Taalportaal - the digital language portal

Dutch
Frisian
Saterfrisian
Afrikaans

Show all

/{s/z}/-insertion between /{t/d}/ and /jə/

quickinfo

Between /{t/d}/ and /jə/ the consonants /{s/z}/ are inserted, so the sequences /{t/d}jə/ surface as [{ts/dz}jə]. The most common context in which /{t/d}jə/ arises is in the paradigm of verbs of the second weak class. Examples are hytsje ( /hit+jə/) emit heat; to heat (next to hyt hot), wachtsje ( /vaxt+jə/) to wait, wiidzje ( /vi:d+jə/) to widen (next to wiid wide), and wyldzje ( /vild+jə/) run about (wildly) (next to wyld wild).

There appears to be a general ban on the sequences /tjə/ and /djə/. If these arise as a result of morphological processes or loanword adaptation, they are in need of repair, which is executed by the insertion of an obstruent. Obstruent sequences in general must meet two constraints: preferably, the obstruents do not agree in continuancy, whereas they must agree in voicing (see complex onset: sequences of two obstruents). This means, first, that only fricatives qualify as repairing obstruents here and, second, that their voicing value must equal that of the preceding plosive. The fricatives act as connective sounds, which is why they are homorganic with the plosive.

extra

The phenomenon at hand does not occur in the dialects of Skiermûntseach/Schiermonnikoog and Hylpen/Hindelopen.

When inflected with the suffix -je (infinitive I, infinitive II, present participle, imperative, the 1st person singular and all plural persons present tense) verbs of the second weak class with a stem ending in /-{d/t}/ end up with final /-{dz/ts}je/. The paradigms of the verbs foarmje to shape, to form, beäntwurdzje to answer, to respond, and achtsje to esteem, to respect are given in the table below:

Table 1: The paradigms of the verbs foarmje shape, form, beäntwurdzje answer, respond, and achtsje esteem, respect

Verb stem	foarm	beäntwurd	acht
Infinitive I	foarmje	beäntwurdzje	achtsje
Infinitive II	foarmjen	beäntwurdzjen	achtsjen
Present tense
1SG	foarmje	beäntwurdzje	achtsje
2SG	foarmest	beäntwurdest	achtest
3SG	foarmet	beäntwurdet	achtet
PL	foarmje	beäntwurdzje	achtsje
Past tense
1SG	foarme	beäntwurde	achte
2SG	foarmest	beäntwurdest	achtest
3SG	foarme	beäntwurde	achte
PL	foarmen	beäntwurden	achten
Imperative	foarmje	beäntwurdzje	achtsje
Past participle	foarme	beäntwurde	achte
Present participle	foarmjend	beäntwurdzjend	achtsjend

extra

Verbs like sintsje [sɪntsjə] to sunbathe, teantsje [tjɛntsjə] to tiptoe; stand on tiptoe, and túntsje [tyntsjə] to garden are related to the nouns sinne sun, tean toe, and tún garden, respectively. Their stems have been augmented with the element /t/. Since the insertion of /{s/z}/ is found in a variety of morphological and phonological contexts − see below − it seems to be conditioned by purely segmental factors.

The finite forms with /-{s/z}-/ must be accounted for. Some (older) Frisian grammars consider the forms with /-{s/z}-/ as basic. This raises two questions: 1) to which part of the finite verb do /{s/z}/ belong and 2) since they do not show up in all forms of the paradigm (see above), how do we get rid of them?

As to the first question, there are two options: either the stem of the verbs in question ends in /-{dz/ts}/, in which case only one suffix is needed, viz. -je, or the suffix is assumed to have three variants, viz. -je, -zje, and -sje, in which case the stems of these verbs end in /-{d/t}/. If the verb stems are assumed to end in /-{dz/ts}/, they have the distributional property that they only occur in combination with the suffix -je. If, on the other hand, three suffix variants are assumed, two of these, viz. -zje and -sje, have a limited distribution in that they only occur in combination with stems ending in /-{d/t}/. Though their intentions are not easy to grasp, these approaches are put forward by Postma and De Clercq (1904:73), Sipma (1913:70), Sipma (1949:27) and Fokkema (1967:58).

