• Dutch
  • Frisian
  • Afrikaans
Show full table of contents
Case - the partitive construction

The bound morpheme -s, a remnant of the old case system, can be used as a marker of a partitive genitive, as in iets groen-s something green and een heleboel lekker-s a lot of sweets where -s is attached to an adjective.


In Middle Dutch, nouns and adjectives after a quantifying expression were marked as (partitive) genitives. This system has mostly disappeared (Scott 2014), but there are idioms where the partitive genitive has been preserved, as in niet veel soups not much soup-s without quality:

Example 1

a. niet veel soep-s
not much soup-s
without quality
b. niet veel soep
not much soup
not much soup

This development in Dutch is parallel to that in English that also lost its partitive genitive construction, and replaced it with a PP construction, as in a spoon of honey, or simply omitted the -s after quantifiers, as in something sweet. However, unlike English, standard Dutch still has a productive construction with a partitive marker -s(Broekhuis 1996, Hoeksema 1998, Booij 2002: 52-54). Noun phrases can have the surface form Quantifier Adjective+s, as is illustrated by the following:

Example 2

a. iets groen-s
something green
b. niets waar-s
nothing true
c. wat leuk-s
something nice
d. allerlei heerlijk-s
all sorts of lovely (things)
e. een heleboel lief-s
a lot of sweet (things)
f. een massa goed-s
a lot of good (things)
Example 3

a. veel mooi-s
a lot of nice (things)
b. weinig prachtig-s
few splendid (things)
c. meer fraai-s
more fine (things)

The difference between Dutch and English is illustrated by the first example iets groens and its English gloss something green: in (standard) Dutch the presence of the -s is obligatory, whereas in English there is no morphological marking on the adjective. However, southern Dutch as spoken in Belgium is similar to English in this respect, and thus exhibits phrases such as iets groen.

[hide extra information]

If the adjective ends in /s/, the suffixed form is indistinguishable from the base form, as illustrated in iets paars something purple (< paars /pars/ purple). The same holds for adjectives in /z/, thanks to final devoicing: iets dwaas something silly (< dwaas /dwaz/ silly).

Besides quantifying expressions, the wat voor what kind of construction also requires this kind of morphological marking of the adjective:

Example 4

Wat voor mooi-s heb je gezien?
What for beautiful-s have you seen?
What kind of beautiful things did you see?

The partitive construction is productive: it applies to adjectives that can be used as prenominal predicating adjectives, as illustrated in the examples above.

As observed in Broekhuis (1996) and Hoeksema (1998), however, there are some classes of adjectives that can be used pre-nominally but yet cannot be used in this construction: relational adjectives (including geographic adjectives in -er), temporal adjectives, material adjectives ending in -en, and adjectives ending in schwa:

Example 5

a. *iets Groninger-s
something from Groningen
b. *iets zaterdag-s
something Saturday-
c. *iets ijzer-en-s
something iron
d. *iets oranje-s
something orange

The adjective in this construction can be separated from the quantifier by modifiers, as in iets absoluut geweldig-s something absolutely fantastic and iets volgens mij ongelofelijk-s something according to me unbelievable-SUF something I cannot believe.

[hide extra information]

The appearance of the -s on the adjectives in (1) looks like a case of contextual inflection in which the quantifier requires the adjective to carry the genitive ending -s. However, this interpretation would raise the question why in present-day Dutch nouns are not subject to contextual inflection when used with a partitive meaning, as shown in (1b).

