Norms for style and grammar in eighteenth-century Dutch prose, and the effect of education and of writing experience

(print instructions)

Gijsbert Rutten (contact)

(University of Leiden)

Received: May 2008, published December 2008 (HSL/SHL 8)

1. Code-switching as a source of grammatical and stylistic norms in eighteenth-century Dutch prose

Style differences have always attracted the attention of historical linguists.¹ The language of farces has been used as a source of for instance spoken language, non-standard language, dialect and lower class language – varieties that are barely recorded, as historical documents are usually written by educated writers. For instance, the Dutch informal form of address singular jij or je appears in informal situations in early-seventeenth-century Hollandic plays, but in other written Dutch, the older form gij was usual even at the end of the nineteenth century (Vermaas 2002:42-52).

While in this particular case reliance on plays is fairly safe as the language of drama indeed reveals the early existence of the jij and je forms, there is general agreement that one cannot be too cautious when dealing with literary representations of non-standardised usages, lower-class language, regionally coloured language, and informal language in general. The Amsterdam playwright G.A. Bredero (1585-1618) does not only depict a fairly negative image of immigrants from Antwerp, he also fails to provide an accurate transcription of their speech.² On the other hand, it is claimed that Constantijn Huygens (1596-1687) in the play Trijntje Cornelis adequately renders the Hollandic and Antwerp sounds, though it is admitted that inaccuracies occur even in the representation of Hollandic (see Huygens 1987 [1653]:54). Be this as it may, one can never rule out that a playwright’s representation of informal or lower-class language is in fact only an interpretation of deviations from his own style, in other words, that his so-called lower-class language is better termed non-upper-class language.

In the eighteenth century, different styles were employed to represent different regiolects and/or sociolects (cf. De Vooys 1951, Daan 1995), most notably by authors such as Justus van Effen (1684-1735) and the joint novelists Betje Wolff (1738-1804) and Aagje Deken (1741-1804). Van Effen, the main author of the periodical De Hollandsche Spectator (1731-1735), is fond of mixing lower styles into his own cultivated prose. He does so by publishing letters from readers which have in fact been written by van Effen himself or by one of his co-authors. In these supposedly original letters, differences in style and grammar appear while the actual author remains the same. In that sense, we can interpret these changes in style and grammar as particular instances of code-switching. The author switches from his own highly educated style to a dialectic and/or sociolectic new code. Similar instances of code-switching appear in the epistolary novels by Wolff & Deken. The bourgeois main characters sometimes mingle with the working classes, for example when they receive a letter from a servant, or give an account of a meeting with a fisherman. Again, changes in style and grammar appear which are apparently sociolinguistically based.

In sections 2 and 3 of this paper, the question will be addressed whether these eighteenth-century instances of code-switching can be of use for historical sociolinguistic research. This seems improbable as – apart from the fact that we cannot know to what extent the new code truly represents the linguistic situation – a specific technique appears to be involved by which the author’s prose is slightly modified in order to suggest dialectic and/or sociolectic authenticity. In section 4, I briefly take stock of the predominantly negative results of sections 2 and 3, and then another example of code-switching is discussed, not taken from a literary text but from a so-called ego-document: a journal written by a non-professional but educated author in which a letter by his less-educated grandmother is included. Here, we find a more reliable source of grammatical and stylistic differences. As the letter is relatively short, we can only use it as a point of departure for further historical sociolinguistic research into eighteenth-century style and grammar; in 4, the direction for this further research is discussed. Some final observations will be made in section 5.

2. Van Effen: a literary pastiche

Justus van Effen counts as one of the best Dutch eighteenth-century authors, and was especially famous for his journal De Hollandsche Spectator (“The Dutch Spectator”)³ in which he shows himself to be a moralist with a good sense of humour. Van Effen is fond of mixing lower styles into his own literary Dutch. In one such fragment, an ethical repudiation of horse racing, van Effen cites a letter he has supposedly received from a certain Klaas Janssen, a somewhat less literate innkeeper and horse trader from the country, supposedly from somewhere around Amsterdam (HS 7, 1731:50-54). In this letter of about 1600 words, linguistic phenomena occur that are usually absent from literary poetry and prose. At first glance, van Effen appears to have used a wide range of non-standard forms as almost every sentence contains at least one uncommon form. Upon closer inspection, however, the matter turns out to be quite different. First, I will present the phenomena (2.1-2.4) and my interpretation of their use (2.5), and then another case is discussed in connection with previous research on the matter (2.6).

2.1 Phonology and orthography

Some 60 phonologically marked items appear in the text. See Table 1.

Table 1. Phonologically marked items in the fabricated letter of Klaas Janssen (HS 7, 1731:50-54); on the left the form in the letter with its frequency, on the right common “standard” Dutch with English translation.

In the left-most column there are five forms (veul, ummers, veur, durp, mot) that are well-known for having been widely used but were absent from cultivated language. These are so common that no specific regional or dialectical label applies, though in Middle Dutch especially Hollandic texts show <eu> in veul (cf. Schönfeld 1954:48). Veur, meugen and durp are considered to be examples of palatalisation of /o:/ and /c/ respectively (van de Ketterij 1980:26). This type of palatalisation would be exemplary for coastal Dutch or Ingweonic (van Loey 1965:124). Mot instead of moet (with /u/) is common in seventeenth-century texts but does not seem to appear regularly in eighteenth-century private documents from less educated people (van der Wal 2007:92-93); in the letter we also find moet. Of instead of af is a well-attested Hollandic and Brabantic, or in general Western, variant in Middle Dutch (van Loey 1965:11). The same goes for bedogt and edogt.

In the middle column there are words commonly spelt with <ee> for /e:/ but here with <ie> for /i/; an instance of, again, palatalisation that already in seventeenth-century literature, especially farces, was associated with “broad Hollandic” (cf. Weijnen 1956:9, 27). In Middle Dutch, it also frequently turns up in Brabantic and sporadically in Flemish (van Loey 1965:46-47). In the last example, mit with /I/ instead of /ε/, the short vowel is different but the phenomenon is the same.

The column on the right contains typical examples of phonological reduction: strong vowels are reduced to schwa (especially in frequent words such as pronouns and articles, cf. Bybee & Hopper 2001:10-13). In als /l/ is deleted, which is still very common in spoken Dutch. Equally common in Modern Dutch is deletion of the final consonant in hou and ree, showing the filler consonant /w/ in plural zouwen and in the past participle ehouwen. It is likely that phonological reduction would have taken place in van Effen’s own spoken language but he kept spelling the words in their unreduced form. In the fabricated letter, however, he considers the orthographical representation of the reduction appropriate, though we also find the full variants mijn, een and zijn in the letter.

Other orthographical features are a few finite verbs ending in –d instead of common –t such as kend “know” (2sg), woond “lives” (3sg), scheeld “hurts” (3sg) and brengd “bring” (imperative pl). Note that Dutch consonants show final devoicing (Auslautverhärtung) as a result of which there is no phonological but merely an orthographical difference between final -d and final -t. There is also one clear example of enclisis with a filler consonant bridging the hiatus between two vowels: Ik gaader instead of Ik ga er “I go there”.

To the forms in Table 1 only two phonogically remarkable forms can be added from the letter. The first is beterschup instead of beterschap, “recovery”. This is an uncommon instance of rounding in the suffix –schap. In (coastal, Ingweonic) Hollandic, a palatalised form such as –schip (cf. English –ship) would be more probable (Schönfeld 1954:195). The second is wurld for werrelt (with /ε/) or wereld (with /e:/), “world”: again an uncommon case of rounding, sporadically attested for Middle Dutch, esp. in Brabant but also in Holland (van Loey 1965:22). I would say that van Effen uses these forms not because they were common rural Hollandic but because of analogy: as is clear from the first five words in the left-most column in Table 1, van Effen considers the vowels /Y/ and /ø:/ which mainly differ in length (cf. Gussenhoven 2007:338-339) as shibboleths of less civilised speech.

2.2 Morphology

Morphologically, also some 60 marked items appear in the letter. As with the phonological variants, the phenomena can easily be grouped; see Table 2.