As to the second question, Sipma (1913:70,§245) states that

Verbs in -sje, -zje drop s and z of the ending in the 2nd and 3rd person singular of the present, in the imperfect and in the past participle

. This must be a mistake, for the 2nd and 3rd person singular present tense of achtsje and beäntwurdzje, for instance, are (do) achtest/beäntwurdest and not *(do) achtjest/beäntwurdjest (see the table above). This means that /j/ must be dropped as well, which is explicitly put forward in Fokkema (1967:58): the past tense and the past participle of the verbs of the second weak class are formed with -e; the ones ending in /sjə/ and /zjə/ lose /sj/ en /zj/, as do the 2nd and 3rd person singular of the present tense. The analysis in Hoekema (1996:51) is along the lines of Fokkema's.

This, however, cannot be the whole story. It is the stem ending in /-{d/t}/ which shows up in derived forms. A case in point are action nouns, with the suffix -ing, like beäntwurding ( /bəɔntvød+ɪŋ/) answering, response and achting ( /axt+ɪŋ/) regard, respect, esteem, which are not *beäntwurdzing and *achtsing. Another one is compound, like antwurd#nûmer freepost and wacht#keamer waiting room, which are not *antwurdznûmer and wachtskeamer. When we assume that verb stems end in - /{dz/ts}/, another deletion operation, one affecting /{z/s}/ this time, would be called for. Therefore, it seems best to take as a point of departure that the stems of the verbs in question end in - /{d/t}/.

This then leaves the question of the suffix. Is there one suffix, -je, or is -je the basic suffix, which has the allomorphs -zje and -sje? Since suffix allomorphy in Frisian seems to figure within the realm of derivation only − the diminutive suffix is a case in point (see -DIM (diminutive)) − assuming suffix allomorphy here is not a likely option. The only one left then is that inflected forms like antwurdzje and achtsje result from the insertion of /{z/s}/ between the verb stem and the suffix -je. This analysis was already put forward in Tuinstra (1937:20, point 4) and Boersma and Van der Woude (1972:47). It was defended more fully in Tiersma (1979:129-131), Tiersma (1985:33-34) and Tiersma (1999:30), and it is the generally accepted analysis since (see, for instance, Popkema (2006:78)).

Tiersma (1979:129-130) considers the insertion of /{z/s}/ between /-{d/t}/ and /jə/ as the affrication of /{d/t}/: The process here referred to as 'Affrication' involves the conversion of stem-final d and t to the affricates dz and ts before a following suffix -je. He is aware of the fact that assuming affrication as a genuine synchronic process of Frisian phonology is not unproblematical, for he also writes:

There are two possible ways to write a rule to account for this [process]. One is to have the segment t become the affricate ts (analogously for d); the other is to insert an s [or a z] between a stem-final dental stop and a following je. The effect of both is essentially the same, and the main difference seems to be whether one wishes to consider ts and dz as affricates which are, as it were, allophones of t and d before jə; or if one wishes to avoid postulating affricates for Frisian by simply inserting s or z. I will write Affrication as a rule which inserts s or z, depending on the voicing of the preceding consonant

. In complex segments, it is shown that sequences consisting of /s/ + (voiceless) plosive, in both orders, are best considered as (two-rooted) complex segments. Since /ts/ and /dz/ do not function as affricates, but as consonant sequences, an insertion analysis is called for.

extra

The verbs widzje to rock (to and fro) (alongside widze cradle), flodzje to gush, to spout (alongside flodze gush, spurt), belytsje to decrease (alongside lyts small), and ferlytsje to reduce, make smaller (idem) have a stem ending in /-ts/-dz/, so that these clusters show up throughout the whole paradigm, as in mem widze (*widde) it bern yn 'e sliep mother lulled the child to sleep and de foto waard ferlytse (*ferlite) (the size of) the photograph was reduced. They also occur in derivations, like ferlytsing /fərlits+ɪŋ/ reduction; diminutive (*ferliting) and widzer /vɪdz+ər/ who rocks (to and fro) (*widder).

So far, one might gain the impression that /{s/z}/ are only inserted in inflected forms of verbs of the second weak class, that is, preceding the suffix -je. The phrasing above that /{s/z}/ are inserted between /{t/d}/ and the sequence /jə/, however, has been chosen deliberately: as it happens, there are other contexts of insertion.

