pauselyke Benedictie met vollen aflaet… Jan F. Willems 1. Antwoord van J.F. Willems, aen J.B. Buelens, R.C.P.r te Mechelen, personal (letter) schryver en uytgever van een werk getiteld: Briefwisseling tusschen J.F. Willems, schryver van het werk tael- en letterkundige verhandeling, enz. 2. De puyn-hoopen rondom Antwerpen, of Bespiegeling op het rhymed/lyrical voorledene: dicht-schets 3. Hymne aan het vaderland over den veldslag van Friedland en de rhymed/lyrical daaropvolgende vrede van Tilsit: dichtstuk 4. Keur van Nederduitsche spreekwoorden en dichterlyke zedelessen rhymed/lyrical 5. Lykrede op Joannes Abraham Terbruggen (gestorven den 12 secular (linguistics) september 1819) stigter van het Antwerpsch tael- en dichtlievend genootschap onder de zinspreuk: Tot nut der jeugd. 6. Over de Hollandsche en Vlaemsche schryfwyzen van het secular (linguistics) Nederduitsch 7. Over het gedrag der Belgen by de scheuring der Nederlandsche secular (argumentation) Provinciën in de zestiende eeuw 8. Redevoering over het karakter van den Nederlandschen Schilder, secular (argumentation) gehouden in het Koninglyk Museum van Antwerpen 9. Reinaert de Vos: episch fabeldicht van de twaelfde en dertiende rhymed/lyrical eeuw 10. Verhandeling over de Nederduytsche tael- en letterkunde: secular (argumentation) opzigtelyk de Zuidelyke provintien der Nederlanden Jan F. Willems & Jan B. Brief-wisseling tusschen J.F. Willems, schryver van het werk: Tael- en personal (letter) Buelens letterkundige verhandeling, opzigtelyk de zuydelyke provintien der Nederlanden, en J.B. Buelens, R.C.P.r Jan V. Wouters Herstelden luyster van den H. Gummarus, beschermheyligen der stad rhymed/lyrical Lier Eugeen Zetternam Het bestuer en de natie statutory (politics)

220 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

Google (Scholar) texts

Table 68 PDDW sources – Northern Dutch

Author name Text name Text type Jaap Frouws Mest en macht: een politiek-sociologische studie naar belangen- statutory (politics) behartiging en beleidsvorming inzake de mestproblematiek in Nederland vanaf 1970 (library.wur.nl) Ronald Kroeze Een kwestie van politieke moraliteit: Politieke corruptieschandalen en statutory (politics) goed bestuur in Nederland, 1848-1940 (books.google.com) Henk te Velde Van regentenmentaliteit tot populisme (openaccess.leidenuniv.nl) statutory (politics)

Table 69 PDDW sources – Southern Dutch (part 1)

Author name Text name Text type Bisschoppen van België Pastorale brief: Christen zijn in deze tijd (www.kerknet.be) personal (letter) Olivier Boehme Voor de natie, tegen het nationalisme: Hendrik de Man (1885-1953) statutory (politics) en de Vlaamse Beweging (repository.uantwerpen.be) Rik Coolsaet Buitenlandse politiek als georganiseerde aanpassing of als voortzetting statutory (politics) van de binnenlandse politiek? (rikcoolsaet.be) FOD België als voortrekker van de circulaire economie (economie.fgov.be) secular (science) Sarah De Geyter, e.a. Delegatie en taakverdeling in de NV (biblio.ugent.be) statutory (administration) Saskia De Groof & Frank Over onderwijs (vub.ac.be) secular (pedagogy) Stevens Yves Haeck & Clara Procederen voor het Europees Hof voor de Rechten van de Mens statutory (administration) Burbano Herrera (biblio.ugent.be) Peter Heyrman Middenstandsbeweging en beleid in België 1918-1940: tussen vrijheid statutory (politics) en regulering (kadoc.kuleuven.be) Geert Van Hoorick Handboek ruimtelijk bestuursrecht (biblio.ugent.be) statutory (administration) Luc Huyse Over politiek (lirias.kuleuven.be) statutory (politics) Dirk Jacobs 1. Nieuwkomers in de Belgische politiek? Bedenkingen over politieke statutory (politics) participatie van vreemdelingen. (homepages.ulb.ac.be) 2. Stemrecht, nationale identiteit en diversiteit (researchgate.net) statutory (politics) 3. Waarom polemisering? Het parlementaire debat over lokaal statutory (politics) kiesrecht voor vreemdelingen in Nederland en België in de jaren zeventig en tachtig (researchgate.net) Gustaaf Janssens Paul-Henri Spaak en het begin van de Belgische statutory (politics) onafhankelijkheidspolitiek (1936-1937) (persee.fr) François Levrau & Het janusgezicht van het onderwijs: Een literatuurstudie over het secular (pedagogy) Christiane Timmerman onderwijs in een interculturele samenleving. (repository.uantwerpen.be) Niels Matheve Selectie van ministers in het tussenoorlogse België (lirias.kuleuven.be) statutory (politics) Inge De Meyer & Jan Wetenschappelijke vaardigheden voor de toekomst: De eerste secular (science) Pauly resultaten van PISA2006. (pisa.marlon.be) Gert-Jan Put, e.a. De geografische spreiding van kandidaten op de Vlaamse kieslijsten statutory (politics) (2003-2010). (lirias.kuleuven.be) Rik Röttger Een rode draad voor een blauw verhaal. De links-liberale uitwerking statutory (politics) van mimetische representatie en de opvattingen over democratisch burgerschap in België, 1893-1900. (library.uu.nl)

