• Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all

Compounding is the concatenative word formation process where two or more roots (specifically combination forms), stems, words, and phrases are combined to form a new word. For instance, the word eet+kamer dining room is a compound formed from the nouns eet to eat and tafel table, while morf·o+logie is a compound formed from the roots morf- and -logie, together with the interfix-o-.

In analyses, a plus sign (+) is used to indicate boundaries between stems/roots in compounds; the middle dot (·) is used before interfixes (and before/after other affixes); the division sign is used for affixoids (÷); an underscore ( _ ) is used for univerbations.

Adapting Van Huyssteen and Verhoeven's taxonomy of compounding for Afrikaans and Dutch (2014), which was based on the work of Scalise and Bisetto (2009), the taxonomy in the figure below presents an overview of compounding in Afrikaans. This taxonomy also serves to structure the topics discussed on Taalportaal.

Van Huyssteen and Verhoeven (2014:36) excluded reduplication as a form of compounding, while it is included here under the rubric reduplicative compound. What they called a "compounding compound" is called a compositional compound here.
Figure 1: Taxonomy of compounding in Afrikaans
[click image to enlarge]

In addition to the classification above, Afrikaans compounds can be classified according to the word class of their heads. The most productive types are nominal compounds and adjectival compounds, while verbal compounds are in principle unproductive. These kinds of compounds are discussed in detail under subordinative compounds.

Terminology and compound classification

Scalise and Bisetto's (2009) seminal chapter on the classification of compounds opens with an important observation: "Classification of compounds has been and still is a much-debated issue in the linguistic literature." This is due to issues related to terminology (which is often used and/or applied differently in and for different languages); differences in what is considered to be a compound or not; inconsistent application of (inconsistent) criteria; diverging view-points from different theoretical frameworks (e.g. on the boundaries between morphology and syntax); differences in methodological approaches (e.g. lack of using usage-based data); different purposes (e.g. for computational purposes); etc. One of the most telling examples is the divergent definitions of the term synthetic compound (Von Schroeder 1874; Bloomfield 1933), which is also used synonymously (or near synonymously) with verbal-nexus compound(Marchand 1969), parasynthetic compound, verbal compound(Selkirk 1982), deverbal compound, secondary compound, and syntactic compound. To add to the confusion, terminology and definitions in different languages might also vary, e.g. samestellende samestelling vs. samestellende afleiding in Afrikaans and Dutch (Booij and Van Santen 2017:185-194), or Zusammenbildung in German (see Neef (2015) for a discussion) for the same or related phenomena.

Even though Scalise and Bisetto's (2009) universal taxonomy of compounding is widely accepted, it is still not uniformly adopted in morphological circles. Van Huyssteen and Verhoeven (2014:36) provided an overview of some of the shortcomings and criticism of the Scalise and Bisetto taxonomy, and provided a slightly adapted version of the taxonomy for Afrikaans and Dutch, specifically with the purpose of annotating data for the CompoNet database. Since 2014, new data and theoretical insights necessitated an updated version of Van Huyssteen and Verhoeven's (2014:36) taxonomy, which was published in Van Huyssteen (2017).

The taxonomy used here on Taalportaal is an adaptation of the latter. Where the purpose of the aforementioned taxonomies of Van Huyssteen and Verhoeven (2014), and Van Huyssteen (2017) was for the annotation of data for comparative-linguistic purposes, the purpose here is more of a descriptive-linguistic nature. Also, some of the choices – e.g. to leave outThe most updated version of the taxonomy will always be available here.

[+]Prototypical properties of compounds

In accordance with the generally accepted properties of (Germanic) compounds (e.g. Kempen (1969:83-135); Lieber and Štekauer (2009); Olsen (2000); Ziering and Van der Plas (2020)), the following properties of prototypical Afrikaans compounds can be mentioned:

  • Realisation (form)
    • In spoken language, Afrikaans compounds generally have a uniform tempo and stress contour, where the left-hand constituent receives primary stress, and the right-hand constituent secondary stress, following Booij's (1995)compound stress rule (CSR) – see example (1). If the left-hand constituent is a multisyllabic word, it follows the general stress patterns for simplex words (Wissing 2019) – see example (2). In contrast, the head of a phrase normally receives primary stress, or otherwise numerous words in a phrase could carry equal stress – see the examples in (3).
      swimming pool
      laboratory assistant
      a. Hulle bly in 'n baie klein huisie.
      They live in a very small house.
      kləin ˈɦœi.si
      a.' Hulle bly nie in 'n groot huis nie, maar in 'n baie klein huisie.
      They don't live in a big house, but in a very small house.
      ˈkləin ˈɦœi.si
      a.'' Hy sit in die kleinhuisie.
      He sits in the toilet.
      Wissing (2019) demonstrated that in compound place names "the final segment appropriates stress", which he ascribed to the fact that "the original constituent segments have lost (or have begun to lose) their original meanings and thus also their status as separate constituents, as part of the more general process of the compound’s gaining, or having gained, monomorphemic status".
      Contrary to the observations about Dutch by Visch (1989), all categories of Afrikaans compounds (including adjectival and prepositional compounds) have default stress on their first constituent (bar individual exceptions, of course). Compare with the topic on the compound stress rule for Dutch.
    • In written language, compounds are by default styled as one word, or in some cases with a hyphen (e.g. to avoid legibility problems, or to indicate certain meaning relationships). Orthographical rules have been developed for over more than a century, and which are well-documented in AWS-11. Crucially, AWS-11, 15.1 states that separate words (i.e. in word groups, phrases, and sentences) are written disjunctively, while AWS-11, 15.3 dictates that constituents (i.e. words) in compounds are written conjunctively.
      film+ekstravaganza (> filmekstravaganza)
      film extravaganza
      a. dans·-+ekstravaganza (> dans-ekstravaganza)
      dance extravaganza
      [If written without a hyphen as dansekstravaganza, the possibility arises to read the last component of the word erroneously as ?sekstravaganza sextravaganza.]
      b. onderwyser-boer
      person who is both a teacher and a farmer
    • The constituents of a compound cannot be separated by other constituents. With reference to example (6b) it is impossible to add another adjective (e.g. *wild·e+swart+bees wild·ATTR+black+bovine), or another noun (e.g. *wild·e+afrikaner+bees wild·ATTR+Africander+bovine) between the two constituents of the compound. In contrast, it is possible to insert other words in phrases, e.g. wild·e, swart bees wild, black bovine, or wild·e afrikaner+bees uncontrollable head of Africander cattle.
      Excluded from this is swearword infixation, like kind·er-fokken-kak child·LK-fucking-shit balder-fucking-dash (WKJ).
    • The constituents of a compound are not interchangeable, unlike words in some phrases. Hypothetically, wild·e, swart bees and swart, wild·e bees would be possible phrases with equal meanings. In contrast, the first two constituents in the compound swart+wild·e+bees white-tailed gnu cannot be interchanged, bar a handful of exceptions.
  • Conceptualisation (meaning)
    • The right-hand constituent in a compound mostly functions as both grammatical and semantic head, and is modified by a left-hand constituent (Olsen 2000:908). Such compounds are known as endocentric compounds (a.k.a. determinative compounds, or tatpurusa compounds). Of course, many exceptions abound:
      • A few compounds (most often loan words, or loan translations) are left-headed, e.g. ouditeur-generaal auditor-general, who is a kind of and not a kind of , and hence takes plural marking on the left-hand constituent (i.e. ouditeur·s-generaal).
      • Exocentric compounds (a.k.a possessive compounds, or bahuvrihi compounds) lack a semantic head, e.g. rooi+kop red+head person with red hair.
      • Coordinative compounds (a.k.a. dvandva compounds and/versus kharmadharaya compounds) have a right-hand grammatical head, but two semantic heads, e.g. onderwyser-boer teacher-farmer person who is both a teacher and a farmer, which takes plural marking on the right-hand constituent only (i.e. onderwyser-boer·e). In some definitions (e.g. Bauer 2004) a distinction is made between proper name compounds like Hewlett-Packard as dvandva compounds, and other cases like onderwyser-boer that are considered to be kharmadharaya compounds (or in Bauer's terms also appositional compounds).
      • Appositive compounds (a.k.a. appositional compounds) are semantically and grammatically right-headed compounds, but where there are most often an metaphorical interpretation between the left-hand and right-hand constituents (in contrast with attributive compounds that have a literal relation). In appositive compounds the left-hand constituent expresses an attribute (i.e. descriptive quality) of the right-hand constituent by means of a noun or verb (in contrast with attributive compounds where adjectives and adverbs are normally used to express an attribute). (Also see Scalise and Bisetto (2009:51).)
    • The left-hand constituent serves as a generic/type, or indefinite specification that restricts or modifies the set of possible denotations of the right-hand constituent. Hence, left-hand constituents are generally bare words or phrases, i.e. words or phrases unmarked for grammatical categories like number, degrees of comparison, case, tense, etc.; see examples in (5). Noticeable exceptions regularly occur in especially phrasal compounds (example (6a)), and attributive compounds (example (6b)).
      a. boek+winkel
      book store
      [Despite the fact that a book store sells a variety of books, the left-hand constituent in the compound is not marked for number, e.g. *boek·e+winkel book·PL+store.]
      b. twee+persoon+span
      two-person team
      [The bare phrase twee persoon two person is used as left-hand constituent, instead of the grammatically marked phrase twee person·e two person·PL two people.]
      a. medies·e+fonds+bydrae
      medical aid contribution
      b. wild·e+bees
      [The phrase wild·e bees refers to any , whereas the compound (with the attributively marked adjective wild·e) refers to a .]

