• Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all
Cohering vs. non-cohering affixes

Like Dutch, Afrikaans has cohering and non-cohering affixes (the terms were introduced in Dixon 1977). The distinction between cohering and non-cohering affixes pertains to the interface between morphology and phonology: cohering affixes form a prosodic word with their stem, while non-cohering affixes are prosodic words of their own. Thus, the morphological boundary of a cohering affix does not necessarily coincide with a prosodic boundary. All non-native suffixes and most native suffixes are cohering. A number of native suffixes, however, are non-cohering. These are the following suffixes, which were all historically independent words: -agtig, -baar, -dom, -heid, -ling, -loos, -skap, -saam . For the same historical reason, affixoids are always non-cohering.

The distinction is easy to see by comparing the syllabification of the words geel·agtig like yellow and gel·erig yellowish. Since -agtig is a non-cohering suffix, the syllable boundary of geel·agtig falls after the final consonant of the stem geel: /xelˈɑx.tǝx/. By contrast, -erig is cohering, and so the stem and suffix are syllabified as a unit and the final consonant of the stem geel becomes the onset of the second syllable: /ˈxe.lǝ.rǝx/.

Afrikaans prefixes are all non-cohering, that is, they do not form one prosodic word with the stem. This can be seen in the way prefixed words are syllabified: the syllable boundary between a prefix and stem is the same as the morpheme boundary. An illustrative example is the derived verb ver·as /fərˈɑs/ [[ver](VBZ)[as](N)](V) VBZ·ash incinerate, which forms a minimal pair with the stem verras /fəˈrɑs/ to surprise (the two [r] sounds are reduced to one by degemination).


The system of Afrikaans cohering and non-cohering affixes overlaps with that of Dutch to a large extent. The analysis of the Afrikaans system below is therefore based on the same Dutch topic.


While Afrikaans prefixes are non-cohering, the morphological boundary may be blurred in fast speech, because syllabification then applies across the prefix boundary. Compare for instance ver·en·ig·ing society, which might be pronounced in fast speech as /fəˈre.nə.xəŋ/.

The boundary between a prefix and stem also blocks the process of pre-vocalic schwa-deletion that applies without exceptions within a prosodic word. For instance, the schwa of be- is not deleted in be·їn·druk /bǝˈǝn.drœk/ to impress somebody.

Note that prefixes, although non-cohering in nature, only form a prosodic word of their own if they contain at least one full vowel. Hence, a prefix like be- /bə/ is not a prosodic word.

If a cohering suffix is attached to a compound, it will form a prosodic word with the last prosodic word of the compound. For instance, the prosodic and morphological structure of the word god·s·dienst·ig religious (derived from the compound god·s·diens religion) are as follows:


The prosodic difference between these two types of suffix can also be observed in gapping in complex words.


The distinction between these types of suffixes in Dutch is discussed in detail in Booij (1995).


  • Booij, Geert1995The phonology of DutchOxfordOxford University Press
  • Dixon, R.M.W1977A grammar of YidinCambridge UP
printreport errorcite