APPENDIX 1 221

Table 69 PDDW sources – Southern Dutch (part 2)

Author name Text name Text type Henk de Smaele Eclectisch en toch nieuw. De uitvinding van het Belgisch parlement in statutory (politics) 1830-1831. (library.uu.nl) Anna Vanhellemont België quo vadis: de zevende staatshervorming (lib.ugent.be) statutory (politics) Frederik Verleden De toegang tot de parlementaire elite: politieke rekrutering en statutory (politics) lijstvorming in België in historisch perspectief (lirias.kuleuven.be) Bruno de Wever, e.a. De Vlaamse patriotten en de natie-vorming. Hoe de Vlaamse natie statutory (politics) ophield ‘klein’ te zijn. (biblio.ugent.be) Els Witte Centrumvorming in België. De rol van Brussel tijdens de stichtingsfase statutory (politics) van de Belgische staat (1830-1840). (library.uu.nl)

Abstract

(English)

This thesis presents a detailed analysis of the re-autonomization process of the modal auxiliaries in Dutch. It aims to investigate the timing, the grammatical and semantic properties, and the factors that play a role in the emergence and development of the new autonomous uses of the modals. The study is corpus-based and works with representative samples from three different stages of New Dutch, i.e. Early New Dutch (END), New Dutch (ND) and Present Day Dutch (PDD). It shows how kunnen, mogen and moeten are increasingly used ‘independently’ in the clause in PDD (i.e. without the presence of another main verb), while in hoeven the upsurge of new autonomous uses already happened in ND. It also reveals that the new autonomous uses are no continuation of the original main verbs, but instead have emerged from the purely auxiliary uses (through a stage with an implied main verb), suggesting we are dealing with a case of (collective) degrammaticalization. From a grammatical perspective, the study shows that the development of the new autonomous uses occurs along two alternative pathways. On the one hand, there are instances whose first argument refers to a state of affairs. On the other hand, there are instances with a (pro)nominal first argument. From a semantic point of view, the new autonomous uses are shown to feature (highly) subjective and intersubjective readings (comparable to the auxiliary uses), and no ‘lexical’ meanings (as opposed to their original main verbs). As to possible explanations for the re-autonomization process, the study demonstrates that although directionals do play a role in triggering new autonomous uses of the type with a (pro)nominal first argument (as in other Germanic languages), they are not the determining factor in causing the general re-autonomization trend in Dutch. Negation, on the other hand, does seem to play a role in the entire re-autonomization process. Yet, how exactly it affects the process remains unclear so far.

Keywords modal auxiliary, diachrony, Dutch, re-autonomization, grammaticalization, degrammaticalization, subjectification, intersubjectification, directionals, negative polarity 224 THE RE-AUTONOMIZATION OF THE MODAL AUXILIARIES IN DUTCH

(Nederlands)

Deze thesis beoogt een grondige analyse van het herautonomiseringsproces van de Neder- landse modale hulpwerkwoorden. Ze heeft tot doel de timing, de grammaticale en semantische eigenschappen, alsook de factoren die een rol spelen in het ontstaan en de ontwikkeling van de nieuwe autonome gebruiken van de modale werkwoorden te onderzoeken. De studie is corpusgebaseerd en maakt gebruik van representatieve steekproeven uit drie verschillende stadia van het Nieuw Nederlands, te weten: Vroegnieuwnederlands, Nieuwnederlands en huidig Nederlands. Ze toont aan hoe kunnen, mogen en moeten in het huidige Nederlands in toenemende mate zelfstandig gebruikt worden in de zin (d.w.z. zonder de aanwezigheid van een ander hoofdwerkwoord), terwijl in hoeven die toename van nieuwe autonome gebruiken al plaatsvond vanaf het Nieuwnederlands. Ze laat ook zien dat de nieuwe autonome gebruiken geen voortzetting zijn van de oorspronkelijke hoofdwerkwoorden, maar integendeel ontstaan zijn uit de zuiver hulpwerkwoordelijke gebruiken (via een tussenfase met een geïmpliceerd hoofdwerkwoord), hetgeen suggereert dat we hier te maken hebben met een geval van (collectieve) degrammaticalisatie. Vanuit grammaticaal perspectief toont de studie aan dat de ontwikkeling van de nieuwe autonome gebruiken langs twee alternatieve paden verloopt. Enerzijds zijn er gevallen waarvan het eerste argument verwijst naar een stand van zaken. Anderzijds zijn er gevallen met een (pro)nominaal eerste argument. Vanuit semantisch oogpunt wordt aangetoond dat de nieuwe autonome gebruiken (zeer) subjectieve en intersubjectieve betekenissen uitdrukken (vergelijkbaar met de hulpwerkwoorden), en geen lexicale betekenis (in tegenstelling tot hun oorspronkelijke hoofdwerkwoorden). Wat de mogelijke verklaringen voor het herautonomiseringsproces betreft, bewijst de studie dat hoewel directionals een rol spelen in het stimuleren van nieuwe autonome gebruiken van het type met een (pro)nominaal eerste argument (net zoals in andere Germaanse talen), zij niet de bepalende factor zijn die de algemene herautonomiseringstendens in het Nederlands veroorzaakt. Negatie daarentegen lijkt wel een rol te spelen in het algehele herautonomiseringsproces, maar hoe precies blijft vooralsnog onduidelijk.

Trefwoorden modaal hulpwerkwoord, diachronie, Nederlands, herautonomisering, grammaticalisatie, degrammaticalisatie, subjectificatie, intersubjectificatie, directionaliteit, negatieve polariteit