Bildts is a speech variety spoken by around 10,000 persons as a first or second language in the province Fryslân, in the north of the Netherlands. It is commonly claimed to be a dialect of Dutch containing some Frisian loan words. This article provides an analysis of Bildts based on a comparison of Bildts with its source languages: Frisian on the one hand and specific dialects of the province of South Holland on the other hand. It argues that Bildts combines a core lexicon mainly derived from Hollandic dialects with a grammar mainly derived from Frisian. However, the core lexicon also contains some Frisian words and the grammar has to some extent been levelled. The precise mixture is not easily described, let alone accounted for, in most models of language contact. Our approach combines sociological-historical information with linguistic factorisation in order to arrive at a plausible view of how Bildts came into existence. It is argued that Bildts is a mixed language, comparable to better known cases such as Ma’á, spoken in Tanzania, and that the specific nature of the mix involved the mutual accommodation of two groups of speakers: a group of mother tongue speakers of Frisian who acquired Bildts as a second language and a group of balanced bilingual speakers of Bildts and Frisian.
1.1 Bildts as a mixed language
This article presents a sociolinguistic analysis of Bildts, a language variety spoken in the rural municipality of het Bildt in Fryslân (Dutch: Friesland) in the North of the Netherlands. In 1505, dykes were built around an alluvium, following an agreement between several Dutch lords and duke George of Saxony. These dykes were built by Hollandic labourers. The independent development of Bildts may be dated to the timeframe of the genesis of this polder.
Present-day Bildts is generally considered to be a mixed Frisian-Dutch dialect (Winkler 1874: 488; Kloeke 1927: 74; Daan and Blok 1969: 36),  but the exact nature of the mix has not been specified. There has been a controversy about this and about the origin of Bildts in older dialect geographical sources (see Section 4.1). These sources generally neglected the investigation of syntactical and morphological features, as they focused almost exclusively on idiosyncratic vocabulary and on the development of Old Germanic phonemes in words which had been preserved. As a result, they arrived at the conclusion that the dialect sounds Dutch, but shares a substantial number of words with Frisian as well. The 1970s saw an increase in attention for morphology and syntax. This led, among other things, to sharper analyses of mixed languages, especially in those cases in which the two languages involved were genetically unrelated. An example of this is Ma’á, which uses a Cushitic core lexicon in a Bantu grammatical matrix, see Mous (2003). Such a split between core lexicon and grammar is not uncommon. It is typical of mixed languages as described by Thomason (2001) and Matras (2003), among others. Although Bildts is usually referred to as a dialect of Dutch, its genesis has been described as a “blending” of Dutch and Frisian in the past by Buwalda (1964: 8). It seems unlikely, however, that he meant to confront the case of Bildts with currently-known cases of language mixing when speaking of languages “blending”. Koldijk (2004: 191) also speaks of a Frisian/Hollandic “mixed language” in his conclusion on Bildts, although he does not specify what he means by this.
Studies on mixed languages have hitherto focused on cases in which the source languages were genetically unrelated or, at least, not very closely related. Naturally, though, the linguistic and social processes underlying the formation of mixed languages are not restricted to contact between genetically unrelated languages. The source languages of Bildts are Hollandic dialects  and Frisian, and these are genetically quite close, both belonging to the subgroup of West Germanic languages spoken in the Dutch coastal area. Recently, the grammar-vocabulary split has been noticed for Bildts in Hoekstra and Van Koppen (2001a, 2001b), and a similar split has been observed for Stedsk ‘Town Frisian’, the dialect spoken in a number of towns in the province Fryslân (Van Bree and Versloot, 2008).  Hoekstra and Van Koppen, however, focused mainly on linguistic aspects of Bildts without paying much attention to sociolinguistic and historical aspects. The aim of the present article is to reassess the linguistic character and social history of Bildts within a framework of language mixing, and to present Bildts as a potential case study for mixed languages.
1.2 Present-day Bildts
Bildts is spoken in and around the historical polder and present-day municipality of Het Bildt. It has some 10,000 speakers. The provincial administration considers Bildts to be a regional language, different from both Frisian and Dutch (Provinsje Fryslân 2014). The municipality of Het Bildt has filed a request with the Ministry of the Interior and Kingdom Relations on 19 February 2016 for the Bildts language to be officially recognised as a regional language under Part II of the European Charter for Regional or Minority Languages (Krol and Goeman 2016). The provincial administration supports this application. Bildts speakers consider Frisian an invasive language, and consider themselves threatened by it (Koldijk 2004: 189–190).
Bildts is stable in terms of grammar and lexicon, which has been codified over the course of the 20th century. Nowadays, in terms of native speakers, Bildts is a minority language in Het Bildt, with 47 percent of its population speaking Frisian as a first language, 28 percent speaking Bildts as a mother tongue, 21 percent speaking Dutch and 4 percent speaking another language as a first language (Provinsje Fryslân 2014: 15).  Mother tongue speakers of Frisian also speak Dutch and can be considered balanced bilinguals Frisian-Dutch. Bilingualism and trilingualism is commonplace: just over half of the inhabitants of Het Bildt speak Bildts on a daily or near-daily basis, and a similar number indicate speaking Bildts very well (Provinsje Fryslân 2014: 14, 16). This implies that Bildts is acquired as a second language by the majority of its speakers. Intergenerational transmission between native speakers of Bildts is strong, and it continues to serve as a community language at the local level, although its use in the public domain is under pressure (Van Sluis 2015: 26). 
Since Het Bildt is completely surrounded by Frisian-speaking areas, Bildts has been in contact with neighbouring Frisian dialects throughout its existence. Moreover, the national language of the Netherlands, Standard Dutch, has affected all dialects, including Bildts. As we will see in Section 2, many Bildts linguistic features may be traced back to either South Hollandic dialects, where its settlers originated from (Koldijk 2004: 211) or to Frisian. This linguistic analysis is corroborated by contemporary historical documents (Section 3), which make it clear that the population of Het Bildts is of mixed Frisian and South Hollandic origin. This provides external evidence for our claim that Bildts is a mixed language, displaying the characteristics of mixed languages mentioned by Thomason (1997b: 80), as will be further explained in Section 4. Section 5 concludes our article.