      If the left-hand constituent is a noun, it is by default (and in principle, for the reasons mentioned above) the unmarked form of the noun (i.e. without plural suffix). The default analysis for a word like perd·e+stal horse stable is therefore [[[perd](N)[e](LK)](allo)[stal](N)](N) (i.e. to analyse the -e- as an interfix, and not as a plural suffix).

      Kempen (1969:96) explicitly stated that a plural analysis is only valid if the compound could be contrasted with another compound that is in form and meaning in the singular. Compare the contrastive examples below where it is illustrated that universiteit·e+span could be contrasted with universiteit+span (and a plural interpretation of the e is therefore possible), but since *perd+stal is not an existing word in Afrikaans, the e in perd·e+stal should be analysed as an interfix.

      a. universiteit·e+span
      team consisting of players from several universities
      a.' universiteit+span
      team consisting of players from a university
      b. perd·e+stal
      stable for animals of the type horse
      b.' *perd+stal
    • The meaning of a compound is not overtly expressed in the combination of the constituents, and are therefore oftentimes semantically ambiguous (or potentially ambiguous). The intended meaning is therefore construed through conventionalised schemas, and encyclopaedic knowledge. Compare the N+N compounds in (7), all with tafel table as right-hand constituent, but with different meaning relations with their left-hand constituents.
      a. kombuis+tafel
      kitchen table = table in kitchen
      b. hout+tafel
      wooden table = table made of wood
      c. operasie+tafel
      operation table = table used during an operation; table used to operate on
    • Contrary to what is possible in phrases, the left-hand constituent in a compound cannot be modified independently from the right-hand constituent. In a phrase like klein kind small child the adjective can be modified by an adverb like baie very to denote a very small child. In the compound klein+kind grandchild such modification is not possible: both *baie klein+kind and *baie+klein+kind would be ungrammatical, while in baie klein+kind·er·s many small+child·LK·PL many grandchildren the modifier baie is a indefinite numeral, and not an adverb modifying klein.
  • Usage
    • Compounds are typically used as a name of a coherent concept, rather than a description of it. For example, the compound klein+kind child of your son or daughter denotes a category of kinship, while the phrase klein kind small child describes a child in terms of size. Also compare English grandchildchild of your son or daughter, versus grand childvery pleasing child. As Olsen (2000:899) observed: "… compounds often express a more consolidated meaning than an equivalent syntactic phrase."
    • Endocentric compounds form the largest and most productive class of compounds. Of these, nominal compounds, and specifically N+N compounds, constitute the largest subclass (i.e., circa 80% of all Afrikaans compounds are N+N compounds).

Compounds as words vs. phrases – morphology vs. syntax

The definition of what a compound is, is at the heart of some of the controversies around the definitions of the concepts word and phrase, and subsequently around the boundaries between morphology and syntax. Literature about the topic abound (see Ziering and Van der Plas (2020) for a recent overview), with opposing theoretical viewpoints being defended quite strongly.