Table 2. Morphologically marked items in the fabricated letter of Klaas Janssen (HS 7, 1731:50-54); on the left the form in the letter with its frequency, on the right common “standard” Dutch with English translation.

By far dominant is the phenomenon in the left-most column, the past participle formed with the prefix e- instead of ge-. In seventeenth-century texts, this phenomenon (and its extreme variant: non-realisation of the prefix) occurs almost exclusively in Hollandic texts (Hermkens & van de Ketterij 1980:122; Weijnen 1956:9). Historically, e- and ø were common in larger parts of the Low Countries, especially along the coast, in the north and in the east (Schönfeld 1954:157).

The middle column of Table 2 contains well-known deviations from the verbal inflection commonly displayed in written Dutch. Some are typically associated with Hollandic (heyt, zel, bennen; cf. van der Wal 1992:229). Wou is still generally in use in spoken Modern Dutch. Maggen is a not very frequent form analogical to the singular mag which has never reached the standard; maggen does not (or hardly?) appear in Middle Dutch. The two participles ezeid and ekonnen are both common in Middle Dutch and disappeared in the Early Modern period (van Loey 1955:86; van de Ketterij 1980:61, 64), though they are still attested by the grammarian Sewel (1712:261, 311). Loof is called ordinary language; it appears in seventeenth-century farces situated in Holland (WNT s.v. “geloven”). It might be a case of analogical, hypercorrect prefix deletion (cf. the left-most colum of Table 2).

Another such case (-ezond) is found in the right-most column of Table 2, where the few non-verbal deviations attested are recorded. The contraction aare is, again, termed ordinary language (WNT s.v. “ander”). The pronoun mijn instead of mij is (supposedly) a mainly Hollandic variant, not accepted in seventeenth-century written Dutch but attested in eighteenth-century ego-documents (van der Wal 2007:91-92; see also below, 4.3).

2.3 Syntax and style

In the field of syntax, only one remarkable item appears: five times van Effen uses the so-called double, or preferably bipartite, negation consisting of the clitic en and a negative particle (geen or niet), comparable with French ne … pas. Double negation is common in Middle Dutch. It disappears from written Dutch during the sixteenth and seventeenth centuries.

Stylistic features which are frequently found include semantically empty phrases and useless clichés such as of zo “or so”, maar geloof me “but believe me”, je kend Klaas ummers wel “you know Klaas after all”, zei ik zo “I said so”, zei hy zo “he said so”, ’t gaat hoe ’t gaat “things go/are as they go/are”, men kan niet meer doen als zen best “one cannot do more than one’s best”, zo lang als ’er leven is, is ’er hoop “as long as there is life, there is hope”, verstaje wel “do you understand”, als ik boos ben dan deug ik niet, en dat weet ’t hiele Durp “if I am mad I am no good, as the whole village knows”, moet je weten “you know”, dat weet je “you know”.

2.4 Lexicon

Two words which were probably considered old-fashioned in the eighteenth century are found: ereis “first” (3) and eevel “after all” (2). Additionally a few badly spelt French words are used: chees (< chaise), pasientie (< patience), apperpos (< à propos), condisie (< condition), resenabel (< raisonable) and also the wrongly used infaam (< infâme); in a note, van Effen explains that befaamd “famous” is meant; infaam means “dishonourable” – a feeble joke as the writer Klaas Janssen is so immoral as to defend horse racing.

2.5 Van Effen’s technique

The ca. 60 phonologically marked items mainly correspond to three strategies: (1) turning vowels into <eu>, <u> and <o>, (2) palatalisation of <ee> into <ie> and (3) the orthographical representation of phonological reduction. Morphologically, an equally limited development seems to occur: (1) an abundance of past participles with e-, (2) a few other verbal items which were probably considered old-fashioned and/or vulgar and (3) the pronoun mij with –n; in total again ca. 60 items. Whereas the text (1600 words) displays a vast amount of non-standard forms, these can easily be arranged in a way that suggests that van Effen is just drawing on a few principles (see Tables 1 and 2). In other words, it seems that van Effen uses a technique in order to represent what he counts as non-standard, or in eighteenth-century jargon, uncultivated language usage. The technique consists of applying these few principles to lexical items. This technique is most easily employed in the fields of phonology and morphology. He tries to present the language more realistically by mixing a few badly spelt French words into it and by using two older words a few times. He also chooses one syntactic feature, and a few superfluous stylistic and pragmatic turns of phrase. All in all, we may conclude that in a text of 1600 words about 140 marked grammatical items (some 60 phonological, 5 orthographical, 60 morphological, 5 syntactic and 10 lexical) plus a range of stylistic and pragmatic superfluities indeed suffice to give the impression of dialectic authenticity.

If we compare the language of the letter to the findings of recent historical sociolinguistic research, one thing strikes us: the complete absence of “stylistic breakdown” (Stilzusammenbruch, cf. Mattheier 1986:247-248), that is, the writer’s inability to meet the needs of the written medium syntactically and stylistically. Instead, Klaas Janssen’s style and syntax is noticeably fluent. Apparent shibboleths of non-standard Dutch are woven into this well-written piece of prose. Because of these seemingly dialectal features, the few orthographical mistakes, and the superfluous pragmatic turns of phrase (“you know”), the illusion of a written down version of spoken Dutch is created, of which the letter’s fluent style could seems to be a characteristic. However, the fluency of the style is only a proof of van Effen’s literary talent.

In sum, the fabricated letter is a fine example of stereotyping. It is impossible to know whether it presents a true image of a certain lower-class form of written Dutch, though the fluency of style and syntax suggests it does not. What we do learn from it, however, is which linguistic items van Effen considers to be no part of well-written Dutch, in other words, which phenomena he rejects (cf. te Winkel 1924:231). This aspect is particularly important because van Effen aims at writing common Dutch as opposed to the prevailing learned and puristic style of poets and grammarians (HS 2, 1731:9-16), by which he, being a man of letters, has inevitably been influenced (de Vooys 1951:75-76).

But even if we do know which phenomena van Effen rejects, we will be unable to discover where he heard or read them. Do they belong to the relatively recently developed urban dialect of Amsterdam and to what extent does van Effen himself use these features in his speech? Are they part of the older rural dialects from Amsterdam’s surroundings? Or has he encountered them in seventeenth-century plays in which similar phenomena are employed in the representation of lower-class people? Even if we would know that they were still part of the spoken language,⁴ we have no way of knowing where, when, by whom, let alone how often they were used. As there are more fake letters from lower-class members in van Effen’s Hollandsche Spectator, we might get a grasp on these matters by comparing the letters. In fact, this has been done and I will discuss this in the next paragraph.

2.6 Another letter – the same technique?

It should be noted that most of the fake letters in the Hollandsche Spectator are not written by van Effen but by co-authors (de Vooys 1951:78). Van Effen composed the letter by Klaas Janssen discussed in 2.5 but not the one by the weaver Japik Schietspoel (HS 188, 1733:305-308) which was written by Pieter Merkman and which will be discussed here. Daan (1995) examines some of the supposedly lower and middle-class language fragments in the Hollandsche Spectator and draws sociolinguistic conclusions on the basis of her findings. Unfortunately, she does not deal with Klaas Janssen’s letter.

Like Klaas Janssen’s letter, the one by Japik Schietspoel (1300 words) displays a number of items not common in written Dutch. I counted about 170 grammatically marked items, even more than in the previous letter of 1600 words which had about 140 such items. See Table 3.

Table 3. Grammatically marked items in the fabricated letter of Japik Schietspoel (HS 188, 1733:305-308); on the left the form in the letter with its frequency, on the right common “standard” Dutch with English translation.

The left and the middle columns show the most frequent (> 3 tokens) deviations from what is normally found in Schietspoel’s letter. It is immediately clear that these are highly similar to the ones in the previous letter: orthographically represented phonological reduction (sen, men, houw, as), the typically Hollandic verbal inflections (even in Modern Dutch) het and sel and the object pronoun myn, palatalisation (bin, mit) and the /ø:/ realisation in veul and veur, the latter of which gave rise to the analogical forms on the right in which also common <oo> (/o:/) appears as <eu> (/ø:/). The middle column shows an additional item compared to Janssen’s letter (ken), though this is also a well-attested Hollandic variant which is still in use.

