• Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all
Final devoicing

The fact that voiced obstruents are absent in syllable-final position in phonetic forms in Afrikaans and some other Germanic languages has led to the formulation of a phonological rule known as Auslautsverhärtung, also named Final Devoicing. Whereas the plural form of the word hoed is [hu.də], with a voiced plosive [d], the singular is [hut] – with a final voiceless [t], frequently described as the result of a process of devoicing. Whereas this is a nearly obligatory phenomenon in Afrikaans in the case of /b/ and /d/, the situation is a bit more complex for /v/.


Final Devoicing in Afrikaans is not seen as a phonological process in all cases. De Villiers and Ponelis (1987) are of the contention that it has already been morphologised. Wissing (1989) shows that it is still a creative rule in Afrikaans by referring to a number of less known devoicing processes of Afrikaans, such as Daaf for the first name Dawie. Furthermore, foreign names with voiced /b/ and /d/ as codas are regularly pronounced voicelessly, e.g. Habib as [hɑbip]. With regard to /f/ and /v/, Grieshaber (1981) and (1987) proposed that an underlyingly /f/ must be postulated which becomes voiced in the appropriate environments (i.e. the opposite of devoicing). Lubbe and Zonneveld (1996), on the other hand, maintain the traditional standpoint that an underlying /v/ is devoiced in final position, but suggest an alternative approach within the theoretical framework of Lexical Phonology; see below for more details. Wissing (2017) holds the same view in his description of this phenomenon in an Optimality Theory framework.

Final devoicing in the English of persons who speak less than perfect Afrikaans may serve as external evidence for the existence of this rule in Afrikaans. Some examples are the English words bad, cab, have which are often subject to final devoicing in Afrikaans-English.

The following examples show final devoicing of underlying /d/ (1) and /b/ (2), and to a lesser extent of /v/ (3) in Afrikaans.

/ɦɑnd/ > [ɦɑnt] hand vs. [hɑn.də] hands
/rəb/ > [rəp] rib vs. [rə.bə] ribs
/vɔlv/ > [vɔlf] wolve vs. [vɔl.və] wolves

Devoicing of /d/ as a singleton coda, as well as in coda clusters, occurs quite commonly in Afrikaans. The examples below are a subset of a larger group, restricted to examples of words with preceding long /a/ and /o/ (1.a), and preceding short /ɑ/ and /ɔ/ (1.b). In all of these cases, resyllabification takes place in plural forms, with, as a consequence, the obstruents in question occurring in syllable onset position, where final devoicing is no longer applicable:

(1.a) daad; draad; graad; raad; saad; aand; maand; naald; baard; brood; dood; nood; potlood; oond; oord; koord; moord; woord

(1.b) band; brand; eiland; hand; land; pand; rand; sand; stand; strand; toestand; verband; vyand; wand; bond; hond; mond; pond; wond

Only a few words with /b/ in the syllable coda exist in Afrikaans. Example (2.a) contains nouns of this type, arguably with underlying /b/. Examples (2.b ) – (2.d) comprise some special cases in support of the existence of final devoicing.

(2.a) eb; kibab; klub; rob; skub; sub; tob; web.

(2.b) A small number of proper names: Jakob, Job and Namib, some of them nonlocal: Gottlieb; Habib; Rob; Saab.

(2.c) Abbreviations of foreign words, nativised in Afrikaans, e.g: bib (biblioteek), fab (fabulous) and lab (laboratorium).

(2.d) Loan words from English, e.g: bob; job; slab.

Unlike the cases of /d/ and /b/, /v/ is more complicated. Though, in earlier works, /v/ is included in the set of voiced obstruents that form the underlying coda, and that undergo devoicing in surface forms (e.g. by Wissing (1971) as well as by Combrink and De Stadler (1987)). More recently a number of alternative frameworks have been proposed. De Villiers & Ponelis (1987) reject the existence of Final Devoicing overall, suggesting instead that all relevant cases have already been morphologised. Grieshaber (1987), on the other hand, proposes a /f/-voicing rule, thus advocating underlying forms with /f/ as coda. Lubbe and Zonneveld (1996) take a separate position, in that they reject both explanations. While they do maintain Final Devoicing as a rule, they differentiate between two types, in the process recognising two sources of [f]: firstly, a morphological rule on Level 1 of the lexicon devoices the underlying /v/ only before +foreign affixes (such as in argi[f]-aris, where -aris is a foreign affix), and a general devoicing rule on Level 2 of the lexicon which applies at the end of syllables (e.g. in [slaf] from /slav/ slave; plural [sla.və]).