What is the categorial status of the adjective+s words? Van Marle (1996) has argued that they are nouns, and that the presence of the suffix -s hence has a category-changing effect. Arguments for that position are that Dutch has deadjectival nouns in s that can be used as the head of an NP, without a quantifying determiner being present:

Example 6

a. wat soep
what soup
some soup
b. allerlei rommel
variety rubbish
all sorts of rubbish
c. een massa snoep
a mass confectionery
a lot of sweets

However, the nouns iets something and niets / niks nothing do not occur with nouns in a partitive construction, only with the -s-marked adjectives:

Example 7

a. *iets soep
some soup
b. *niets soep
nothing soup
c. *niks soep
nothing soup

The other option to be considered is that the -s-marked words are adjectives. This is the position taken by Schultink (1962: 79-80) who states that the only function of -s is to mark this specific construction. Evidence for this interpretation is that the -s-marked words can head an adjectival phrase with adverbs and verbal/adjectival complements that occur before the head; such modifiers do not occur before nouns (Paardekooper 1958: 161-70, Broekhuis 1996, Hoeksema 1998).

Example 8

a. iets heel lastig-s
something very difficult-s
something very difficult
b. iets volgens mij ongelofelijk-s
something according.to me unbelievable-s
something that is unbelievable according to me
c. niets in het oog vallend-s
nothing in the eye falling-s
nothing that catches the eye
d. iets op ons doel afgestemd-s
something on our goal tuned-s
something tuned to our goal

The word heel very in (8a) is an adverb, which suggests that the next word is an adjective. In (8b) the prepositional phrase volgens mij according to me functions as an adverbial modifier. In (8c) the word vallend falling is a present participle used as an adjective, as is the past participle afgestemd tuned in (8d). Hence, vallends and afgestemds allow for verbal complements. The occurrence of such specifiers and complements that are typical for adjectives and participles is a problem for assigning the -s-marked adjective the status of noun, unless one can show that deadjectival nouns systematically inherit the syntactic valency of their adjectival bases. This is not the case, however. For instance, the deadjectival noun nieuw-s news cannot be modified by adverbs, only by adjectives, as illustrated by the contrast between het goed-e nieuws the good news and *het goed nieuws (goed is the adverbial form) versus iets heel nieuws something that is very new, with the adverb heel very.

Another problem for Van Marle’s hypothesis that the suffix -s used here is a suffix that creates nouns, is that many of these words in -s cannot occur in other nominal syntactic slots, except for some lexicalized nouns in -s such as nieuws news: noun phrases such as *het lastigs the difficulty or *het in het oog vallends that what catches the eye are impossible. Yet, the construction as a whole is an NP, hence there must be a noun that can function as the head of an NP. Van Marle was aware of this problem, and therefore proposed a rule that makes this category shift from adjective to noun dependent on the specific syntactic context discussed here.

Kester (1996) and Hoeksema (1998) have proposed to analyze this construction in a different way, as a determiner phrase with a quantifier in the determiner position, and an empty head noun. Thus, the phrase iets lastigs would receive the following structural analysis (with a DP analysis of noun phrases):

Example 9


where e stands for ‘empty element’. The -s is then considered a case of contextual inflection of adjectives, and will be interpreted as licensing an empty noun in the head position of the NP. This analysis has the advantage of explaining why these words in -s do not have the normal distribution of nouns. Those cases where the adjective+s combination clearly behaves as a noun, as is the case for nieuws news can then be seen as lexicalized cases of a nominal reinterpretation of the adjective on the basis of the surface structure in which an overt noun is lacking.

The drawback of this analysis is that the appearance of the -s is not a normal case of contextual inflection, since it is restricted to NPs with empty nouns. The normal contextual inflection of adjectives in prenominal position is by means of the suffix -e [ә] or Ø.

Therefore, the analysis proposed by Broekhuis (1996) in which the quantifier noun is the head of the NP and is followed by an AP is to be preferred. Hence, the -s is a bound morpheme attached at the right edge of an AP in post-nominal position. This AP is a modifier of the head noun that denotes an indefinite quantity. This analysis captures the insight that in modern Dutch APs cannot appear as post-nominal modifiers, with the exception of these APs when marked by -s. Thus, there is a parallel with the Determiner construction where the -s licenses the use of NPs as pre-nominal modifiers.