A few more less frequent items can be added to these, such as sukken < zulken “such” with <l> deletion, a very common form in colloquial and dialectical Dutch (cf. WNT s.v. “zulk”) similar to veul, durp and mot in Table 1, and also dikkels < dikwijls “often”, again a reduced form (28 tokens in the Woordenboek der Nederlandsche Taal (WNT), from the seventeenth to the twentieth centuries). Another example is the reflexive pronoun ’em (< hem) which was replaced by zich in Early Modern Dutch and which probably counted as old-fashioned (cf. the use of double negation and of a word like eevel in the previous letter; evel is also recorded in the current letter). Then there is begost, a sporadically attested variant of begon “began” in the seventeenth century (Hermkens & van de Ketterij 1980:135), and verkost instead of verkocht “sold” which is an unusual variant: the Hollandic form would be verkoft (Schönfeld 1954:96). Maybe verkost is formed analogically to begost, or is the result of confusion with another money related verb: kosten “to cost”.

As with Klaas Janssen’s letter, some old-fashioned (evel) and badly spelt French words appear, e.g. pissiensi (< patience), consolaesi (< consolation). Syntactically, no remarkable phenomena appear (no double negation, no stylistic breakdown). Again, semantic emptiness is at hand: ik seg het so rond uyt, en so as het me op het hart legt “Thus, I say it straightforward, and as I have it on my mind”.

The letter by Schietspoel displays the same characteristics as the one by Klaas Janssen and we can thus conclude that we are dealing with a technique. The authors employ a reservoir of phonological, morphological and to a lesser extent lexical and stylistic non-standard phenomena and they mix these into their own syntactically and stylistically flawless prose. Most of the items are already attested in seventeenth-century plays, especially farces, so that we cannot know whether they represent early-eighteenth-century spoken Dutch or a literary tradition of stereotyping.

There is no reason, in my opinion, to assume that these letters and other similar fragments from the Hollandsche Spectator provide reliable material for historical sociolinguistic research. Yet this is what Daan (1995) suggests in her study of eigteenth-century Dutch sociolects (cf. Verdenius 1946:19). She distinguishes three variants: (1) the language of the upper-classes: gentry and bourgeoisie, (2) the language of the middle classes, e.g. small traders and craftsmen (3) the language of elderly people and of the rural population. Japik Schietspoel, a weaver, would have used the second variant, while Klaas Janssen as a writer with a rural background uses the third. Following Daan (1995), Janssen should gutteralise the <nd> cluster, e.g. stongden instead of stonden “stood”, ongder instead of onder “under” but Janssen’s letter does not show this type of gutteralisation. To me, it appears that van Effen’s fake letters are (socio)linguistically as unreliable as Bredero’s farces, which Daan herself rightly points out in her edition of his plays (Daan 1971 [1612-1613]:28-30, 246-248).

All we learn from van Effen’s texts is what variants he refutes. Unfortunately, we do not know to what extent these variants were in use at the beginning of the eighteenth century. Nevertheless, we do learn about a strategy by which the higher middle and upper classes tried to distinguish themselves in written texts from the lower middle classes and the non-urban population: by linguistic stereotyping. The result is an occasionally humorous literary pastiche.

3. Wolff & Deken: the same old story?

Two other famous Dutch prose writers from the eighteenth century, the novelists Betje Wolff and Aagje Deken, also engage in style mixing in their work, and again it is claimed that the different styles represent the sociolinguistic make-up of late-eighteenth-century Holland (de Vooys 1970:145). The claim, however, is rightly accompanied by two important restrictions: first, Wolff & Deken wrote epistolary novels, so there are ample typically written language phenomena in their novels, and secondly, the examples illustrating the claim are all lexical, that is, borrowings from everyday or dialect language into the letters of socio-economically higher characters (de Vooys 1970:145). However, de Vooys still claims that two specific characters “write entirely as they speak” (see 3.2). Besides, Hol (1941:278) uses a passage from Wolff & Deken as linguistic evidence for his study of the past participle prefix (see 3.1), which is an additional reason to look into the novels of Wolff & Deken.

3.1 The (un)reliability of Sara Burgerhart

In Sara Burgerhart (Wolff & Deken 1782), we find two linguistically marked letters. The first one, no 64, is written by Sara’s former servant and nanny Pieternelletje Deegelyk; the second one, no 69, by a negatively evaluated, somewhat superficial lady who is significantly called Charlotte Rien du Tout (1782:289-293; 307-308). At the end of letter no 69, the following passage occurs: “The editor deemed it fit to purify this letter a little, as the one by Pieternel, from language and writing errors, so that one would be able to read them”.⁵ This in itself is more than revealing with regard to the inherent social stereotyping.

A linguistic analysis confirms the dubious status of these two letters. In no 69, the only remarkable items we find are two tokens of myn where my (mij) would be appropriate from a normative perspective (cf. sections 2.2 and 2.6). In no 64, a few more strategies are employed which hardly differ from those van Effen used half a century earlier. We find a few instances of “vulgar” Dutch or broad Hollandic such as kommen instead of komen “come”, lei for legde “lay”, ouwers for ouders “parents”. Some of these are used more than once: het for heeft “has” occurs five times, deuze for deze “this” twice and drok for druk “busy” twice as well. The word evel “after all” also occurs twice (cf. 2.4 and 2.6). With respect to the personal pronoun mij, letter no 64 is similar to no 69: myn and the reduced form men both appear once instead of my (mij) – and both in the first sentence! It seems that the first sentence is used as a sociolinguistic description (stereotype) because from then on regular my appears (6 times). Furthermore, the possessive pronoun myn is written once in its full form and 14 times is reduced to men (cf. 2.1). So far: the same old song as with van Effen.

Yet Wolff & Deken seem to be more reliable witnesses than van Effen, since the language they ascribe to a fisherman from the town of Maassluis in South Holland differs from the North-Hollandic items described in the previous paragraph which may be the reason why Hol (1941) considers Sara Burgerhart reliable material. In the passage in question (1782:246-248), taken from letter no 55, a few remarkable items appear:

These items are not recorded in letters 64 and 69 (but note that van Effen records the e-prefix and palatalised /e:/ for North Holland, see 2.2). Moreover, the marked variants of the third person singular of the verb hebben “to have” (i.e. heeft) are regularly distributed by Wolff & Deken: het in North Holland (letters 64 and 69), eit in the South, that is, Maassluis (letter 55). This certainly suggests that Wolff & Deken have some knowledge of regional linguistic variation, though the objections and doubts raised with regard to van Effen still hold: it is impossible to decide to what extent their observations are true representations of the current variants. However, the 1871 recording by Winkler (1874:149-153) of the dialect of Vlaardingen, which would be very similar to that of the nearby town of Maassluis, shows the following features to be characteristic of the dialect: ø instead of /h/, e- instead of ge-, and eit for heeft. Palatalised /e:/, on the other hand, is not attested, though the century in between Sara Burgerhart (1782) and Winkler’s recording (1871) may be of influence here. All in all, the language in Sara Burgerhart appears to be somewhat more reliable than van Effen’s pastiches in that the authors show awareness of regional linguistic variation. However, it seems to be a little far-fetched to conclude that Sara Burgerhart provides a true image of eighteenth-century variation.

3.2 The novelty of Willem Leevend

In another epistolary novel by Wolff & Deken, the eight-volume Willem Leevend (1784-1785), two characters are presented who allegedly “write entirely as they speak” (de Vooys 1970:145). The characters in question are the older couple Martha and Frederyk de Harde. In order to investigate their language, I selected the letters from Martha in vols. 1 and 2, six letters in all, and the three letters by Frederyk, which appear in vols. 4, 6 and 7.⁶ What is striking in these letters is that Wolff & Deken seem to adopt a new strategy for the representation of non-“standard” Dutch. The phonologically, morphologically and lexically marked items – see Table 4 – are reduced in favour of syntactic and stylistic features.