2 Linguistic comparison of Bildts with its source languages
We will proceed with a systematic comparison of Bildts with its source languages, dealing successively with the lexicon, phonology, nominal morphology, verbal morphology and syntax. The bulk of written (and spoken) Bildts dates from the 20th century. As for South Hollandic, the same applies. As a result, we will compare 20th century Bildts with 20th century Frisian and 20th century South Hollandic, and we will rely on knowledge of the historical development of Frisian and Dutch in order to arrive at a well-motivated interpretation of that comparison.
A South Hollandic dialect of the 20th century has been described in Lafeber (1967). In addition, we can rely on dialect surveys and atlases on phonology (Goossens et al. 1998; Goossens et al. 2000), morphology (De Schutter et al. 2005; Goeman et al. 2008) and syntax (Barbiers et al. 2005; Barbiers et al. 2008). An overview of Bildts grammar and a comparison with several Hollandic dialects and Frisian is given by Koldijk (2004). A summary of his findings on morphology, syntax and phonology is given here, along with findings by Hoekstra and Van Koppen (2001a, 2001b), Hoekstra (2002).
Recognisably Frisian words are jimme ‘you’, and dêr ‘there’. A word like goed ‘good’ is written the same in Frisian and Dutch, but its pronunciation is decidedly Frisian as [ɡu:ət], with a voiced velar stop, as opposed to Hollandic [xut], with a voiceless velar fricative. The voiced velar stop does not occur in Hollandic dialects (and Dutch), except in loans like ‘goal’ and as allophone in regressive assimilation. Recognisably Hollandic words are an ‘to’ and doen ‘do’. Other words may descend from both Frisian and South Hollandic, but this cannot be decided since they are identical in the two languages, such as en ‘and’, op ‘on’, te slaan ‘to hit’. This partial overlap in lexicon between South Hollandic and Frisian obscures the mixed character of Bildts. It is easier to identify a language as a mixed language when the source languages are genetically far apart, as in the case of Michif and other well-known cases from the literature.
Another obviously mixed sentence is given below:
In this sentence, the stems of the words feul ‘much’, loofd ‘believed’, wete ‘know’, and wouste ‘wanted’ may unambiguously be traced back to a South Hollandic origin. Hest ‘have’ and doe ‘then’ are without any doubt Frisian words. Problematic is the word dou ‘you’. It is not Frisian, which has do. It is not present-day South Hollandic either, but we know that this word occurred in the form dou in earlier stages of South Hollandic, and it may still have been in use around 1500 (Schönfeld 1970: 137). Similarly, we cannot claim independently that the -st ending comes from Frisian. Although it is no longer current in either South Hollandic or Standard Dutch, it was probably still in use around 1500, since the ending and the pronoun form a syntactic tandem due to verb agreement. Likewise, the form loofd ‘believed’ lacks the (Standard) Dutch participial prefix ge-, mirroring Frisian leau-d rather than Dutch ge-loof-d, which may or may not have been around in 1500s South Hollandic dialects. However, all of these forms reveal that the development of Bildts subsequent to its introduction to Frisian environment has been such that it converges with Frisian, and not with Dutch (let alone South Hollandic), as far as verbal morphology is concerned. As a result, Bildts verbs regularly show Frisian flection, while displaying South Hollandic roots. Thus South Hollandic remnants are especially found in the primary lexicon. The primary lexicon contains words belonging to the sphere of everyday activities and experiences (Van Bree 1996: 34). Thus, words for ‘do’, ‘much’, ‘believe’ and ‘want’ all have South Hollandic roots. As a result, the Bildts grammatical matrix appears mostly Frisian while its core lexicon appears mostly South Hollandic on a synchronic level.
Here we limit ourselves to Bildts items which are recognisably either Frisian or (South) Hollandic, and which belong to the Swadesh list (Swadesh 1971). The Swadesh list contains 207 items, comprising both the original 200-word Swadesh list as well as 7 items only found in the later 100-word Swadesh list. The Bildts items on this list were systematically compared to Frisian and South Hollandic. The overwhelming majority of words is South Hollandic in etymology. Some words on this list may be traced back to South Hollandic as well as Frisian, but 88 words are still found to be South Hollandic in origin as opposed to 9 certain Frisian etymologies. Notable exceptions to the South Hollandic lexicon are the words for ‘father’ and ‘mother’. These are hait and mim, and may be traced back to Frisian heit and dialectal Frisian mim, respectively.  Other exceptions also relate to intimate domains, such as body parts, bodily functions or subjective terms, e. g. the Bildts word for ‘bad’ is min, and mirrors Frisian min rather than South Hollandic slecht. A salient feature of Bildts lexicon is bra ‘very’, which is South Hollandic in origin, and is found in neither (Standard) Dutch nor Frisian. Another interesting example is pinemond ‘pain in the mouth > toothache’. This calque is Frisian in its semantic and idiomatic outlook, except that the Frisian word mûle ‘mouth’ has been replaced with Bildts and South Hollandic mond ‘mouth’.
Bildts is principally Hollandic not only in its primary lexicon, but also in its secondary lexicon, i. e. terms that are exclusively used in a local setting. An example of this is agricultural terminology. Koldijk gives an overview of typical South Hollandic agricultural terminology (2004: 158), e. g. plok ‘hay rack’, til ‘attic of a barn’, and kússy ‘baby calf’. South Hollandic agricultural terminology is mostly restricted to elements inside farm buildings. Frisian, by contrast, has had lexical influence in agricultural terms referring to larger concepts, such as plaats ‘barn’ and hiem ‘yard’.