Firstly, the unstressed ending ie ( /-i/) of Dutch loanwords shows up as je ( /-jə/) in Frisian. If this results in the word-final sequence /-djə/, the latter shows up as /-dzjə/. Examples, taken from Visser (2000:159), are given in (1):

Examples of loanwords ending in /dzjə/

studie

→

stúdje

→

stúdzje

study

komedie

→

komeedje

→

komeedzje

comedy

The same pattern is found in remeedzje remedy, subsydzje subsidy, grant, and trageedzje tragedy.

extra

The loanword keraadzje /kəra:dzjə/ courage − the current form of which is koeraazje − is peculiar in that the word-final sequence /-dzjə/ is the result of the insertion of /d/ and not of /z/. The same holds for nedzje /nɛdzjə/ garnish(ing), which is related to nedding /nɛd+ɪŋ/ seam, attachment (both of which are likely to have become obsolete). They are adapted forms of the French loanword neige. It has been reported that nedzje was also realized with a long vowel, as nêdzje /nɛ:dzjə/; the commonest realization, however, was without /d/, as nêzje /nɛ:zjə/ (see Veen (1984-2011, volume 14), s.v. nedzje II). The latter pronunciation linked up with the one of the original French word. It makes it all the more clear that it is /d/, and not /z/, which has been inserted here, probably due to some sort of analogy.

This also occurs in native words. Originally, the words jierdei birthday and middei afternoon are compounds, the literal meanings of which are year day (jier#dei) and mid day (mid#dei). The literal meaning has become opaque, which has paved the way for a morphological restructuring. Nowadays, jierdei and middei count as simplex words. A clear indication of the latter is that the old plural forms jierdagen birthdays and middagen afternoons have generally been replaced by jierdeis and middeis, with -s, whereas the noun dei day in isolation has kept the irregular plural form dagen days. This restructuring also had phonological repercussions (see Visser (1992)). Being in unstressed position, the original diphthong /ɛj/ of dei has turned into the monophthong /i/: /jɪdi/ (jierdei) and /mɪdi/ (middei). In areas where final /i/ has turned into /jə/, jierdei and middei got the forms /jɪdjə/ and /mɪdjə/. These may be realized as such, but /jɪdzjə/ and /mɪdzjə/, with /-dzjə/, occur as well. This is also an indication that /djə/ is a disfavoured sequence (though not a prohibited one).

Unstressed word-internal /-ti-/ in Dutch loanwords turns into /-tjə-/, which provides a context for the insertion of /s/. Take tuttifrutti /tʏtifrʏti/ tutti-frutti, which turns into Frisian /tøtjəfrøtjə/, eventually surfacing as /tøtsjəfrøtsjə/.

The sequence /-tjə-/ as such can also be part of the underlying representation of Dutch loanwords (see Visser (1992:74, note 10)). Take netjes /nɛtjəs/ neat; decent, which is rendered as netsjes /nɛtsjəs/ in Frisian. When /-tjə-/ arises in Frisian words, it turns into /-tsjə-/ as well. Examples are kultsjefater /køltsjəfa:tər/ cultivator (cf. Dutch cultivator) and the wish wol bekomme 't je much good may it do to you (expressed at the end of a meal), which may show up as wolbekomtsje /volbəkomtsjə/. Both the noun and the wish have become obsolete.

extra

More and more, Dutch /i/ is taken over as such in Frisian. At present then tuttifrutti and kultivator by and large have the same pronunciation in Dutch and Frisian.

extra

The seven numerals ending in -tjin /-tjən/ -teen− fyftjin /fiftjən/ fifteen and santjin /sɔntjən/ seventeen, for instance − show a dialectal alternation with forms ending in -tsjin /-tsjən/: fyftsjin /fiftsjən/ and santsjin /sɔntsjən/. Note that the numerals without /s/ have the disfavoured sequence /-tjə-/.

extra

The Dutch diminutive suffix -tje /-tjə/ has the Frisian counterpart -tsje /-tsjə/. The Frisian equivalents of Dutch kaartje /ka:rt+tjə/ ticket and potje /pɔt+tjə/ little pot, for instance, are kaartsje /ka:t+tsjə/ and potsje /pɔ:t+tsjə/. This is also indicative of the disfavoured status of the sequence /tjə/ in Frisian.

If a Dutch loanword ends in /-sti/ ( stie), the change of unstressed word-final /i/ into /jə/ and insertion of /s/ would result in word-final /-stsjə/The latter sequence is ill-formed and it is simplified to /-sjə/. This means that Dutch kwestie /kʋɛsti/ question and suggestie /sʏɣɛsti/ suggestion are rendered as kwestje [kwɛsjə] and suggestje [søɡɛsjə] in Frisian.