Table 4. Phonologically, morphologically and lexically marked items in the language of Martha and Frederyk de Harde (Wolff & Deken 1784-1785).

Most items in Table 4 only appear once or a few times in the texts. The verbs and the possessive pronoun jen appear more often. To the items in Table 4 the same applies as to those in Tables 1, 2 and 3. Phonologically, this means incidental palatalisation, realisations of <eu>, phonological reduction in the pronouns (still very common in spoken Dutch) and softening of <d> into <w>, /j/ or ø (also common in spoken Modern Dutch). Morphologically, there are the “vulgar” verbal inflections we met before, and the pronoun jen which is a very remarkable form almost exclusively attested in the writings of Wolff & Deken. This –n-variant does not just appear before vowels or <h>, but is used generally throughout the letters of Martha and Frederyk. The uniqueness of jen makes it improbable that it was a common spoken form. The WNT (s.v. “jouw”) suggests Wolff & Deken created it with the reduced forms men and zen. As for lexical items, we once again find ereis and evel, and some other colloquial lexical items: eigenste and allegaar are deemed old-fashioned and/or vulgar by the WNT (s.v. “eigen” and “allegaar”), koeteren is colloquial for speaking badly and especially for speaking French badly. I encountered one additional example of bad French: apperepo, “à propos”; this kind of misspelling had already been employed by van Effen.

To these grammatical phenomena also applies what was stated before: these are standard vulgarisations of what was considered correct Dutch, and they are part of a tradition which goes back to van Effen at the beginning of the eighteenth century, and beyond, to the seventeenth-century farces. But Wolff & Deken elaborated on these compulsory deviations from the norm and included much more striking syntactic and stylistic features. The authors use a range of techniques in order to create the illusion of speech in the evidently written letters. The most peculiar example is, to my opinion, the self-correction ei, ik verspreek my “oops, I made a slip of the tongue”, as if a less-educated person is likely to confuse speaking with writing.

Syntactically, the techniques come down to two basic principles: first, typically spoken-Dutch constructions such as NP-dislocation, as in (1), and additional, from a normative perspective superfluous, supportive subordinators such as of and dat, for which see example (2) (cf. 4.2 below); secondly, short and often paratactically connected sentences as in (3).

Stylistically, the letters are characterised by a wide range of semantically empty phrases and useless clichés similar to those listed in 2.3 above, proverbs as in (4), idioms as in (5) and quotations which are often proverbial, derived from the bible and from popular moral literature as in (6). Examples (4) - (6) contain just a few examples; (5a-b) illustrate captain Frederyk’s language, which was evidently influenced by his profession.

The syntactic and stylistic techniques employed here certainly cause the impression of naturalistic speech: the syntactic phenomena illustrated by (1) - (3) emphasise the spoken character of the language; the stylistic ones in (4) - (6) would be exemplary of less-proficient, less-educated language. The fundamental problem, however, is that this supposedly spoken Dutch turns up in an evidently written genre: the personal letter. The abundance of the marked phenomena suggests that Martha and Frederyk and the social group they represent have hardly an idea of what it means to write a letter, which is of course highly improbable. Further, by concentrating on syntactic and stylistic phenomena, Wolff & Deken certainly make progress with respect to the representation of non-standard Dutch in that they move away from the obvious stereotyping based upon phonology and, to a lesser extent, morphology. Also, it has been shown that less-experienced writers tend to use a fair amount of fixed expressions and constructions (Elspaß 2005) and in this respect the representation of middle-class language is certainly improved. However, it remains a fact that syntax and style are remarkable but also remarkably flawless: the syntax is perfectly grammatical, semantics is never endangered, and the style is simply fluent. Again: it seems to me that stereotyping techniques are woven into Wolff’s & Deken’s own literary Dutch, and consequently, that Martha and Frederyk definitely do not write as they speak.

4. Van der Does and his grandmother

The conclusion that the texts by van Effen (see 2 above) and Wolff & Deken (see 3) are unreliable sources for historical sociolinguistic research – not necessarily for historical linguistic research of course – does not have to keep us from using them to formulate hypotheses. As stated above, since writers adopt certain linguistic forms for the purpose of stereotyping their characters, we could investigate to what extent these forms were indeed used, and to what extent they are part of a literary tradition. In order to do so, a large corpus of preferably non-published texts by lower- and middle-class writers should be analysed. Such a corpus, however, does not yet exist. Virtually all the above-mentioned linguistic phenomena could be subject to research but the most interesting would be those that consistently turn up such as the variation of mij and mijn (see also 4.3), the apparently wide-spread variation in verbal inflection (see 4.4), syntactic features such as double negation and multiple connectives (wat of, hoe of, etc.) and fixed expressions and constructions.

Due to the absence of large corpora, case studies are necessary. In the late-eighteenth-century ego-documents I have collected so far, similar code-switching passages occur as in van Effen’s and Wolff’s & Deken’s texts. These can provide a more reliable picture of grammatical and stylistic usages and norms in eighteenth-century written Dutch. To this topic I will turn in this section.

Not only are well-established literary authors sensitive to the existence of differences of style and grammar in the language, also non-professional and less experienced yet educated writers show an awareness of different levels of linguistic competence. When writing they produce linguistic forms that must be the result of education and/or experience. This is clearly suggested for instance by the fact that they tend to write in what grammars used to consider complete sentences: beginning with a capital, ending with a full stop, or by the fact that they often employ a fairly regularised orthography. If ego-documents by such non-professional writers show cases of code-switching as discussed above (sections 2 and 3), then maybe we are able to get a better grasp of stylistic and grammatical variation across social groups, as it is improbable, though not impossible, that the author’s goal was to create a pastiche.

In 1787, the twenty-year old Willem van der Does kept a journal recording his experiences during the troubled times the Low Countries were then suffering from. Van der Does was the son of a regent from ‘s-Hertogenbosch, an important city in the province of Brabant. As an upper-class youngster, he spent his teens in Latin school and then went to Leiden to study law. During the patriot (republican) revolution and the monarchic counter-revolution, he lived with his family in Brabant, though at times they were separated from each other. If that was the case, they communicated by letter, and one such letter, from van der Does’s grandmother, young Willem reproduced in his journal, creating a similar narrative structure as van Effen above (section 2), the important difference being that here we have a real rather than a fictional letter. In this section, stylistic and grammatical differences between the grandmother’s letter and van der Does’s own text will be studied, as well as some instances of quotations from actual speech.⁷

The letter by van der Does’s grandmother counts only 235 words. As with van Effen and Wolff & Deken, however, the text can be used as a means to formulate hypotheses for research. Besides, the variation displayed in the document by van der Does, his grandmother’s letter included, provides a telling insight into sociolinguistic variation within one family. Though we have to reckon with the fact that the letter was copied by Willem, in the process of which some changes may have been made, this letter is an interesting source, and I will use it as evidence to formulate a number of hypotheses for further research. Next, I will make a start with this research by comparing the outcome of the van der Does study with other ego-documents, most importantly the letters that Christina van Steensel wrote to her husband in the years 1782-1800 (see van Steensel 1994 [1782-1800]). Van Steensel was a middle-class woman with some education and writing skills. Her letters of about 34,000 words represent a fairly large source for the late-eighteenth-century language of middle-class, less-experienced writers. Occasionally, I will also use other sources.

Whereas van Effen, and Wolff & Deken to a lesser extent, mainly focus on those linguistic phenomena that attract the most attention when one is confronted with a dialect, namely the phonological differences, in the present case we find syntactic items to be of great importance. Of the following four sections two are therefore concerned with syntax (4.1, 4.2) and two with morphology (4.3, 4.4).