Seven words in the English-language Swadesh list have two Bildts equivalents, one of which may be traced back to South Hollandic and another to Frisian. This could be due to additive borrowing, but may also point towards remnants of an earlier parallel lexicon. If Bildts ever had parallel lexica, the South Hollandic-origin lexicon would serve as an alternative register for Frisian. An example of such a pair of synonyms is bonke ‘bone’, which stems from Frisian and exists side by side with South Hollandic been in Bildts. Bildts similarly has two words for ‘to vomit’: brake (South Hollandic: braken) beside spoie (Frisian spuie).
In functional categories, Frisian-origin words are more common. The Bildts paradigm of personal pronouns corresponds exactly with that of neighbouring spoken Frisian, with two exceptions: the oblique first-person plural pronoun ôns ‘us’, South Hollandic ons, Frisian ús, and the third person singular masculine, if used as a clitic following a verb: Bildts -ie, South Hollandic -ie, Frisian -er (Hoekstra et al. 2001a). Demonstratives vary in origin, even for semantically closely related words: the Bildts word hier ‘here’ stems from South Hollandic hier while the word dêr ‘there’ probably stems from Frisian dêr, although dêr cannot be completely ruled out for 16th century South Hollandic. Similarly, interrogative wie ‘who’ corresponds with South Hollandic wie, while wêr ‘where’ corresponds with Frisian wêr.
Some Bildts words exist in both Frisian and Dutch, but retain exclusively South Hollandic semantics. An example is the use of al as a conditional ‘if’ (Frisian as, Dutch als). In Frisian and Dutch, this conjunction typically carries a concessive meaning. Hoekstra and Van Koppen (2001a) regard this form as evidence of a Hollandic substrate.
Bildts phonology is principally South Hollandic, but follows some Frisian restrictions. Bildts is like Frisian as the Dutch word-initial voiceless stop /x/ does not occur. Instead, it has the voiced velar stop /ɡ/ (Koldijk 2004: 77–78, 114). Furthermore, the adoption of some Frisian loanwords in Bildts has led to the incorporation of a quantitative distinction between long /i:/ and short /i/ (Koldijk 2004: 76). Quantitative distinction is not adopted from Frisian wholesale, however. Otherwise, the vowel inventory of Bildts roughly reflects that of South Hollandic, but it is more similar to Frisian in its phonetic realisation (Koldijk 2004: 76). An example is Bildts /ɛi/, which is realised as [ɛi] in Hollandic. In Bildts, this diphthong is realised as [ɔi], reflecting the pronunciation of neighbouring Frisian dialects (Koldijk 2004: 106). An example of non-convergence with Frisian phonology is that Bildts has not adopted Frisian breaking. Breaking causes some historically long vowels to be pronounced as diphthongs ending in schwa. These vowels are a salient feature of present-day Frisian phonology, and typically occur in stressed position. Breaking developed in Frisian in the late 17th century, and spread through the language slowly (Van der Meer 1985: 49). On a phonetic level, Bildts has a similar system of diphthongs ending in schwa (Koldijk 2004: 75–76).
2.4 Inherent flection
A striking aspect of Bildts inherent flection is its diminutive system. Hoekstra and Van Koppen (2001b) state that a diminutive system is at the interface between a language’s morphology and its lexicon. Diminutives in Bildts are South Hollandic in origin, but differ from South Hollandic. Rather, they correspond closely with the diminutives used in the local dialects of South Holland (Hoekstra et al. 2001b; Koldijk 2004: 132–143). The Bildts diminutive system is rather complex, and has four different endings depending on the phonological environment of the ending. This is shown in the Table 1, based on the descriptions and examples given in the references above:
|-y /i/ is used after /ŋ/, /ft/, /xt/, /st/, /f/, /γ/, /s/, /k/, /p/ and /b/. After fricative-/t/ clusters, /t/ may be deleted and the fricative may be reduplicated.|
|-py /pi/ is used after /m/|
|-(e)chy /xi/ is used after /jə/, and after /m/, /n/, /ŋ/, /r(ə)/, /b(ə)/, /p/ if these sounds are preceded by a short vowel|
|-tsy /tsi/ is used after vowels, liquid consonants, /d/, and /t/|
As can be seen, some phonological environments are represented more than once in this overview. This is due to the fact that some words may have more than one diminutive form, e. g. Bildts boom-py exists beside boom-tsy ‘little tree’. The ending -tsy differs from its South Hollandic counterpart -tjie. This difference in ending may be seen as an example of phonetic convergence of Bildts towards Frisian: Dutch lacks the phonotactic combination /ts/ in syllable-initial position, except in some loan words.
Frisian influence is found in Bildts derivational morphology. An example of this is the suffix -ens, which may be used to derive nouns from adjectives. It is found in Bildts and in Frisian, but it does not occur in South Hollandic. This suffix -ens exists side by side with the suffix -hyd, which is found both in Bildts, Frisian and South Hollandic (Koldijk 2004: 143). These two suffixes often compete for the same root. Thus both Frisian and Bildts display morphological doublets like eerlikens/eerlikhyd ‘honesty’, failigens/failighyd ‘safety’, and so on.
Another example of inherent flection is plural formation. Koldijk notes that South Hollandic and Frisian equally contributed to Bildts plural formation (2004: 130–131). Thus, Bildts displays plurals in -s in words ending in a schwa, just like certain South Hollandic dialects. Some examples illustrating this are Bildts karres ‘carts’, bolles ‘bulls’, which in Frisian have an n-plural: karren ‘carts’, bollen ‘bulls’.
A South Hollandic element in the inherent flection of Bildts is that it has only one class of verbal flection. Frisian, by contrast, has two classes for weak verbs: those whose infinitive ends in -e and those whose infinitive ends in -je. Bildts has not taken over the -je-category (Buwalda 1960; Koldijk 2004: 128–129), but this may be due to lack of infinitives in -je(n) in South Hollandic, meaning there were no lexemes that could take these endings.  As such, what seems a reorientation towards South Hollandic may also be explained as a consequence of relexification with Dutch verbal stems.