The same is at issue with verbs of the second weak class with a stem ending in /-st/ (see Visser (1993)). In verbs like haastje to hurry (alongside haast /ha:st/ haste; hurry) and ynlistje to frame (alongside list /lɪst/ frame), there is no insertion of /s/ between stem-final /-t/ and /-jə/ (*haastsje and *ynlistsje). Instead, stem-final /t/ deletes: [ha:sjə] (haastje) and [ĩlɪsjə] (ynlistje). Stem-final /t/ is realized in the forms with the other inflectional endings, as in hy haastet /ha:st+ət/ [ha:stət]him net he takes his time (literally: he is not in a hurry) and hja hat de foto kreas ynliste /inlɪst+ə/ [ĩlɪstə] she has neatly framed the photograph. See also derivations like ynlisting /inlɪst+ɪŋ/ framing and oerhaasting /uərha:st+ɪŋ/ rush, hurry, which are always realized with /-st-/.

extra

The verbs hoastje to cough and kostje to cost have lost their stem-final /-t/ altogether − save in the spelling −, so (hja) hoasten (de hiele tiid) (they) coughed (all the time) and (it hat hûndert Euro) koste (it) cost (a hundred Euro's) are realized as [vwasn̩] and [kɔsə], respectively. See /h/-deletion for more on the phonological status and the realization of initial hoa.

Finally, some remarks on the place of /{s/z}/-insertion between /{t/d}/ and /jə/ in the grammar of Frisian are in order.

Word-final -je is realized as either [-jə] or [-i], which show a dialectally complementary distribution (see Sipma (1913:41,§175), Tiersma (1979:161-164), and Visser (1992)). As noted above, the most common context in which the sequences /{t/d}jə/ arise is the paradigm of verbs of the second weak class. One might thus expect these verbs to end in /{-ts/dz}jə/ in the [-jə]-dialect and, without insertion, in /-{t/d}i/ in the [-i]-dialect. This, however, is not borne out by the facts. Verbs like wachtsje to wait and skodzje to shake are realized as either [vaxtsjə]/ [skodzjə] in the [-jə]-dialect or [vaxtsi]/ [skodzi] in the [-i]-dialect (whereas [*vaxti] and [*skodi] are out). As to /{s/z}/-insertion then both dialects behave symmetrically. This led Visser (1992) to postulating a single underlying representation for [-jə] and [-i], viz. a (short) rising diphhtong, the left-hand vowel of which is /i/ and the right-hand one has no content. The latter can get the default specification /ə/, resulting in [jə], whereas [i] is the result if default specification does not take place. So, /{s/z}/-insertion is triggered by 'undivided' -je. This means that it must be an early process, in line with the fact that it is morpho-lexically conditioned.

extra

Diminutive formation (see -DIM (diminution)) also takes up 'undivided' -je. Nouns ending in a velar consonant take -je, which can be realized as [-jə] or as [-i]. So, eachje /ɪəɣ+jə/ little eye and boekje /buk+jə/ booklet are either [ɪəxjə]/ [bukjə] or [ɪəxi]/ [buki]. Between word-final /-ŋ/ and -je the velar plosive /k/ is inserted, no matter whether -je is realized as [-jə] or [-i], so rinkje /rɪŋ+jə/ little ring is pronounced as [rɪŋkjə] or as [rɪŋki]. Diminutives thus show symmetrical behaviour with respect to [-jə] and [-i]. Note that the choice of the diminutive suffix and the insertion of /k/ are also morpho-lexically conditioned.

There are also phonological processes with respect to which [-jə]- and [-i]-final verbs show asymmetrical behaviour. Vowel Nasalization is a case in point (see Tiersma (1979:163-164), Tiersma (1985:34, note), Visser (1992:80)). Verbs like wenje /vɛn+jə/ to live and trúnje /tryn+jə/ to urge are realized as either [vɛ̃jə]/ [trỹjə] or [vɛni]/ [tryni]. This means that vowel nasalization only has access to 'divided'-je, which is in line with the fact that it is a purely (automatic) phonological process.

The same holds for the loanwords in (1) above. Take the Dutch word studie /stydi/ study; it is rendered as stúdzje /stydzjə/ in the [-jə]-dialect, but as /stydi/, and not as /*stydzi/, in the [-i]-dialect. And the native words jierdei birthday and middei afternoon are /jɪd(z)jə/ and /mɪd(z)jə/ in the [-jə]-dialect and /jɪdi/ and /mɪdi/, but not /*jɪdzi/ and /*mɪdzi/, in the [-jə]-dialect.