4.1 Stylistic “breakdown”

The concept of Stilzusammenbruch, stylistic breakdown, was introduced by Mattheier (1986:247-248; cf. Mattheier 1990) in connection with German labourers’ letters from the nineteenth century. It refers to the syntactic and stylistic deviations from educated written language which come about as a result of the confrontation of speech and only partly acquired norms for written language. In other words, lack of education and of writing experience leads to ungrammatical and pragmatically marked language. Problematically, the phenomenon was so wide-spread that communication was apparently not hampered by it. Instead of analysing the features as deviations from the norm it seems better to interpret them as regular features; this certainly does more justice to the communcative skills of the writers (cf. Vandenbussche 2007:284-285). Nevertheless, when juxtaposing educated and experienced language usage with less-educated and less-experienced language, or when comparing everyday language with standard and/or formal language, stylistic “breakdowns” remain important phenomena that need to be studied seriously.

Contrary to van Effen’s texts, the author’s grammatical and stylistic characteristics in the van der Does manuscript are not restricted to the word level (phonological, morphological, lexical) but are also displayed at sentence level. Whereas van der Does himself writes syntactically impeccable Dutch (from the perpective of the standard, that is), his grandmother’s texts show some instances of stylistic “breakdown”. Consider the following passage:

The word daar (“because”) in the first line is a causal conjunction introducing a sentence; it is coordinated with the phrase introduced by en (“and”) in the second line. Both phrases (“the nasty situation in which now in am” and “the departure of your husband who did not have time to eat”) would be an example of something which remains unexpressed: the discourse is not ended. Instead, a statement follows as to what the husband was forced to do; this statement does not contain a subject unless it is die (“who”) but then the link between eeten and zoo maar fails. Besides, the phrase daarin nu in ben lacks a subject while the preposition is doubled.⁸ Further, the last sentence (“and people spoke”) is coordinated with the previous utterance though a semantic relation seems to be lacking.

Despite the formal syntactic irregularities the message is quite clear. Moreover, a pragmatic pattern appears in the formulation of semantically relevant phrases in that a syntactically redundant and semantically empty conjunction is used as a marker: daar as well as the first and second en introduce a phrase. Perhaps this pattern is related to the (paratactic) main clause continuation of unfinished complex (hypotactic) embedded structures known from German (Mattheier 1990:293), and more generally to the paratactical chaining of phrases into a long sentence with several semantic functions, also attested for German (cf. Grosse 1990:305).

Further, though subject deletion may be inconsistent with normative grammar, from the text it is easily inferred which NP applies. Both phenomena appear once again in the grandmother’s letter. Consider (8):

Again, conjunctions such as maar and en are used to introduce phrases while the syntactic and semantic function is unclear, that is, does not correspond to normative grammar. The subject of nam and is is formally lacking but the NP uw knegt having just been mentioned – be it within a PP within another NP (het geval van uw knegt) – it is clear what the subject is. See also (9) in which the words in square brackets are added by me:

The pronouns [zij] and [hij] are subjects the author omitted; again, despite the problematical syntax, the meaning of the sentence is clear. Also, subordinating dat is lacking, which is a common phenomenon in the history of written Dutch (cf. van der Horst 1990) though certainly not in the history of educated prose, let alone in normative grammars.

The instances of stylistic “breakdown” – uncommon syntax from a normative perspective – in the grandmother’s letter are at once remarkable and understandable. While they are uncommon and rarely attested in historical linguistics, the comprehensibility of the message does not suffer as a result. Further analyses into the history of unpublished written Dutch will have to throw more light on these phenomena. How common is this kind of syntax for less educated, less experienced authors? In what way is it related to spoken-Dutch syntax? Can we recognise a historical-linguistic development in the features identified here? Two examples from the seventeenth and the nineteenth centuries are presented in order to illustrate the probable commonness of chaining phrases with the help of semantically somewhat empty connectives:

4.2 Conjunctions

The syntactic and functional development of connectives such as conjunctions has not received much attention in the history of Dutch. It appears, however, that major changes in form and use have taken place, in earlier as well as in more recent linguistic periods. In this section, I will discuss some conjunctions.

4.2.1 Want and want dat

In Modern Dutch, the causal conjunction want (“since”, “because”) functions as a coordinator introducing a main clause with V2, as in (12a). In Middle Dutch, however, subordinating want introducing a dependent clause with end-V seems to have been more common (12b).⁹

Somewhere during the late Middle Dutch or Early Modern Dutch periods, (12a) becomes regular and (12b) disappears. Burridge (1993:60-61) states that after 1500 subordinating want disappeared from Hollandic and after 1600 also from Brabantic. Van Megen (2002) shows that subordinating want (12b) is used late in the seventeenth century in private correspondence; cf. also (13):

In van der Does’s grandmother’s letter, from the end of the eighteenth century, want appears twice and both times as a subordinator; see (14) - (15).

In (15), the subordinator is in fact want dat. In so-called non-standard Dutch, dat (and of) are sometimes used as additional supportive subordinators (see 3.2, esp. (2) above). Besides, the subordinator want dat (“because”) with SOV has been attested for Middle Dutch, though it has not been recorded for later periods. Van der Does himself only uses want once (16): as a coordinator (cf. 12a).

The difference is striking. By the beginning of the eighteenth century, subordinating want (12b) had definitely become marked. One of the most important grammarians of the century, Moonen, in his Nederduitsche spraekkunst (1706:291-292), does mention (12b) but he explicitly refers to one specific author: Joost van den Vondel (1587-1679). Moonen himself consistently uses (12a). Nylöe’s grammar, of which the third edition came out in 1723, provides one example which also has (12a) (1723:70). In 1764, the grammarian de Haes contradicts Moonen: he rejects (12b) and claims (12a) is “much more appropriate and fluent” (1764:143). Other grammarians do not devote many words to the matter. Stijl, for instance, merely hints at the different syntactic functions of conjunctions, and states these can be acquired from usage (1778:144). Weiland, then, who at the beginning of the nineteenth century composes a national grammar for Dutch, elaborates a little on this when he writes that want does not, contrary to most conjunctions, change the so-called narrative word order (1805:300-301). In sum, in eighteenth-century published texts and grammar books, want (12b) gives way to want (12a); however, this is not the case in van der Does’s grandmother’s letter.

Is the use of want (12b) related to inexperience or lack of education? In that case, we would expect that other inexperienced or less-educated late eighteenth-century writers would use it. However, I have not come across a single instance so far. Christina van Steensel consistently used want (12a) (van Steensel 1994 [1782-1800]); likewise the little more educated author Arie Knock (1994) [1784-1797] in his journal on his travels and experiences as a soldier also uses this form (ca. 68.000 words).

Is want (12b) then related to age? Findings for want dat (cf. 15) indirectly support this hypothesis: although want dat is considered a Middle Dutch subordinator, it appears in late eighteenth-century material with SOV-order (cf. 12b and 15), not just with less-educated writers such as the soldier Knock (17) but also in the diary of a very-well educated, upper-class boy: Otto van Eck (18).

The use of want dat, then, appears not to be restricted to age: van Eck was a teenager when writing his diary. Both Knock and van Eck, however, consistently distribute want and want dat as in (19):

In published texts from the eighteenth century, and as a result in histories of eighteenth-century Dutch, (12b) and (19b) are not recorded. In unpublished texts, however, both appear. Note that the functional and syntactic flexibility of connectives has been attested for German everyday language in connection with weil (Elspaß 2005:304-316). Perhaps we can hypothesise that (12b) became less and less frequent during the seventeenth and eighteenth centuries, and that the competetion between (12a) and (12b) was decided in concordance with the equal distribution of different word orders. In other words: because (19b) existed, the choice between (12a) and (12b) worked out in the favour of (12a) (= (19a)). As a consequence, the two semantically and phonologically very similar connectives want and want dat were syntactically distinguished. Maybe this new distribution (19a-b) was well-known, or even introduced, by younger generations but not employed by the older ones among which van der Does’s grandmother.

4.2.2 Maar, doch and dan

As in Modern Dutch, eighteenth-century maar expresses adversativity and it functions as a coordinating conjunction (cf. English but). As such, maar is synonymous with doch and dan.¹⁰ The distribution of these three variants in van der Does’s manuscript is as follows:

                Table 5. Distribution of adversative coordinators with van der Does.