In a seminal article, Haspelmath (2011) examined ten criteria for defining a cross-linguistically valid concept of "(morphosyntactic) word", viz.: potential pauses; free occurrence; mobility; uninterruptibility; non-selectivity; non-coordinatability; anaphoric islandhood; nonextractability; morphophonological idiosyncrasies; and deviations from biuniqueness. He showed that "none of them is necessary and sufficient on its own, and no combination of them gives a definition of "word" that accords with linguists' orthographic practice." He concluded that "we do not currently have a good basis for dividing the domain of morphosyntax into "morphology" and "syntax", and that linguists should be very careful with general claims that make crucial reference to a cross-linguistic "word" notion." This viewpoint is fully compatible with a constructionist definition of grammar, where there is no difference between morphology and syntax, bar the "size" of the constructions being described.

[+]Prototypical examples: Subordinative compounds

Table 1
Morphosyntactic type Morphosemantic type POS of constituent 1 POS of constituent 2 POS of compound Example Gloss Translation
Verbal-nexus Endocentric N V·ADJZ ADJ hand+vervaardig·de hand+produce·ADJZ hand-made
N V·NMLZ N gras+sny·er grass+cut·NMLZ lawn mower
V V·NMLZ N eet+staak·ing > eetstaking eat+strike·NMLZ hunger strike
ADJ V·NMLZ N kaal+nael·er naked+run·NMLZ streaker
ADV V·PTCP ADJ dig+bebos·te thick+afforest·ADJZ thickly wooded
Ground Endocentric N N N tafel+poot table+leg table leg
N ADJ ADJ kleur+blind colour+blind colour blind
N ADV (POSTP) ADJ/ADV pad+langs road+along via road; directly
N V N dag+breek day+break dusk
N V V raad+pleeg advice+commit to consult
NUM N N twee+klank two+sound diphthong
V N N stryk+plank iron+board ironing board
V ADJ ADJ werk+sku work+shy lazy
ADJ N N siek+bed sick+bed sickbed
INTERJ N N sjoe+broek ouch+pants hot pants
Exocentric V N V knip+oog snip+eye to wink
V N N suip+lap booze+cloth drunkard
V N ADV druip+stert drip+tail embarrassed
V ADV N trap+suutjies tread+softly chameleon
N N N tronk+voël jail+bird jailbird
N N ADJ hoender+kop chicken+head drunk
ADV V N mal+trap crazy+step crazy, jolly person
Phrasal Endocentric VP N N skop-skiet-en-donder-film kick-shoot-and-hit-movie action movie
NP N N kaas-en-wyn-onthaal cheese-and-wine-party cheese and wine party
sentence N N God-is-dood-teologie God-is-dead-theology "God is dead" theology
(Neo)classical Endocentric root root N hidro+logie hydro+logy hydrology
root N N bio+brandstof bio+fuel biofuel
N root N Japan+logie > Japannologie Japan+logy Japanese studies

[+]Prototypical examples: Attributive compounds

Table 2
Morphosyntactic type Morphosemantic type POS of constituent 1 POS of constituent 2 POS of compound Example Gloss Translation
Ground Endocentric ADJ N N blou+draad blue+wire galvanised wire
ADJ (PREP.INTR) N N buite+kamer outside+room outside room
ADV N N terug+weg back+way the way back
ADV ADJ ADJ donker+blond dark+blonde dark blonde
N ADJ (PREP.INTR) ADJ bek+af beak+off disappointed
Exocentric ADJ N N rooi+kop red+head person with red hair
ADJ N ADJ kaal+voet bare+foot barefooted
ADJ N ADV vroeg+dag early+day early in the morning
NUM N N tien+kamp ten+camp decathlon
Phrasal Endocentric AP N N los-en-vas-praatjies loose-and-set-talks random chatting
NP N N kop-aan-kop-botsing head-on-head-collision head-on collision
PP N N in-die-kol-humor in-the-spot-humour spot-on humour

[+]Prototypical examples: Appositive compounds

Table 3
Morphosyntactic type Morphosemantic type POS of constituent 1 POS of constituent 2 POS of compound Example Gloss Translation
Ground Endocentric N N N treffer+liedjie hit+song hit song
N ADJ ADJ hond+mak dog+tame as tame as a dog
N ADJ ADJ hond+lelik dog+ugly very ugly
V ADJ ADJ spring+lewendig jump+lively alive and well
Phrasal Endocentric VP ADJ ADJ kielie-my-maag-lekker tickle-my-stomach-nice as nice as tickling my stomach
NP ADJ ADJ sonsak-in-Ibiza-mooi sunset-in-Ibiza-pretty sunset-in-Ibiza-pretty