/{s/z}/-insertion, then, seems to operate in two different ways: as a morpho-lexically conditioned, early process, which treats /jə/ and /i/ symmetrically and as a purely phonologically conditioned process, with respect to which /jə/ and /i/ behave asymmetrically. One and the same process may thus apply under different conditions and with different outcomes (which is in line with one of the premises of the theory of Lexical Phonology).

References

Boersma, Johannes. & Woude, Goasse van der1972Spraeklear I. Lesboekje foar de oplieding ta de Fryske AkteLjouwertAfûk
Fokkema, Klaas1967Beknopte Friese SpraakkunstFryske Akademy, Leeuwarden
Fokkema, Klaas1967Beknopte Friese SpraakkunstFryske Akademy, Leeuwarden
Hoekema, Teake1996Beknopte Friese vormleerLeeuwardenAfûk
Popkema, Jan2006Grammatica FriesUtrecht/ LjouwertUitgeverij Het Spectrum BV Prisma Woordenboeken en Taaluitgaven/ Fryske Akademy
Postma, Gertjan & Clercq, Pieter de1904Lytse Fryske spraekleare, it Westerlauwersk om 1900 hinne oangeandeLjouwertR. van der Velde
Sipma, Pieter1913Phonology and Grammar of Modern West FrisianLondon, New YorkOxford University Press
Sipma, Pieter1913Phonology and Grammar of Modern West FrisianLondon, New YorkOxford University Press
Sipma, Pieter1913Phonology and Grammar of Modern West FrisianLondon, New YorkOxford University Press
Sipma, Pieter1949Ta it Frysk IIIR. van der Velde, Ljouwert
Tiersma, Pieter M1979Aspects of the phonology of Frisian based on the language of GrouMeidielingen fan de stúdzjerjochting Frysk oan de Frije Universiteit yn Amsterdam4
Tiersma, Pieter M1979Aspects of the phonology of Frisian based on the language of GrouMeidielingen fan de stúdzjerjochting Frysk oan de Frije Universiteit yn Amsterdam4
Tiersma, Pieter M1979Aspects of the phonology of Frisian based on the language of GrouMeidielingen fan de stúdzjerjochting Frysk oan de Frije Universiteit yn Amsterdam4
Tiersma, Pieter M1979Aspects of the phonology of Frisian based on the language of GrouMeidielingen fan de stúdzjerjochting Frysk oan de Frije Universiteit yn Amsterdam4
Tiersma, Pieter M1985Frisian reference grammarDordrechtForis Publications
Tiersma, Pieter M1985Frisian reference grammarDordrechtForis Publications
Tiersma, Pieter M1999Frisian Reference GrammarAfûk, Ljouwert
Tuinstra, Uilke S1937Eenvoudige Friese SpraakkunstBatavia
Veen, Klaas F. van der et al1984-2011Wurdboek fan de Fryske taal - Woordenboek der Friese taalFryske Akademy
Visser, Willem1992Oer -je en -JE. De morfology en fonology fan it einichste wurddiel -jeTydskrift foar Fryske Taalkunde769-87
Visser, Willem1992Oer -je en -JE. De morfology en fonology fan it einichste wurddiel -jeTydskrift foar Fryske Taalkunde769-87
Visser, Willem1992Oer -je en -JE. De morfology en fonology fan it einichste wurddiel -jeTydskrift foar Fryske Taalkunde769-87
Visser, Willem1992Oer -je en -JE. De morfology en fonology fan it einichste wurddiel -jeTydskrift foar Fryske Taalkunde769-87
Visser, Willem1992Oer -je en -JE. De morfology en fonology fan it einichste wurddiel -jeTydskrift foar Fryske Taalkunde769-87
Visser, Willem1993In kwestje fan haastjen: oer hoe't yn it Frysk de sekwinsjes -sts en -tj mijd wurdeTydskrift foar Fryske Taalkunde8123-130
Visser, Willem2000Frjemd wurdt eigener. Oer de âlde Frânske lienwurden yn it FryskIt Beaken62141-218

⇑ Back to top

About this article

Author(s): Willem Visser

Category: Frisian Phonology

Keywords: {s/z}-insertion

Version: 1.0, December 2015

Version history:

version	editor(s)	date	remarks
1.0	Willem Visser	December 2015

printreport errorcite

Jump to:

introduction
references
related articles