Another South Hollandic element in Bildts inherent flection is the ablaut pattern of most strong verbs. An example is the Bildts verb blieve – bleef – bleven ‘stay-stayed-stayed’, which is quite different in Frisian: bliuwe – bleau – bleaun ‘stay-stayed-stayed’. Hoekstra and Van Koppen (2001a), Koldijk (2004: 128) note that irregular verbs that form their past and perfect tense by alternating the root vowel tend to follow South Hollandic alternation patterns.
2.5 Contextual flection
Since Bildts has no case system, contextual flection is mostly found in verbal morphology and syntax. Some phenomena are shared between Frisian, Bildts and South-Hollandic alike. For example, the Bildts monosyllabic verbs gaan ‘go’, staan ‘stand’, doen ‘do’, sien ‘see’, and slaan ‘hit’ retain an ending in -n in the first person singular, e. g. ik gaan ‘I go’. The same applies to the Frisian monosyllabic verbs, as exemplified by ik gean. Koldijk (2004: 129) notes that this ending is also found in several dialects of South and North Holland. Similarly, the Bildts second person singular ending in -st is found in Frisian, but not in present-day South Hollandic. In Middle Dutch, however, -st still occurred, and it disappeared in written Dutch later on. It may have existed for a longer period in spoken Dutch, however (Hoekstra et al. 2001a). These considerations suggest that phenomena such as the above do not provide unequivocal evidence that Frisian played a major role in the genesis of Bildts, though it is interesting to note that Frisian and Bildts both retained certain linguistic phenomena which are no longer found in South Hollandic.
However, Bildts and Frisian exhibit a phenomenon which South Hollandic dialects do not exhibit. This phenomenon is characteristic of the Frisian language family (Hoekstra 2012), which includes West-Frisian as spoken in The Netherlands, East Frisian in Germany (of which Saterlandic Frisian is the sole survivor) and North Frisian, spoken in Germany near the border with Denmark. All Frisian languages and their dialects have a complex system of two infinitive morphemes (-e and -en in Frisian as spoken in The Netherlands). The choice of morpheme depends on its syntactic context, more specifically on the element selecting the infinitive.
The following examples are taken from Hoekstra and Van Koppen (2001a: 97–99), and show the syntactic distribution of both infinitive endings. In all cases, Bildts corresponds exactly with Frisian with respect to the contextually determined distribution of -e and of -n. Below, the major syntactic contexts are listed, accompanied by an example from Bildts.
Infinitives governed by a modal verb take an ending in -e:
When the infinitive is preceded by preposition te ‘to’, the ending in -en is used.
Infinitives without any syntactic context take the -e ending.
After verbs of perception, infinitives in -en are found:
Infinitives governed by causative late ‘leave’ take the infinitive ending in -e:
Nominalised infinitives take the -en ending:
If an infinitive precedes auxiliary doen ‘do’, it ends in -e:
In spoken Dutch and in South Hollandic, only one infinitive suffix exists: -e, which existed side by side with the variant pronunciation in -en. The choice for either -e or -en has never had a syntactic motivation in Dutch as far as infinitives are concerned.
Hoekstra and Van Koppen (2001a) note that Southern Holland has never had such a distribution of infinitives, so this feature cannot have been brought over by settlers. Moreover, they state, such a complex feature cannot have been the result of language contact through trade relations. Rather, this complex system points towards a Frisian substrate, and is evidence that native Frisian speakers had input in the grammar of Bildts. 
In some contexts, Frisian uses a past participle where Dutch would use a so-called Ersatzinfinitiv. The phenomenon in question is also referred to as Infinitivus pro Participio or IPP (Haeseryn et al. 1997: 954–958), since the infinitive shows up in contexts in which logically a past participle is expected. Bildts follows the Frisian pattern yielding a past participle, which is found in the following example:
The infinitive would have been late in Bildts.
In Frisian, second person singular do ‘you’ may be dropped, leaving the verbal ending as the only clue for person and number. This pronoun may also be dropped in Bildts, as is shown in the following example (Buwalda 1948: 43):
This behaviour of the Bildts second person singular corroborates the impression that the verbal morphological material in Bildts is in fact Frisian in origin, even where some aspects of it may also be traced back to some historical and dialectal varieties of Dutch based on form alone. The development towards pronoun dropping is a shared Frisian-Bildts innovation.
When Bildts interrogative pronouns are used as relative pronouns, epenthetic ‘t is added, like in Frisian, e. g. Bildts wie’t ‘who’ (rel.), Frisian wa’t. This is a specifically Frisian development, not attested in South Hollandic dialects (Koldijk 2004: 178–179). It only became common in Frisian during the 19th and 20th century (Van Coetsem 1960: 327; Van der Woude 1960), so insertion of ‘t is a development shared by Bildts and Frisian, and probably arose as a result of widespread bilingualism.
2.6 Word order
Word order is similar to contextual flection in that the relative position of an individual word within a clause is conditioned by its relationship to other words. In his chapter on Bildts syntax, Koldijk (2004: 181) wonders whether Bildts is primarily a Frisian dialect. Bildts word order indeed patterns like Frisian, which is apparent in verb clusters in subordinate clauses, for example.
In South Hollandic, the positions of the participle and the auxiliary verb could be reversed, yielding heeft gezegd ‘has said’ (Koldijk 2004: 169), though the order found in Bildts and Frisian is also allowed (Barbiers et al. 2008: 184.108.40.206).
In main clauses, the placement of verbs within a predicate also patterns like Frisian (Koldijk 2004: 165):
In Hollandic dialects, the order of the infinitives would typically be reversed: horen praten. In Frisian, the word order given above is obligatory, as it is in Bildts.
