Case morphology in Dutch
At the start of its existence as a language in its own right, Dutch had a morphological case-marking system comprising nominative, genitive, dative and accusative cases (although this system was in decline and syncretic to some degree, as noted by van der Horst 2008:145), and a three-way gender distinction (masculine, feminine and neuter). Case and gender agreement within the noun phrase was marked on the determiner, any adjectives, and on some nouns (particularly masculine and neuter singulars in the genitive). This is exemplified in (3).
(3) de alley bloeyenste Eeuwe des
the all blossoming.SUPER century the.GEN.NEUT.SG
Roman.GEN.NEUT.SG empire.GEN. neut.sg
‘the most blossoming century of the Roman Empire’ (EMDC, 16th-17th century, Academic)
Deflection caused the loss of morphological case marking from spoken Dutch during the Middle Dutch period, and the weakening and eventual loss of the masculine-feminine nominal gender distinction in northern Dutch, which is attested from the 16th century onwards (van der Horst 2008: 803), by the early modern period. Both morphological case marking and the masculine-feminine distinction remained in the written language until the early 20th century as part of the norm of the standardised written code. The manifestations of deflection most relevant to the adnominal use of the genitive case were the competition from the van-construction (4), whose suppression of the adnominal genitive gained strength throughout the Middle Dutch period (Weerman and de Wit 1999:
- 1158), and the exaptation of the possessive -s construction (5) from the originally masculine and neuter singular genitive suffix -s which, in its new role, could also be attached to feminine nouns and was only marked once in the noun phrase, meaning that a determiner in the noun phrase would remain uninflected (see, e.g., Booij2010: 216-222). The construction exemplified in (4) is now the default means of connecting two noun phrases in a possessive or partitive relationship, rather than two adjacent noun phrases (one of which is morphologically marked as genitive), which is possible in a language with morphological case marking (see, e.g., Weerman 1997: 437).
- (4) eene vande nieuwe predicanten one of-the new ministers
‘one of the new ministers’ (EMDC, 16th-17th century, Diaries)
(5) voor-by de Bruydts huys quam ryden
past the.0 bride.poss house came ride
‘came riding past the bride’s house’ (EMDC, 16th-17th century,
From the 16th century onwards, starting with the first true grammar of Dutch, Twe-spraack (1584), grammarians customarily included a morphological case system - sometimes with six cases after the Latin model (e.g. Twe-spraack 1584; van Heule [1625/1626] 1953), or with four cases (e.g. Bilderdijk 1826; Weiland 1805) - and the three-way gender distinction in their norm; indeed, knowledge of the latter was a necessity for accurate use of the former (van der Wal and van Bree 2008: 241-244, 294-296). Over time, a number of the grammarians started to take usage into account; for instance, by noting - and, often, prescribing - the register-based division of labour between synthetic case marking in formal language and analytic alternatives in less formal language (e.g. ten Kate 1723: 334-340; Weiland 1805: 76). By the 19th century there was a discrepancy between spoken language and the old fashioned written language (Willemyns 2003:110). Morphological case marking was long gone from speech (e.g. Muller 1891: 201-202); as a written phenomenon, it had become an orthographical matter (van der Horst and van der Horst 1999: 311) and was finally lost from the written language as a result of spelling reforms in the first half of the 20th century.
-  Van der Horst’s exact description is of “vier min of meer te onderscheiden naamvallen” [“fourmore or less distinguishable cases”] (2008:145).
-  Notice that, since the gender of huis ‘house’ is neuter, this example does not involve a compound whose constituents are written apart; for this to be the case, the example would have toread het Bruydts huis ‘the bride-house, i.e. ?bridal house’. The determiner de in (3) clearly refersto Bruydt and not to huis.