[+]Prototypical examples: Coordinative compounds

Table 4
Morphosyntactic type Morphosemantic type POS of constituent 1 POS of constituent 2 POS of compound Example Gloss Translation
Ground Endocentric N N N skrywer-boer writer-farmer writer-farmer
ADJ ADJ ADJ stom+verbaas mute+surprised very surprised
V V V sit-lê sit-lie sit and lie
PREP PREP PREP van+af from+of from
PREP PREP ADV (PREP.INTR) voor+op before+above firstly
Exocentric PR PR PR Bosnië-Herzegowina Bosnia-Herzegovina Bosnia-Herzegovina
N N N ma-kind mother-child mother-child

[+]Prototypical examples: Compositional compounds

Table 5
Morphosyntactic type Morphosemantic type POS of constituent 1 POS of constituent 2 POS of compound Example Gloss Translation
Phrasal Endocentric NP N N groot+skerm+televisie big+screen+television big screen television
NP N N twee+sitplek+motor two+seat+car two-seater
PP N N buite+boord+motor out+board+motor outboard motor
AP N N ou+jong+kêrel old+young+dude bachelor
VP N N lekker+sit+stoel nicely+sit+chair comfortable chair
VP N N haar+sny+skêr hair+cut+scissors hairdressing scissors

[+]Prototypical examples: Parasynthetic compounds

Table 6
Morphosyntactic type Morphosemantic type POS of constituent 1 POS of constituent 2 POS of compound Example Gloss Translation
Phrasal Endocentric NP NMLZ N groot+skaal·s large+scale·ADJZ large-scale
NP ADJZ ADJ een+blaar·ig > eenblarig one+leaf·ADJZ monopetalous
VP NMLZ N alleen+loop·er > alleenloper alone+walk·NMLZ loner
VP NMLZ N ter+aarde+bestel·ing > teraardebestelling to+earth+deliver·NMLZ burial

[+]Prototypical examples: Separable complex verbs

Table 7
Morphosyntactic type Morphosemantic type POS of constituent 1 POS of constituent 2 POS of compound Example Gloss Translation
Ground Endocentric PREP V V in+gooi in+throw to throw in
ADJ V V goed+keur good+grade to approve
ADV V V neer+gooi down+throw to throw down
N V V vleis+braai meat+roast to barbeque
Exocentric ADV (PREP.INTR) ADJ V aan+dik on+thick to exaggerate
ADV (PREP.INTR) N V in+perk in+bound to confine; to limit

[+]Prototypical examples: Reduplicative compounds

Table 8
Morphosyntactic type Morphosemantic type POS of constituent 1 POS of constituent 2 POS of compound Example Gloss Translation
Ground Endocentric V V V vat-vat touch-touch to touch tentatively
ADJ ADJ ADJ lang-lang long-long very long
ADV ADV ADV nou-nou now-now just now
N·PL N.PL N·PL bakk·e-bakk·e bowl·PL-bowl·PL many/countless bowls
NUM NUM ADV twee-twee two-two in pairs of two
Exocentric INTERJ INTERJ N hoep-hoep hoop-hoop Upupa africana (name of a common South African garden bird)
INTERJ INTERJ ADV doef-doef boom-boom pit-a-pat
INTERJ INTERJ V tjirp-tjirp chirp-chirp to chirp iteratively
N N N huis-huis house-house playing house
V V ADV lag-lag laugh-laugh easily
N N ADV plek-plek place-place sporadically

[+]A note on compounding vs. univerbation

Compound words of non-lexical categories, such as conjunctions, have not been formed through compounding, but through univerbation. This is the case for conjunctions like omdat because and deurdat because, and complex prepositions such as agterop at/on the back, behind. For a discussion of cases like these, see the topic on univerbation.

  • Booij, Geert1995The phonology of DutchOxfordOxford University Press
  • Booij, Geert and Ariane van Santen2017Morfologie. De woordstructuur van het Nederlands.AmsterdamAmsterdam University Press
  • Marchand, Hans1969The categories and types of present-day English word formationMünchenBeck
  • Visch, Ellis1989The rhythm rule in English and DutchUtrecht UniversityThesis
printreport errorcite