2.7 Concluding remarks
Examination of the Swadesh list led us to conclude that most of the core lexicon of Bildts is Dutch. Inherent flection in Bildts is largely Dutch in origin, but has converged with Frisian to some degree. Contextual flection of Bildts is complicated: Bildts verbal flection corresponds exactly with Frisian on a synchronic level. However, some of these morphemes could individually be explained as remnants of various dialects of earlier Dutch. It is therefore possible, but improbable, to assume the Bildts paradigm of verbal flection is a merger of Dutch dialects which happens to exactly mirror Frisian. Even if this were the case, the result is nevertheless a Frisian grammatical matrix in verbal flection, which explains why pronoun dou ‘you’ may be dropped. In addition, the distribution of Bildts infinitive morphemes corresponds exactly with Frisian. Similarly, word order behaves like contextual flection as the relative position of a word within a clause is determined by its relation to other words rather than by its lexical semantics, and therefore patterns like Frisian.
What we see, then, is a South Hollandic lexicon, mixed South Hollandic/Frisian inherent flection, and Frisian contextual flection and word order. Due to its lexicon, Bildts is traditionally classified as a phylogenetically Dutch dialect. However, many morphological and syntactic features are Frisian in origin, to the extent that one may speak of a Frisian grammatical system in a Dutch lexical jacket.
3 Origins of the settlers of Het Bildt and subsequent immigration
Early 16th century documents mention migration of Hollandic, Zeelandic, Brabantian and other settlers towards Het Bildt (Kuiken 2013: 1). These settlers set in motion the development of Het Bildt and the independent development of Bildts. Personal names, historical documents and the linguistic character of Bildts all corroborate that a significant part of these migrants came from Southern Holland, roughly from the rural area around The Hague and Rotterdam. Less substantial migration came from Westfriesland in North Holland and the (now former) island of Wieringen, as has been registered in historical documents (Koldijk 2004: 44–45). The labourers who built the dykes did not play a significant role in populating Het Bildt. Instead, investors in the project contracted migrants from Holland, who in turn brought about a subsequent wave of chain migration by bringing over their families and relatives (Kuiken 2013: 63–68). Thus the settlers mainly had a Hollandic origin. This ties in with the fact reported in Section 2 that most Dutch elements in Bildts can be traced back to rural dialects used in Southern Holland.
Demographic sources in Het Bildt document a severe drop in Frisian personal names from 89 percent to 27 percent between 1506 and 1557 (Kuiken 2013: 61). This period lasted at most until the beginning of the Eighty Years War with Spain. Some sort of “balance” between roughly 25 percent Frisians and 75 percent non-Frisians existed from the 1550s onwards (Kuiken 2013: 62). Labour migration from neighbouring Frisian villages has been continuous since then. Initially, economic ties between the settlers and their homeland in Holland were strong. In the early years following its settlement, grain produced in Het Bildt was sold in Amsterdam. During the Eighty Years War (1568–1648), however, access to this trading hub was disrupted, causing Het Bildt to trade with neighbouring Frisian towns and villages instead (Kuiken 2013: 77–88). Thus Hollandic speakers in Het Bildt were cut off from their ancestral homeland. Significant migration of Hollandic speakers towards Het Bildt was a one-off occurrence during the century following the reclamation of this land. Labour immigration of Frisian speakers from neighbouring land, by contrast, has been continuous and continues until this day.
Throughout most of the 18th century, Het Bildt was relatively poor, and the percentage of paupers in Het Bildt was higher than in neighbouring areas, or Friesland in general. Between 1689 and 1744, the overall population grew by 4 percent, while the amount of paupers grew by 26 percent in the same time frame. Moreover, Het Bildt took particularly long to recover from the economic downturn of the early eighteenth century: the 1744 census shows that 13 percent of the Bildts population was pauper, as opposed to 8 percent for Fryslân as a whole (Kuiken 2013: 138–139). This poverty lasted for much of the Bildts history, and was exacerbated by absentee landownership, which caused economic surpluses to be extracted from Het Bildt. Because of this poverty, Het Bildt could not maintain a splendid isolation from the surrounding Frisian area. Instead, it provided cheap seasonal labourers for farmers in the surrounding Frisian area.
Het Bildt is a traditionally agricultural area, although mechanisation has led to a shift towards agricultural industry in the twentieth century (Koldijk 2004: 14). As a result, Het Bildt, being a small area, depends on the surrounding Frisian area for trade in general. Furthermore, the provincial authorities were and are located in Leeuwarden, outside Het Bildt. Likewise, the cattle market used to be held in Leeuwarden, the capital of the province. The Bildt region itself does not feature any cities. Within this context, competence in Frisian was an economic necessity for most Bildts speakers. This is corroborated by the adoption of Frisian agricultural terminology, as seen in Section 2.
To sum up, Het Bildt originally held a majority of South Hollandic settlers and a minority of Frisian settlers. In addition, there has been a steady trickling of immigration of Frisians throughout the centuries. It is likely that the population of Het Bildt was bilingual, mainly for two reasons. First, the population started out as a mix of South Hollandic and Frisian settlers who were in daily contact. Second, bilingualism was enhanced by the economic and political dependence of Het Bildt on the surrounding Frisian area. Seeing that the two languages are genetically very close, it would not be difficult for speakers of Frisian to acquire the language of the newcomers and vice versa. The presence of a mixed bilingual population thus suggests that Bildts may be analysed as a mixed language. This hypothesis is further explored below.
4 Mixed population, mixed language
4.1 Previous analyses
Presently, two competing accounts exist on the origin of Bildts. One theory says that Bildts is a Dutch dialect acquired by Frisian speakers (Kloeke 1927). The existence of Bildts as a native dialect may be seen as an example of the expansion of Dutch as a language of official and every-day communication in Friesland. Another theory claims that Bildts is the result of the import of Dutch, more specifically Hollandic speakers (De Vries 1927). Within the latter theory, Bildts is mainly the result of uninterrupted transmission of native speakers from Southern Holland.
