• Dutch
  • Frisian
  • Afrikaans
Show full table of contents
Intrusive stop formation

Derivational and inflectional processes may give rise to heterorganic sequences of nasal + plosive (which hardly occur in underlying representations). Such sequences undergo a repair, resulting in homorganicity. Within words, the repair mechanism is so-called Intrusive Stop Formation (ISF). Inflected verb forms like do komst /kom+st/ you come and do sjongst /sjoŋ+st/ you sing, with the heterorganic cluster /-ms-/ and /-ŋs-/, are realized as [kompst] and [sjoŋkst], with a [p/k] inserted between [m/ŋ] and [s]. Intrusive Stop Formation is the topic of this section.


It is a virtually exceptionless pattern that a nasal and a plosive have the same place of articulation when making up or being part of a final sequence; the bilabial nasal /m/ is the only one to enter into heterorganic clusters, few though the cases may be (see word-final sequences of a nasal and an obstruent). A second source of heterorganic clusters are derivational and inflectional processes, which seem to be blind to the homorganicity demand.

In general, heterorganic clusters do not reach the surface unchanged. Instead, they undergo a repair, resulting in homorganicity. Across words, the repair mechanism is Regressive Place Assimilation (Nasal Assimilation), which sees to it that the coronal (word-final) nasal /n/ adopts the place specification of the following (word-initial) non-continuant segment, resulting in an organic, inseparable consonant sequence.

[hide extra information]

In case the word with final /n/ is followed by a word beginning with a continuant segment, the sequence vowel /-n/-continuant segment provides a context for Vowel Nasalization, as a result of which /n/ disappears, passing on its nasality to the preceding vowel.

Within words the repair is achieved through so-called Intrusive Stop Formation (ISF), which is exemplified in (1):

Example 1

Examples of Intrusive Stop Formation
a. Following /m/
a.1 In underived words
himd /hɪmd/ [hɪmpt] vest, singlet
ramt /ramt/ [rampt] casing, frame(work)
jamk /jamk/ [jampk] often; very; possibly
nimf /nɪmf/ [nɪmpf] nymph
tjems /tjɛmz/ [tjɛmps] sieve (for milk)
boems /bums/ [bumps] bang, boom, bounce
a.2 In derived and/or inflected forms
beamke /bjɛm+kə/ [bjɛmpkə] small tree; bush, shrub (cf. beam /bɪəm/ [bɪəm] tree )
glimkje /ɡlɪm-k+jə/ [ɡlɪmpkjə] smile (cf. glimme /ɡlɪm+ə/ [ɡlɪmə] to shine, to gleam )
ymker /im-k+ər/ [impkər] bee-keeper (cf. ime /imə/ [imə] honeybee )
komst /kom+st/ [kompst] coming, arrival (cf. komme /kom+ə/ [komə] to come, to arrive )
om't /om+t/ [ompt] because (cf. omdat /om+dɔt/ [omdɔt] because )
do komst /kom+st/ [kompst] you come (cf. komme /kom+ə/ [komə] to come, to arrive )
hy komt /kom+t/ [kompt] he comes (cf. komme /kom+ə/ [komə] to come, to arrive )
wat stoms /stom+s/ [stomps] something stupid (cf. stom /stom/ [stom] stupid )
b. Following /ŋ/ (only in derived and/or inflected forms)
salang't /salaŋ+t/ [salaŋkt] as long as (cf. salang as /salaŋ+ɔs/ [salaŋ ɔs] as long as )
eangst /ɪəŋ+st/ [ɪəŋkst] fear (cf. eang /ɪəŋ/ [ɪəŋ] afraid )
fangst /faŋ+st/ [faŋkst] catch(ing) (cf. fange /faŋ+ə/ [faŋe] to catch )
opbringst /opbrɪŋ+st/ [obrɪŋkst] yield, revenue (cf. opbringe /opbrɪŋ+ə/ [obrɪŋə] to yield )
hy sjongt /sjoŋ+t/ [sjoŋkt] he sings (cf. sjonge /sjoŋ+ə/ [sjoŋe] to sing )
neat langs /laŋ+s/ [laŋks] nothing long (cf. lang /laŋ/ [laŋ] long )
[hide extra information]

The context of ISF is largely confined to non-coronal nasals, due to the fact that a) virtually all derivational and inflectional suffixes are coronal, hence homorganic with /n/, and b) that [+continuant] obstruents following /n/ trigger Vowel Nasalization, as a result of which /n/ does not reach the surface.

[hide extra information]

ISF is also triggered by syllabic sonorant consonsonants, whether or not the latter have undergone Progressive Place Assimilation. Examples are provided below:

Example 2

sikens /si:k+əns/ [si(:)kŋks] illness
lekkens /lɛkən+s/ [lɛkŋks] cloths, sheets
stommens /stom+əns/ [stomm̩ps] dumbness
tsjeppens /tsjɛp+əns/ [tsjɛpm̩ps] prettiness

This is the only instance in which a heterorganic cluster of a nasal and a plosive is not due to a derivational or inflectional process, but to a phonological one. It shows once more that a syllabic sonorant consonant is a 'normal' consonant in all respects (except for its syllabic affiliation).

In cases like goedens /ɡuəd+əns/ [ɡuədn̩ts] goodness and wiettens /vjɪt+əns/ [vjɪtn̩ts] wetness ISF occurs following the coronal nasal /n/, in fact syllabic [n] (compare the previous Extra).