A closer examination results in a fairly different view. Generally, van der Does uses doch and dan throughout the text. In his grandmother’s letter, however, neither conjunction appears. She solely uses maar, and does so five times; see (20) for an example.

Elsewhere, van der Does reports his mother’s oral account of a trip. The passage is written by van der Does, yet it is supposedly a citation. The three adversative statements his mother expresses are introduced by doch (twice) and maar (once). Elsewhere again, when allegedly quoting a conversation, van der Does uses maar; see (21).

The last instance of the eight tokens of maar is quoted in (22).

It seems that here maar is used in indirect speech. On the basis of these findings, Table 5 can be adjusted as follows; see Table 6.

Table 6. Distribution of adversative coordinators; “(indirect) speech” refers to the explicit and implicit quotations from oral language in Van der Does’s text.

However small the absolute figures are, a pattern can be discerned: in the spoken language and in the written language of a less educated (and/or less experienced) writer, such as van der Does’s grandmother, maar is regular. In educated written language, by van der Does himself, doch and dan occur as stylistic variants instead. Sometimes, education influences speech: doch incidently appears in quoted speech. Maybe dan was never anything but a written variant and therefore, perhaps, it disappeared from Standard Dutch.¹¹

If maar is the prefered variant for less educated and/or inexperienced writers, we would expect the above- mentioned van Steensel to use it abundantly – which indeed she does, as is evident from the figures in Table 7. The total absence of dan from her language can be explained by the fact that dan is only a written variant for educated writers.

           
            Table 7. Distribution of adversative coordinators with van Steensel.

4.3 The pronouns mij, mijn and me

In Modern Standard Dutch, the oblique form of the first person singular pronoun ik is mij. In Middle Dutch, mijn is a well-attested variant (van Loey 1955:33; van de Ketterij 1980:38) and in seventeenth-century literary as well as informal texts mijn appears as well (Weijnen 1956:44; van der Sijs 2004:486), according to some especially in Hollandic texts (Hermkens & van de Ketterij 1980:89). Today, mijn is very common in dialect and not just in Holland: it is also widely used in large parts of the provinces of Zeeland, Brabant, Utrecht and Gelderland (Weijnen 1966:291; van der Sijs 2004:485). As early as the beginning of the seventeenth century, mijn was stigmatised in written Dutch, though not by grammarians, who did not comment upon the matter (van der Wal 2007:91). In eighteenth-century literary prose, mijn was certainly regarded as non-standard (see 2.2, 2.6 and 3.1). Mijn is absent in most published texts and in some unpublished texts from the eighteenth century but a high percentage (55%) of mijn was attested in a collection of letters from the years 1776-1780 written by a sailor’s wife (van der Wal 2007:91-92).

In van der Does’s and his grandmother’s texts, again a fairly regular distribution of competing forms appears. Whereas van der Does himself consistently uses mij as the oblique form, his grandmother’s letter shows considerable variation: first she writes mij four times, as in examples (23) - (26), then three times mijn, for which see (27) - (29).

Whereas printed texts and grammar books consistently present mij as the oblique form, with which (23), (24), (26) and (25) comply, mijn also appears, as object form as in (28) and (29), as well as in a PP (27). It is as if the grandmother uses mijn in speech but knows she is supposed to write mij, which she does, as in examples (23) - (26). A slip of the pen in (27), however, immediately triggers two more cases of mijn (28-29).¹²

The results so far lead to the hypothesis that inexperienced, less-educated writers use mijn in prescribed mij positions. Again I used van Steensel as a test-case. For prescribed mij she writes myn 137 times and many more times mijn (in the first letter of only 733 words mijn appears already as often as 13 times), as an object form (30) as well as after a preposition (31). In fact, she uses mij only once (32a), in an expression in which she otherwise wrote mijn/myn (32b).

But van Steensel does not only use mijn instead of mij, she also employs the (regularly) phonologically reduced form me 11 times, see (33-35). Of these 11 tokens, no less than 6 are of the not very frequent reflexive pronoun type, which means that at least in written Dutch phonological reduction is not per se related to frequency. On the contrary, the relative semantic emptiness of the reflexive pronoun, which only specifies the meaning of the verb, is a better predictor of reduction.

Another hypothesis can be formulated: the next stage – following the phonological reduction of the pronoun – would be its disappearance. The data from van Steensel support this hypothesis: (33) - (34) are not reflexive in Modern Dutch, while (35) is not even recorded in the WNT. The verb in (33) is usually non-reflexive with van Steensel. (33) and (35) show that the reflexive pronoun is not used to compensate the absence of a direct object.¹³

In what way are these observations historically related to the use of reflexives and of personal pronouns in indirect object position? Such loss of reflexivity - which may be incidental - is a fact of Modern Dutch as becomes clear from (36-37) where I put the non-realised zich in square brackets:¹⁴

In connection with the mij/mijn-variation, another important observation can be made on the basis of van Steensel’s letters. Van Steensel also adopts me as a reduced variant, not of personal or reflexive mij this time but of the possessive pronoun mij. See for instance (38) in which the contrast between the reduced possessive Me and the ungrammatical indirect object myn is illustrated. Here, frequency does seem to trigger the reduced form: of the 21 tokens, 15 are possessive pronouns followed by a noun denoting a family member, as listed in (39). Perhaps these are regularly used denotations in daily life, consequently appearing in the spoken language, in other words, these phrases are subject to phonological reduction.

From the perspective of published texts and normative grammar books, the variation of mij and mijn had come to an end in Standard Dutch by the seventeenth century, while mijn as object form survived in dialects (van der Sijs 2004:485-486). The data from eighteenth-century ego-documents discussed here suggest otherwise: the authors of these letters and diaries do not write a dialect but use some sort of general Dutch: a written, non-dialectical variant, not the standard of poets and grammarians but a so-called intended standard (Mihm 1998): a variant that allows more variation than the language of the well-educated but at the same time is stripped of explicit local or regional features. In this non-standard writing practice, mijn and mij are variants, occurring not just well into the eighteenth century. In a collection of letters and fragments from letters from Dutch immigrants into the United States from 1847-1920, the latest example dates from 1904, see (40).

Even so, what I have called ungrammatical mijn was stigmatised not just by the upper classes. The above-mentioned Knock, who was educated enough to write a full journal on his experiences in the army, was a middle-class soldier and officer (Knock 1994 [1784-1797]:8). In his journal, he consistently uses mij except for 4 irregular instances of mijn. One is from his own text (41), which is an exception to the rule, the other three are used in connection with an opponent within the army, Engelbertus Lucas. Knock cites a letter by Lucas in which mijn instead of mij appears (Knock 1994 [1784-1797]:168) but even more telling is the account of the last conversation the two had before “being friends no more” (cf. Knock 1994 [1784-1797]:177); see (42) with twice mijn.

We cannot tell whether Lucas used mijn in his speech, but the passage clearly suggests a delicate sense of stereotyping.

With regard to van Effen and Wolff & Deken, who employed ungrammatical mijn as a technique for stereotyping, we can conclude that the attribution of mijn to less educated speakers and writers is not necessarily wrong, though its precise distribution and frequency are still unclear.

4.4 Verbal inflection: zou and zoude, wou and wilde

Verbal inflection shows considerable variation in the text in question but it is not always as clear-cut as in the previous cases. In fact, variation is displayed in printed as well as in unpublished texts and it seems hard to discover a regular pattern (van der Wal 2007:92-93). I will discuss only two items; a reliable interpretation of the whole range of variants attested (compare also Table 2 and Table 4) is not yet possible, as quantitatively relevant data are simply lacking.

The preterite singular of the verb zullen is (> sulde/solde) zou – at least in Modern Dutch. It is not used in the preterite but as an auxiliary of the subjunctive mood, which has been its function for centuries. In the first and third persons singular, most of the important eighteenth-century grammarians have the longer form zoude,¹⁷ for instance Moonen (1706:156-157), de Haes (1764:61), van der Palm (1769:74) and Stijl (1778:112). Sewel mentions zoude alongside zou (1712:249), while Ten Kate only provides zoude as the regular form, considering zou a contraction of zoude (1723-I:573-574).