The reason for this debate is the seemingly contradictory nature of Bildts. Within all of these theories, Bildts is considered a direct descendant of Hollandic as it was spoken in the early 16th century. The classification of Bildts as a dialect of Dutch has most likely been made on the basis of its core lexicon and its basic pronunciation, which is mostly Hollandic, suggesting continuous native transmission. However, salient features such as verbal flection and some nominal morphology are Frisian (see Sections 2.4 and 2.5), so based on its morphology and syntax, Bildts would be classified as a Frisian dialect. Bildts has incorporated a relatively complex infinitive system from Frisian, which seems an unlikely candidate for borrowing by non-native Frisian speakers (Hoekstra et al. 2001a). Frisian native input therefore has to be assumed.
The paradox is only apparent if we consider the possibility that both scenarios are to some extent true. This is in line with the demographic analysis, which allows us to distinguish two groups at the period of the initial settlement: a majority of South Hollandic speakers, and a minority of Frisian speakers (Section 3) who acquired South Hollandic as a second language. The South Hollandic language was gradually frisianised as a result of widespread bilingualism and as a result of slow but steady Frisian immigration. If this is correct, the earliest Bildts must have been only slightly different from South Hollandic. The earliest source in Bildts comes from Dirck Jansz, a farmer who kept a diary over the course of several years in the early 1600s. The language used in his diary looks significantly more Hollandic than present-day Bildts. 
Thus those who took sides in the debate on the origin of Bildts failed to grasp the implication of the fact that Bildts is truly a mixed language, displaying a grammatical matrix which may be considered Frisian, while its core vocabulary is mostly South Hollandic. The pattern of facts exhibited in Bildts is characteristically found in mixed languages, as will be clarified in the following section.
4.2 Characteristics of mixed languages
Mixed languages are all characterised by some sort of split that may not be explained by conventional models of either interference or borrowing. Thomason (1997b: 80) presents the following positive characteristics of mixed languages:
various subsystems of the resulting language can be separated according to the language of origin;
there is little to no simplification of either component in the resulting mixed language;
they evolve or are created in two-language contact situations;
bilingualism is widespread in at least one and usually in both speaker groups;
the origin and preservation of the language involves the desire to maintain a separate identity
The first two properties (see Section 2) make mixed languages qualitatively different from other instances of language contact. The split in linguistic origin between various subsystems of a mixed language is too clear-cut to be explained on the basis of more orthodox patterns of borrowing and interference. The prototypical mixed language has been described as one where one language provides finite verb inflection, rules for word order in the VP, and typically also structures of clause combining, whereas the other language is the lexifier of most of the potentially unbound core lexicon (Matras 2003). This basic distinction is claimed to be found in all mixed languages. However, the language providing finite verb inflection is better defined as a language providing contextual flection, at least in the case of Bildts, seeing that inherent verbal properties such as Frisian weak declension classes were not preserved in Bildts (cf. Section 2). We offer this as a tentative remark, since it is not the purpose of this article to review the various definitions of mixed languages.
Properties (3) and (4) have already been discussed in Section 3. Mixed languages typically arise from situations of characterised by bilingualism (Bakker 1994; Thomason 1997a; Mous 2003), so incorporation of words from the lexifier language does not necessarily undergo interference from the matrix language. Because mixed languages typically arise in an environment of bilingualism, they do not serve to bridge a communicative gap, but rather serve as markers of an in-group identity.  This identity may be created through the emergence of a new group, or as a form of maintenance of an old identity under pressure. The emergence of such a language as an in-group marker is commonly the result of conscious effort to maintain or transform a language (see the section “Language and social identity” in Krauss and Chiu ). Evidence for property (5) is discussed in the section below.
4.3 Identity and mixed languages
There is ample evidence that Bildts speakers retained consciousness of their Hollandic roots, which played a large role in the formation of a Bildts identity. An example is the circulation of a 16th-century affidavit from Holland in the 17th and 18th century. This document, which only survives in a copy, relates the origin of four settlers in the South Hollandic town of Sassenheim, near Leiden (Kuiken 2013: 169). Another example is the popular belief that Steven Huygen is the common ancestor of all “genuine” people from Het Bildt. This Steven never lived in Het Bildt himself, but some of his descendants are thought to have been among its initial settlers (Kuiken 2013: 278). Furthermore, many Bildts-speakers claim to be direct descendants from the original settlers (Koldijk 2004: 42–44), even though continuous Frisian immigration and Frisian surnames cast severe doubt on such a claim. Thus there is clear evidence of a desire for maintaining a Bildts identity.
Mixed languages arise out of social needs rather than communicative ones (Thomason 2001: 198). This is all the more true for Bildts, which is mutually intelligible with Frisian. Identity is under threat within a bilingual community. Speakers belonging to a small community may have a wish to maintain their identity, and express it by means of an in-group language (Thomason 2001: 198). This has also been found to be the case with other mixed languages. Thomason and Kaufman (1988: 225–226; but cf. Mous 2003) state that it was in such a context that Ma’á received its mixed character. Ma’á, like Bildts, functions as a marker of retained identity: its existence is the outcome of resistance to shift; its mixed character is the outcome of pressure to shift.
Due to the factors mentioned in Section 3, such as economic decline, South Hollandic lost prestige in Het Bildt. This led to increased pressure to switch to Frisian among the remaining population. Despite pressure to shift, inhabitants of Het Bildt maintained a separate identity from their Frisian-speaking neighbours by preserving the most salient elements from their heritage Hollandic dialect. The pressure from Frisian prevented maintenance of more abstract structures in their inherited language. Instead, the lexical elements that were maintained came to be imported into a framework of Frisian contextual flection and syntax due to widespread bilingualism among native inhabitants of the region and Frisian immigrants. The resulting lect eventually fossilised into Bildts.