The effect of ISF may be that formations which differ in their underlying representation become indistinguishable at the surface (cf. Booij (1995:137) as to Dutch). Some examples of this are given in (3):

Example 3

Examples of ISF erasing underlying differences
do dampst /damp+st/ [dampst] you steam
do damst /dam+st/ [dampst] you play draughts
hy swinkt /swɪŋk+t/ [swɪŋkt] he swerves; turns
hy swingt /swɪŋ+t/ [swɪŋkt] he swings
wat stomps /stomp+s/ [stomps] something blunt
wat stoms /stom+s/ [stomps] something stupid
[hide extra information]

The word betingst /bətɪŋ+st/ condition is linked to the verb beting(je) /bətɪŋ/ insist on, to require. The great majority of Frisian-speakers are unaware of this and since, moreover, they do not regularly write Frisian, the wordt betingst is often misspelled as betinkst, in line with its pronunciation, viz. [bətɪŋkst]. The derivation at hand therefore may have been reinterpreted as /bətɪŋk+st/, which is now connected to the verb betink(e) think of, to invent, to devise, so that betinkst is assigned the meaning what has been invented, devised. Be this as it may, it is indicative of the phonological reality of the intrusive stop /k/ in betingst.

It is not without reason that the stop in question is called 'intrusive' and, by the same token, that the name of the process makes mention of 'stop formation' and not of 'stop insertion'. There is no segment insertion in a literal sense here. What strikes the ear as an inserted stop, is a by-product of phonological retiming. Take wat stoms /stom+s/ something stupid. Due to retiming of the phonetic content of the phoneme /m/ at the melodic level with respect to its underlying position at the structural phonological level, the place feature [labial] is linked to the structural position of the following obstruent /s/. The latter, thus, ends up associated with two place specifications: in its first phase it is a labial sound and in its final phase a coronal one. Retiming only involves the place features, leaving the specification for continuancy, sonorancy and voicing unaffected. The upshot of this is that /s/ is interpreted as a voiceless (bi)labial plosive − nasal consonants are [-continuant] − in its first and as a voiceless coronal fricative in its final phase: [stom(ps)]. The same applies to the place feature [velar] of /ŋ/. In hy sjongt /sjoŋ+t/ he sings, for instance, [velar] spreads to /t/, turning the latter into a segment with a sequence of two different place specifications: [sjoŋ(kt)]. Due to ISF, heterorganic final clusters of a nasal and a plosive attain homorganicity at the surface level, in line with the general pattern.

[hide extra information]

There are two repair mechanisms for heterorganic nasal plus obstruent clusters, viz. Regressive Place Assimilation (Nasal Assimilation) and Intrusive Stop Formation. The former is a phonological process, the latters one of phonetic implementation, hence they display huge differences. For one thing, Nasal Assimilation only targets the coronal nasal /n/, whereas ISF operates with respect to all nasals.

[hide extra information]

If the above analysis is correct, then ISF does not affect the number of structural phonological positions of a form. This means that one may expect a durational difference between, for instance, (wat) stomps /stomps/ something blunt (six structural positions) and (wat) stoms /stoms/ something stupid (five such positions). Though the relation between (abstract) phonological length and (concrete) phonetic duration is not as straightforward as suggested, there nevertheless seems to be an interesting subject for phonetic research here.

[hide extra information]

In the sequence /-mf/ − as in nimf /nɪmf/ nymph − the difference between bilabial /m/ and labiodental /f/ seems to be big enough for them to count as heterorganic with respect to ISF.

[hide extra information]

Being a matter of phonetic implementation, ISF is likely to operate at the surface level. This means that in a word like himd /hɪmd/ vest, singlet, it affects the surface form [hɪmt], with final [t], which explains the voiceless [-p-] of [hɪmpt]. In contexts where Regressive Voice Assimilation applies − like ramt#belied framework policy and de nimf betsjoent my the nymph enchants me − the left-hand obstruent ends up as voiced: [ramdbəliət] and [nɪmv bətsjuənt]; this explains the voiced intrusive stop here: [rambdbəliət] and [nɪmbv bətsjuənt].

[hide extra information]

The lateral approximant /l/ also seems to trigger ISF. The word els /ɛls/ alder; (brad)awl may be pronounced as [ɛlts], which renders it indistinguishable from elts [ɛlts] each (one). This seems to be indicative of the fact that /l/ and /s/ form a heterorganic sequence. As to this, it should be noted that Popkema (2006:63) characterizes /l/ as a dental/prepalatal and not as a coronal segment.

[hide extra information]

Clements (1987:47, footnote 4) suggests that in order for ISF to apply the nasal and the following obstruent must be tautosyllabic. This means that there should be a difference between monosyllabic beamt [(bjɛmt)] and disyllabic beamte [(bjɛm)(tə)], both meaning trees. If such a difference exists at all, it is hard to perceive.

  • Booij, Geert1995The phonology of DutchOxfordOxford University Press
  • Clements, George N1987Phonological Feature Representation and the Description of Intrusive StopsCLS23:229-50
  • Popkema, Jan2006Grammatica FriesUtrecht/ LjouwertUitgeverij Het Spectrum BV Prisma Woordenboeken en Taaluitgaven/ Fryske Akademy