In line with the arguments given above, the Partitive Construction is characterized as follows in Booij (2010: 227):

Example 10

Partitive -s Construction
[NP(i) [… [x-s](A)](AP)(j)]NP(k) ↔ [Quantity(i) with Property(j)](k)

The construction in (10) has an open slot x for adjectival stems. All qualifying adjectives can be inserted into this slot, except those of the few subclasses mentioned below (4). This analysis correctly predicts that nouns like iets and niets can co-occur in this construction even though they cannot occur before nouns.The variable x stands for the phonological form of the stem of the adjective. We thus express that the -s is a bound morpheme that is phonologically part of the phrase-final adjective.

The interpretation of the partitive -s as an affix of an adjective is confirmed by the observation that in this construction the -s can only appear on the adjective:

Example 11

a. We zoeken iets [op ons doel afgestemd-s] AP
We search something to our goal tuned-s
We look for something tuned to our goal
b. *We zoeken iets [afgestemd op ons doel-s]AP
We search something tuned to our goal-s
c. *We zoeken iets [afgestemd-s op ons doel]AP

In (11b), the -s is affixed to the final noun, which leads to ungrammaticality. In (11c) the adjective is not at the right edge of the AP, which also leads to ungrammaticality.

The Partitive Construction is similar to the Possessive Construction in that in both cases a bound morpheme -s that is a relic of the genitive case marker functions as the morphological marker of a specific syntactic construction.

  • Booij, Geert2002The morphology of DutchOxfordOxford University Press
  • Booij, Geert2010Construction morphologyOxford/New YorkOxford University Press
  • Broekhuis, Hans & Strang, Anke1996De partitieve genitiefconstructieNederlandse taalkunde1221-238
  • Broekhuis, Hans & Strang, Anke1996De partitieve genitiefconstructieNederlandse taalkunde1221-238
  • Broekhuis, Hans & Strang, Anke1996De partitieve genitiefconstructieNederlandse taalkunde1221-238
  • Broekhuis, Hans & Strang, Anke1996De partitieve genitiefconstructieNederlandse taalkunde1221-238
  • Hoeksema, Jack1998Adjectivale inflectie op -s: geen geval van transpositieHoekstra, Erik, Smits, Caroline & Marle, Jaap van (eds.)Morfologiedagen 1996, Cahiers van het Meertens InstituutAmsterdam46-72
  • Hoeksema, Jack1998Adjectivale inflectie op -s: geen geval van transpositieHoekstra, Erik, Smits, Caroline & Marle, Jaap van (eds.)Morfologiedagen 1996, Cahiers van het Meertens InstituutAmsterdam46-72
  • Hoeksema, Jack1998Adjectivale inflectie op -s: geen geval van transpositieHoekstra, Erik, Smits, Caroline & Marle, Jaap van (eds.)Morfologiedagen 1996, Cahiers van het Meertens InstituutAmsterdam46-72
  • Hoeksema, Jack1998Adjectivale inflectie op -s: geen geval van transpositieHoekstra, Erik, Smits, Caroline & Marle, Jaap van (eds.)Morfologiedagen 1996, Cahiers van het Meertens InstituutAmsterdam46-72
  • Kester, Ellen-Petra1996The nature of adjectival inflectionUtrechtUniversity of UtrechtThesis
  • Marle, Jaap van1996The unity of morphology; on the interwovenness of the derivational and inflectional dimension of the wordBooij, Geert & Marle, Jaap van (eds.)Yearbook of morphology 1995Dordrecht/Boston/LondonKluwer Academic Publishers67-82
  • Paardekooper, Piet. C1958ABN-Spraakkunst. VoorstudiesDen BoschL.C.G. Malmberg
  • Schultink, Henk1962De morfologische valentie van het ongelede adjectief in modern NederlandsDen HaagVan Goor Zonen
  • Scott, Alan2014The Genitive Case in Dutch and German. A Study of Morphosyntactic Change in Codified LanguagesLeidenBrill