Bildts exists as the result of the settlement of native Hollandic speakers, and these speakers have managed to transfer a large part of the core lexicon of their dialect to subsequent generations. Widespread Frisian-South Hollandic bilingualism, as well as immigration of speakers of Frisian, put the South Hollandic dialect spoken by the initial settlers of Het Bildt under pressure, and caused a near-complete grammatical convergence with Frisian. Within such a linguistic setting, it is not the change towards a Frisian grammatical framework that is surprising, but the maintenance of a South Hollandic lexicon. This is where the identity issue comes in. Mixed languages typically emerge as a marker of identity. In the case of Bildts, it is shown that its formation may be seen in light of an effort of its speakers to retain their identity. The original Hollandic grammar, in contrast, was eroded by widespread Frisian bilingualism. Thomason (1997a: 483) states that “language is a salient group marker, powerful enough that some new (sub-)ethnic groups have invented their own […], so it is not surprising that a group clinging to their separate identity would also cling as much as possible to their ethnic language” (see also Krauss and Chiu 1998). It is a very peculiar thing for a language to undergo so much change in core morphology and syntax, and at the same time not undergo a similar change in core lexicon. Mixed languages have a peculiar set of social circumstances in common. Bildts offers one more view on the prototypical mixed language, and it stands out as a case in which the two languages being mixed are genetically very close.
The authors would like to thank Kees Kuiken and Cor van der Meer for commenting on an earlier version of this article.
Bakker, P. 1994. Michif. In P. Bakker & M. Mous (eds.), Mixed languages: 15 case studies in language intertwining. Amsterdam: IFOTT.Search in Google Scholar
Barbiers, S., H. Bennis, G. de Vogelaer, M. Devos & M. van der Ham. 2005. SAND. Syntactische Atlas van de Nederlandse Dialecten. Deel I. Amsterdam: Amsterdam University Press.Search in Google Scholar
Barbiers, S., J. van der Auwera, H. Bennis, E. Boef, G. de Vogelaer & M. van der Ham. 2008. SAND. Syntactische Atlas van de Nederlandse Dialecten. Deel II. Amsterdam: Amsterdam University Press.10.5117/9789053567791Search in Google Scholar
Buwalda, H. S. 1948. Út ‘e ouwe doas. Drie Bildtse teksten fan Waling Dykstra. Nije útgave, fan inlaiding en ferklarings foorsien deur H.S. Buwalda. St. Annaparochie: Drukkerij Van Leer & De Jong.Search in Google Scholar
Buwalda, H. S. 1960. Over de verba in het Bildts. In K. Dykstra, K. Heeroma, W. Kok & H. T. J. Miedema (eds.), Fryske Studzjes. Oanbean oan Prof. Dr. J. H. Brouwer, 361–370. Assen: Van Gorcum.Search in Google Scholar
Buwalda, H. S. 1963. Hoe skrive wy ’t in ’t Bildts? ’n Hanlaiding foor de spelling, en inkele dingen út ‘e spraakkûnst. St.-Anna-Parochy: Van Leer & de Jong.Search in Google Scholar
Buwalda, H. S. 1964. Aanvullingen op H. Sannes, Geschiedenis van Het Bildt, deel I. Franeker: Wever.Search in Google Scholar
Daan, J. & D. P. Blok. 1969. Van Randstad tot Landrand. Bijdragen en Mededelingen der Dialectcommissie van de KNAW. Amsterdam: Noord-Hollandsche Uitgevers Maatschappij.Search in Google Scholar
De Schutter, G. T. De Schutter T. de Jong Goeman & B. van den Berg. 2005. MAND – Morfologische atlas van de Nederlandse dialecten. Deel I. Amsterdam: Amsterdam University Press.10.5117/9789053566961Search in Google Scholar
De Vries, W. 1927. Zijn Bilts en Vriezenveens ontstaan doordat Friezen van taal veranderden? Tijdschrift voor Nederlandse Taal- en Letterkunde 46. 198–209.Search in Google Scholar
Gerritsen, M. 1991. Atlas van de Nederlandse Dialectsyntaxis (AND). Amsterdam: P. J. Meertens-Instituut.Search in Google Scholar
Goeman, T., M. Oostendorp, P. van Reenen, O. Koornwinder & B. van den Berg. 2008. MAND – Morfologische atlas van de Nederlandse dialecten. Deel I. Amsterdam: Amsterdam University Press.10.5117/9789053567746Search in Google Scholar
Golovko, E. V. 1994. Copper Island Aleut. In P. Bakker & M. Mous (eds.), Mixed languages: 15 case studies in language intertwining, 113–121. Amsterdam: IFOTT.Search in Google Scholar
Goossens, J. 1970. De indeling van de Nederlandse dialecten. In J. Goossens, K. Heeroma, G. H. Kocks & H. Entjes (eds.), Dialectindeling en structuurgeografie, 56–73. Groningen: Sasland.10.21825/kzm.v30i0.21000Search in Google Scholar
Goossens, J. 1977. Inleiding tot de Nederlandse Dialectologie. Groningen: Wolters-Noordhoff.Search in Google Scholar
Goossens, J., J. Taeldeman en G. Verleyen. 1998. FAND – Fonologische atlas van de Nederlandse dialecten. Deel I: Het korte vocalisme. Gent: Secretariaat van de Koninklijke Academie voor Nederlandse Taal- en Letterkunde.Search in Google Scholar
Goossens, J., J. Taeldeman en G. Verleyen. 2000. FAND – Fonologische atlas van de Nederlandse dialecten. Delen II+III. Deel II: De Westgermaanse korte vocalen in open syllaben; Deel III: De Westgermaanse lange vocalen en diftongen. Gent: Secretariaat van de Koninklijke Academie voor Nederlandse Taal- en Letterkunde.Search in Google Scholar