Likewise, ten Kate (1723-I:570) considers wou’ (sic) a contraction of woude which in its turn is a variant of wilde the past tense of will. As woude stems (> wolde) from wollen, an extinct infinitive replaced by willen, and wilde is a regular weak inflection of willen, usually wilde is considered standard while wou, frequently used in informal and/or spoken Modern Dutch, remains non-standard. Ten Kate already only mentions woude and wou’ in a passage devoted to willen and wilde, and he explicitly states that wollen is a historic form. Moonen (1706:158) has both wilde and woude. He considers wou to be a case of “afkapping” (“chopping”), his translation of apocope. Van Belle (1755) has the forms in <il> and adds that in “Straat-Taal” (“street language”) the <ou> forms appear. Table 8 shows that zoude and zou are singular, with the plural always being zouden, and that wilde and wou represent singular as well as plural forms (wilde and wilden resp. wou and wouden).

Table 8. Distribution of zoude, zou, wilde and wou in van der Does’s manuscript.

The material may be scarce but a pattern appears to exist. While the grandmother uses the so-called contracted forms, van der Does himself prefers the longer variants. Is this distribution confirmed by the van Steensel material? In other words: does van Steensel also use the more colloquial forms zou and wou and not the normatively higher esteemed variants zoude and wilde? And what about Knock, also a member of the middle class, unlike van der Does, but whose texts nevertheless show that fine sense of stereotype with regard to mij and mijn (see above, 4.3)? See Table 9.

Table 9. Distribution of zou, zoude, wou and wilde with van Steensel and Knock; van Steensel spells sou instead of zou.

As is evident from Table 9, the little better educated Knock, possibly because he was male, generates an interesting difference with respect to the conjugation of the preterite of zullen and willen. Whereas van Steensel almost exclusively uses the colloquial forms sou and wou, Knock uses the grammarians’ zoude and wilde with only one exception, and this exception occurs in a letter he quotes. The difference is not class-related as van Steensel and Knock were both middle-class figures. Van Steensel maybe even had a higher social status than Knock. Her parents, of modest origin, were social climbers and her mother towards the end of her life even obtained a noble title: vrouwe “lady” (see van Steensel 1994 [1782-1800]). Knock, on the other hand, was probably wealthy though certainly not very rich, and his carreer only went down after he had become an officer in the army (Knock 1994 1784-1797]). Next, van der Does and his grandmother both belonged to the highest circles of ’s-Hertogenbosch. The linguistic difference between van der Does and Knock, and van Steensel respectively, then, must be due to the fact that van der Does and Knock were better educated and that they had more experience in the practice of writing. Here we witness the influence of grammarians and/or other normative authors on the writing practice of individuals. This in no way implies Knock is a master of the highest level of style. In his journal, we find for instance twice the plural bennen instead of zijn “are”, which is considered “vulgar” by the grammarians, and five times kommen for komen “come”. Clearly, the results of education and of writing experience differ from time to time, from person to person, and from form to form.

5. Final remarks

Instances of code-switching in eighteenth-century texts can be used as a source for research into grammatical and stylistic norms. Literary texts, however, should be treated with the utmost caution, as it remains unclear to what extent features like non-standard, dialectic or sociolectic usage truly represent the linguistic situation; literary authors especially seem to have a strong preference for phonological and morphological features in order to mark the language of their characters, but we can nevertheless use literary texts for formulating hypotheses.

In ego-documents produced by non-professional writers instances of code-switching can be encountered as well, and these provide more reliable material for historical sociolinguistic analysis. Contrary to the literary pastiches, these show syntactic and pragmatic phenomena that throw new light on the development of written Dutch (4.1, 4.2). Also, they provide insight into the transition from spoken to written Dutch in the attempt at using an (intended) standard, as in the case of avoidance of maar, mijn, zou and wou in well-educated written Dutch (4.2, 4.3, 4.4). Thus, insight is gained into the effect of education on the linguistic output of writers.

Although the van der Does manuscript contains only a brief example of code-switching, it is of importance for a number of reasons. First, the co-existence of different styles and grammatical forms within one manuscript has strong illustrative power. Further, the variational co-existence within one social network, even within one family, demonstrates that we are dealing with social variation not caused by regional or class differences but related to education and writing experience, or perhaps age (cf. 4.2.2). Next, the manuscript is in the hand of a single author who employed the variants systematically. This suggests that the young van der Does consciously avoided certain forms, while freely using others. This is especially the case with maar, zou and wou, and maybe the use of mijn as well, and perhaps also with the syntactic items dealt with in 4.1. These probably belonged to the regular spoken Dutch of the period, but were systematically kept from written Dutch, and the data found correlate with the level of education and of writing experience.

This last observation is of importance for Koch and Oesterreicher’s concept of Sprache der Nähe “language of proximity” as opposed to Sprache der Distanz “language of distance” which has played a crucial role in recent historical sociolinguistic research (e.g. Elspaß 2005). The idea that we get the closest look we can at spoken everyday language by studying ego-documents becomes problematical as the influence of education and of writing experience, in short of linguistic norms in operation at the time, appears to be in evidence even in the prose of a soldier/officer such as Knock, and even more in that of van der Does. Their language is at least partly adjusted to certain norms, as a result of which it can be called formalised in the sense that the difference between spoken and written language approximates that between informal and formal usage. This suggests that informal written language as such did not exist. Alternatively speaking, if we hold on to the difference between language of proximity and language of distance, we have to conclude that what we at first sight would consider to be quite formal and “distanced” Dutch is in fact the appropriate style for an educated young man such as van der Does to adopt in an ego-document. Further research will have to clear these matters up.

References

Belle, J. van. 1755. Korte schets der Néderduitse spraakkonst. Haarlem: Christoph Henrich Bohn.

Bredero, G.A. 1971 [1612-1613]. Kluchten. Ed. by J. Daan. Culemborg: Tjeenk Willink-Noorduijn N.V.

Bredero, G.A. 1974 [1617]. Spaanschen Brabander. Ed. by C.F.P. Stutterheim. Culemborg: Tjeenk Willink-Noorduijn.

Brinks, H.J. 1978. Schrijf spoedig terug. Brieven van immigranten in Amerika 1847-1920. The Hague: Uitgeverij Boekencentrum b.v.

Burridge, K. 1993. Syntactic Change in Germanic. Amsterdam & Philadelphia: John Benjamins.

Bybee, J. & P. Hopper (eds.) 2001. Frequency and the Emergence of Linguistic Structure. Amsterdam & Philadelphia: John Benjamins.

Daan, J. 1995. “Sociolecten in de achttiende eeuw”. In: J. Cajot, L. Kremer & H. Niebaum (eds.), Lingua Theodisca. Beiträge zur Sprache- und Literaturwissenschaft. Jan Goossens zum 65. Geburtstag. Münster: Zentrum für Niederlande-Studien. 263-269.

van Dale = Van Dale Groot woordenboek van de Nederlandse taal. 14th ed. (2005). Utrecht & Antwerp: Van Dale Lexicografie B.V.

Eck, O. van. 1998 [1791-1797]. Het dagboek van Otto van Eck (1791-1797). Ed. by A. Baggerman & R. Dekker. Hilversum: Verloren.

Effen, Justus van. 1984 [1731-1735]. De Hollandsche Spectator. Ed. by P.J. Buijnsters. Deventer: Sub Rosa.

Effen, Justus van. 1999 [1733]. De Hollandsche Spectator. Aflevering 151-195: 7 april 1733 - 7 september 1733. Ed. by M. de Niet. Leiden: Astraea.

Elspaß, S. 2005. Sprachgeschichte von unten. Untersuchungen zum geschriebenen Alltagsdeutsch im 19. Jahrhundert. Tübingen: Niemeyer

Grosse, S. 1990. “Zu Syntax und Stil der deutschen Sprache des 19. Jahrhunderts”. In: A. Betten & C.M. Riehl (eds.), Neuere Forschungen zur historischen Syntax des Deutschen. Tübingen: Max Niemeyer Verlag. 300-309.