In (3.a) below a number of plural forms ending on -e are given, with the preceding consonant being the voiced /v/; thus, like the example of slawe mentioned in the previous paragraph, all of these words have a singular form ending on voiceless [f]. The unmarked theoretical position (i.e. contra Grieshaber) would be to thus postulate an underlying /v/ as in /slav/, from which, via Final Devoicing, [f] in a singular form slaf is derived. (3.a) briewe; diewe; druiwe; duiwe; erwe; grawe; geriewe; griewe; groewe; klowe; lywe; motiewe; olywe; proewe; salwe; siwwe; skerwe; skuiwe; stowe; tariewe; verwe.

A large body of adjectives of the kind listed in (3.a.i) exists. Here too the unmarked position would be /v/ > [f], for example for underlying /ɑktiv/ (Sing. aktief [ɑktif], Plur. aktiewe [ɑktivə] active). In these cases /v/ would then be the basic underlying phoneme, as in the plural form; [f], the surface consonant, would be derived via the phonological process/rule of Final Devoicing.

(3.a.i) aktiewe; alternatiewe; defektiewe; definitiewe; depressiewe; effektiewe; eksklusiewe; ekstensiewe; foutiewe; inisatief; inklusief; intensief; kollektief; negatief; objektief; offensief; passief; positief; relatief; selektief; sensitief; tentatief.

Voiced [v] is found intervocalically in a variety of other derivations too, that is in the contexs V_suffix, where the suffixes start with another vowel, viz. in: -eling, -end, -enier, -er, -erig, -erny, -igheid, -ing, -isme, -isties. Examples are provived below (written w and v both denote [v]).

(3.a.ii) howeling; lieweling (also liefling); nuweling; aangewend; asemrowend; belowend; beskrywend; blywend; drywend; lawend; lewend; oorblywend; oorlewend; oorverdowend; stawend; sterwend; strewend; tydrowend; verslawend; voorskrywend; howenier; beywer; drywer; growwer; hawer; huiwer; jenewer; kewer; klawer; lewer; liewer; nuwer; oewer; power; silwer; skrywer; strawwer; stywer; suiwer; uitgewer; verower; wedywer; ywer; klewerig, stowwerig; slawerny; stowwerigheid; aandrywing; aanskrywing; belewing; beproewinge; beskawing; bewing; drywing; herlewing; inskrywing; meelewing; nastrewing; oorlewing; opgrawing; oplewing; samelewing; skuiwings; stawing; uitgrawing; verdowing; verdrywing; verslawing; werwing; aktivisme; negativisme; positivisme; aktivisties; negativisties; positivisties.

According to Coetzee (2014:705), both [v] and [f] can appear vowel-internally in monomorphemes, as in oewer [uvər] river bank versus offer [ɔfər] sacrifice. In the case of morphologically derived environments— across morpheme boundaries— Coetzee (2014) proposes that there is a process of intersonorant voicing that applies to /f/, rendering [v]; the examples in (3.a.i) and (3.a.ii) are relevant in this regard. Only before the diminutive suffix [-i] does [f] occur exclusively, as in grafie [xrafi] small spade.

Some other cases that could possibly be viewed as evidence in support of the notion of a process of intersonorant voicing (as opposed to a process of final devoicing), as propagated by Coetzee (2014), are:

(3b) Alternating verbs: belowe ~ beloof; lewe ~ leef; sterwe ~ sterf; strewe ~ streef; both members of a pair are allowed, e.g. in Hy beloof om te kom; Hy belowe om te kom He promises to come.

(3c) Adjectives in predicative -f or attributive positions -we – e.g. aktiewe ~ aktief. The following examples behave in a similar fashion: kreatief; lief; motief; negatief; objektief; passief; positief; relatief; sensitief; tentatief (e.g. Dit is 'n lekker aktiewe groep resp. Die groep is lekker aktief It is a very active group).

All in all, while final devoicing is clear in the case of the alternation between /b/ and /p/ as well as /d/ and /t/, the same cannot be said for /v/ and /f/.

[+]Final Devoicing and Regressive Voicing Assimilation (RVA)

In compound words with a voiced obstruent as coda of the first component (commonly the obstruent consonant /d/), followed by a component having a voiced onset (mainly but not exclusively another obstruent – see RVA), final devoicing does not always take place; it may be overridden by the application of RVA, as demonstrated in the following cases:

The phoneme /db/ remains [db]: landbou; Wêreldbank; grondbesit; grondwet.

The phoneme /dv/ remains [dv]: Noordwes; landwyd.

    printreport errorcite