Haeseryn, W., K. Romijn, G. Geerts, J. de Rooij & M. C. van den Toorn. 1997. Algemene Nederlandse Spraakkunst. Groningen: Martinus Nijhoff.Search in Google Scholar
Hoekstra, E. 1994. Oer de oerienkomsten tusken de dialekten fan Noard-Hollân en it Frysk. In Ph. H. Breuker, S. Dyk, D. Gorter & L. G. Jansma en W. Visser (eds.), Philologia Frisica anno 1993, 81–103. Fryske Akademy: Ljouwert.Search in Google Scholar
Hoekstra, E. 2002. A structure-based analysis of morphosyntactic regularities in language contact. In J. W. Zwart & W. Abraham (eds.), Studies in comparative Germanic syntax, 277–291. Amsterdam: John Benjamins.10.1075/la.53.16hoeSearch in Google Scholar
Hoekstra, E. 2012. Fries substraat in de syntaxis van het West-Fries: oorsprong en implicatie van het onderscheid tussen infinitief en gerundium en hun syntactische distributie. It Beaken. Tydskrift fan de Fryske Akademy 74. 151–172.Search in Google Scholar
Hoekstra, E. & M. Van Koppen. 2001a. Het Bildts als resultaat van Fries-Hollands taalcontact. In D. Boutkan & A. Quak (eds.), Language contact. Substratum, superstratum, adstratum in Germanic languages. Amsterdamer Beiträäge zur Alteren Germanistik 54. Amsterdam: Rodopi, 89–106.Search in Google Scholar
Hoekstra, E. & M. Van Koppen 2001b. Het diminutiefsysteem in het Bildts: Fries of Hollands? In D. Boutkan & A. Quak (eds.), Language contact. Substratum, superstratum, adstratum in Germanic languages. Amsterdamer Beiträge zur Alteren Germanistik 54, 107–116. Amsterdam: Rodopi.10.1163/18756719-054-01-90000010Search in Google Scholar
Hoekstra, J. & P. M. Tiersma. 1994. Frisian. In E. König & J. van der Auwera (eds.), The Germanic languages, 505–531. London: Routledge.Search in Google Scholar
Kloeke, G. G. 1927. De Hollandsche expansie in de zestiende en zeventiende eeuw en haar weerspiegeling in de hedendaagsche Nederlandsche dialecten. Proeve eener historisch-dialectgeographische synthese. ’s-Gravenhage.Search in Google Scholar
Koldijk, J. D., 2004. Het Bildts, zijn wezen, herkomst en problematiek: Een dialectgeografisch en historisch onderzoek. Leeuwarden: Fryske Akademy.Search in Google Scholar
Krauss, R. M. & C.-Y. Chiu. 1998. Language and social behavior. In S. T. Fiske, G. Lindzey (eds.), The handbook of social psychology, 41–88, Vol. 2, 4th edn. New York: McGraw-Hill.Search in Google Scholar
Krol, G. & R. C. Goeman. 2016. Official application for the recognition of the Bildts language under part II of the European Charter for Regional and Minority Languages. Sint-Annaparochie, 19 February 2016.Search in Google Scholar
Kuiken, C. J. 2013. Het Bildt is geen eiland: Capita cultuurgeschiedenis van een vroegmoderne polder in Friesland. Groningen & Wageningen: Nederlands Agronomisch Historisch Instituut.10.21827/5e70c0eb417f7Search in Google Scholar
Lafeber, A. P. M. met medewerking van L. B. Korstanje. 1967. Het dialect van Gouda. Oudheidkundige Kring “Die Goude”, Gouda.Search in Google Scholar
Matras, Y. 2003. Mixed languages: Re-examining the structural prototype. In Bakker, P. & Y. Matras (eds.), The mixed language debate, 151–175. Berlin: Mouton de Gruyter.10.1515/9783110197242Search in Google Scholar
Provinsje Fryslân. 2014. Streektalenatlas 2014: Streektalen in beeld. Leeuwarden: Provinsje Fryslân.Search in Google Scholar
Schönfeld, M. 1970. Historische grammatica van het Nederlands. Edited by A. van Loey. Zutphen: Thieme. 2003: http://www.dbnl.org/tekst/scho074hist01_01/Search in Google Scholar
Swadesh, M. 1971. The origin and diversification of language. J. SherzerChicago: Aldine.Search in Google Scholar
Thomason, S. G. 1997b. A typology of contact languages. In A. Spears & D. Winford (eds.), The structure and status of pidgins and creoles, 71–88. Amsterdam: John Benjamins.10.1075/cll.19.06thoSearch in Google Scholar
Thomason, S. G., 2001. Language contact. Edinburgh: Edinburgh University Press.Search in Google Scholar
Van Bree, C. 1996. Historische Taalkunde. Amersfoort: Acco. 2004: http://www.dbnl.org/tekst/bree001hist01_01/Search in Google Scholar
Van Bree, C. & A. Versloot. 2008. Oorsprongen van het Stadsfries. Met medewerking van Rolf H. Bremmer jr. Leeuwarden: Fryske Akademy.Search in Google Scholar
Van Coetsem, F. 1960. De Friese relativa en conjuncties met t. In K. Dykstra & K. Heeroma et al. (eds.), Fryske studzjes, 327–334. Oanbean oan Prof. Dr. J. H. Brouwer op syn sechstichste jierdei 23 Augustus 1960, Assen: Van Gorcum.Search in Google Scholar
Van der Meer, G. 1985. Frisian ‘Breaking’. Aspects of the origin and development of a sound change. Groningen: Stifting FFYRUGSearch in Google Scholar
Van der Veen, K. 1986. Yndieling en relative ôfstân fan Fryske plattelansdialekten. Philologia Frisica 1984. 10–60.Search in Google Scholar
Van Sluis, P. S. 2015. Seven perspectives on Bildts: The Bildts language status in a European context. Leeuwarden: Fryske Akademy.Search in Google Scholar
Woude, G. van der. 1960. Oer it gebrûk fan ‘t by bynwurden. In K. Dykstra, K. Heeroma, W. Kok & H. T. J. Miedema (eds.), Fryske Studzjes, 335–343. Oanbean oan Prof. Dr. J. H. Brouwer. Assen: Van Gorcum.Search in Google Scholar
© 2016 van Sluis et al, published by De Gruyter
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.