Gussenhoven, C. 2007. “Wat is de beste transcriptie voor het Nederlands?”. Nederlandse Taalkunde 12. 331-350.

Haes, F. de. 1764. De nagelaten gedichten, en Nederduitsche spraekkunst. Amsterdam: Peter Meijer.

Hanou, A.J. 1981. “Dutch periodicals from 1697 to 1721: in imitation of the English?”. Studies on Voltaire and the eighteenth Century 199. 187-204.

Hermkens, H. & C. van de Ketterij. 1980. Grammaticale interpretatie van zeventiende-eeuwse teksten. Instructiegrammatica. Groningen: Wolters-Noordhoff.

Hol, A.R. 1941. “Het prefix in het verleden deelwoord”. Tijdschrift voor Nederlandse Taal- en Letterkunde 60. 249-293.

Horst, J. van der. 1990. “Is weglating van het voegwoord dat typisch voor 19de-eeuwse vouwen?” In: J. den Besten et al. (eds.), Vragende wijs. Vragen over tekst, taal en taalgeschiedenis. Bundel aangeboden aan Leopold Peeters. Amsterdam & Atlanta, GA: Rodopi. 95-103.

HS = De Hollandsche Spectator. See van Effen 1984 [1731-1735], van Effen 1999 [1733].

Huygens, Constantijn. 1987 [1653]. Trijntje Cornelis. Vol. 1. Ed. by H.M. Hermkens. Utrecht: HES.

Kate, L. ten. 1723. Aenleiding tot de kennisse van het verhevene deel der Nederduitsche sprake. 2 vols. Amsterdam: Rudolph and Gerard Wetstein.

Ketterij, C. van de. 1980. Grammaticale interpretatie van middelnederlandse teksten. Instructiegrammatica. Groningen: Wolters-Noordhoff.

Knock, A.J. 1994 [1784-1797]. Uit Lievde voor Vaderland en Vrijheid. Het journaal van de patriot Arie Johannes Knock over de periode 1784 tot 1797. Ed. by P.M. Peucker & J.P. Sigmond. Hilversum: Verloren.

Loey, A. van. 1955. Middelnederlandse spraakkunst. I. Vormleer. 2nd ed. Groningen & Djakarta: J.B. Wolters & Antwerp: De Sikkel.

Loey, A. van. 1965. Middelnederlandse spraakkunst. II. Klankleer. 4th ed. Groningen: J.B. Wolters & Antwerp: De Sikkel.

Mattheier, K.J. 1986. “‘Lauter Borke um den Kopp’. Überlegungen zur Sprache der Arbeiter im 19. Jahrhundert”. Rheinische Vierteljahrsblätter 50. 222-252.

Mattheier, K.J. 1990. “Formale und funktionale Aspekte der Syntax von Arbeiterschriftsprache im 19. Jahrhundert”. In: A. Betten & C.M. Riehl (eds.), Neuere Forschungen zur historischen Syntax des Deutschen. Tübingen: Max Niemeyer Verlag. 286-299.

Megen, N. van. 2002. “Dan en want: hun functie en betekenis in zeventiende-eeuws taalgebruik”. neerlandistiek.nl 02.02. <www.neerlandistiek.nl>

Mihm, A. 1998. “Arbeitersprache und gesprochene Sprache im. 19. Jahrhundert”. In: D. Cherubim, S. Grosse & K.J. Mattheier (eds.), Sprache und bürgerliche Nation. Beiträge zur deutschen und europäischen Sprachgeschichte des 19. Jahrhunderts. Berlin & New York: De Gruyter. 282-316.

Moonen, A. 1706. Nederduitsche spraekkunst. Amsterdam: François Halma.

Nylöe, J. 1723. Aanleiding tot de Nederduitsche taal. 3rd ed. Amsterdam: Gerard Onder de Linden.

Palm, K. van der. 1769. Nederduitsche spraekkunst, voor de jeugdt. Rotterdam: Reinier Arrenberg.

Rutten, G. 2008. “Standaardvariatie in de achttiende eeuw. Historisch-sociolinguïstische verkenningen”. Nederlandse Taalkunde 13. 34-59.

Rutten, G. (in prep.). “From adverb to conjunction and back. The case of Dutch dan”.

Sasse van IJsselt, A.F.O. van. 1900. “Eenige bladzijden uit de geschiedenis van den patriottentijd te ’s Hertogenbosch”. Taxandria VII. 234-249.

Schönfeld, M. 1954. Historische grammatica van het Nederlands. 5th ed., by A. van Loey. Zutphen: W.J. Thieme & Cie.

Schutter, G. De. 1999. “Het Antwerps. Een schets van zijn evolutie tussen 1898 en 1998”. In: J. Kruijsen & N. van der Sijs (eds.), Honderd jaar stadstaal. Amsterdam & Antwerp: Contact. 301-315.

Sewel, W. 1712. Nederduytsche spraakkonst. 2nd ed. Amsterdam: Robert Blokland.

Sijs, N. van der. 2004. Taal als mensenwerk: het ontstaan van het ABN. The Hague: Sdu.

Steensel, C. van. 1994 [1782-1800]. “Het is of ik met mijn lieve sprak”. De briefwisseling tussen Jean Malherbe en Christina van Steensel, 1782-1800. Ed. by A. Dik & D. Helmers. Hilversum: Verloren.

Stijl, K. 1778. Beknopte aanleiding tot de kennis der Spelling, Spraakdeelen, en Zinteekenen van de Nederduitsche taal. 2nd ed., by L. van Bolhuis. Groningen: Jan Oomkens.

Vandenbussche, W. 2007. “‘Lower class language’ in 19th century Flanders”. Multilingua 26. 279-290.

Verdenius, A.A. 1946. Studies over zeventiende eeuws. S.l.

Vermaas, J.A.M. 2002. Veranderingen in de Nederlandse aanspreekvormen van de dertiende t/m de twintigste eeuw. Utrecht: LOT.

Vliet, A.P. van. 2006. “Een vriendelijcke groetenisse”. Brieven van het thuisfront aan de vloot van De Ruyter (1664-1665). Franeker: Uitgeverij Van Wijnen.

Vooys, C.G.N. de. 1951. “Opmerkingen over de taal van Justus van Effen’s Hollandsche Spectator”. Nieuwe taalgids 44. 75-78.

Vooys, C.G.N. de. 1970. Geschiedenis van de Nederlandse taal. Groningen: Wolters-Noordhoff.

Wal, M.J. van der. 1992. Geschiedenis van het Nederlands. With C. van Bree. Utrecht: Spectrum.

Wal, M.J. van der. 2007. “Eighteenth-century linguistic variation from the perspective of a Dutch diary and a collection of private letters”. In: S. Elspaß, N. Langer, J. Scharloth & W. Vandenbussche (eds.), Germanic Language Histories “from below” (1700-2000). Berlin & New York: Walter de Gruyter. 83-96.

Weijnen, A.A. 1956. Zeventiende-eeuwse Taal. 2nd ed. Zutphen: W.J. Thieme & Cie.

Weijnen, A.A. 1966. Nederlandse dialectkunde. 2nd ed. Assen: Van Gorcum etc.

Weiland P. 1805. Nederduitsche Spraakkunst. Amsterdam: Johannes Allart.

Winkel, J. te. 1924. De ontwikkelingsgang der Nederlandsche letterkunde V. Geschiedenis der Nederlandsche letterkunde van de Republiek der Vereenigde Nederlanden (3). 2nd ed. Haarlem 1924: De erven F. Bohn.

Winkler, Johan. 1874. Algemeen Nederduitsch en Friesch dialecticon. Vol. 2. The Hague: Martinus Nijhoff.

WNT =Woordenboek der Nederlandsche Taal.

Wolff, B & A. Deken. 1782. Historie van mejuffrouw Sara Burgerhart. The Hague: Isaac van Cleef.

Wolff, B & A. Deken. 1784-1785. Historie van den heer Willem Leevend. 8 vols. The Hague: Isaac van Cleef.