/ instituut voor de Nederlandse taal / Taalportaal
bibliography glossary about
  • Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all
    Phonological evidence for cliticization
    quickinfo

    A clitic and its host word form one phonological word. The phonological evidence testifying to this word-status is the subject of this topic. It is concerned with syllabification and resyllabification facts, the occurrence of syllabic sonorant consonants, and the uninterruptability of the combination of host and clitic.

    readmore

    The assertion that a clitic and its host word form one phonological word presupposes that there are criteria for establishing the clitichood of a function word and, by the same token, the phonological wordhood of the combination of a function word and its host word. The question is therefore when the latter combination can be said to form a phonological word or, put differently, when a function word can be said to show clitic behaviour. As to this, Gussenhoven (1985-1986:180) formulates the, slightly rephrased, criterion:

    1
    Criterion for the clitichood of a function word (after Gussenhoven (1985-1986:180))
    A function word behaves as a clitic if, together with its host word, it forms a unit for which phonological conditions hold which do not hold for combinations of two non-function words [translated from Dutch].

    It does not seem too far-fetched to equate the unit for which phonological conditions hold which do not hold for combinations of two non-function words with the phonological word.

    Taking the above criterion as a point of departure, there appears to be evidence for the clitic behaviour of function words in Frisian. In the first place, there is evidence based on syllabification. Take the examples below:

    2
    Examples of the combination of a preposition and the definite article it 'the (with neuter nouns)'
    a. Oan it dak
    on the roof
    On the roof
    Oer it wetter
    over the water
    Across the water
    b. Op it dak
    on the roof
    On the roof
    Fan it wetter
    of the water
    Of the water

    The phonological word is considered to be the domain of syllabification. In the examples in (2a), the article it /ət/ the (with neuter nouns) makes up a single phonological word with, viz. cliticizes onto, the prepositions oan /oən/ on and oer /uər/ over by which it is preceded. The syllabification of oan it /oən ət/ and oer it /uər ət/ is as in the example below:

    3
    The syllabification of oan it /oən ət/ and oer it /uər ət/
    [(oə)(nət)]
    [(uə)(rət)]

    The final segments of the preposition − /n/ and /r/− form the onset of the right-most syllable, in line with the demand that a word-internal schwa syllable should have an onset (see Schwa). Since these prepositions have a long vocalic sequence here − the centring diphthongs /oə/ and /uə/−, the left-most syllable is in line with the Rhyme Constraint, according to which the rhyme of a word-internal syllable must occupy two structural positions. As to syllabification then oan it (dak) and oer it (wetter) behave just like the inflected words toanen /toən+ən/ [(toə)(nən)] tones and boeren /buər+ən/ [(buə)(rən)] farmers.

    The syllabification of op it and fan it in (2b) is as below:

    4
    The syllabification of op it /op ət/ and fan it /fɔn ət/
    [(op)(pət)]
    [(fɔn)(nət)]

    Due to short vowels of the prepositions, their final segments − /p/ and /n/− end up as ambisyllabic, that is, as both the coda of the first and the onset of the second syllable, simultaneously satisfying the onset demand for word-internal schwa syllables and the Rhyme Constraint. In this respect, the above sequences resemble the inflected words koppen /kop+ən/ [(kop)(pən)] cups and sokken /sɔk+ən/ [(sɔk)(kən)] socks.

    extra

    Of course, the above also holds for other combinations than those of a preposition and an article. Finite verbs and personal pronouns are a case in point. Take (ik) seach 'em /sɪəɣ əm/ (I) saw him, which is syllabified as [(sɪə)(ɣəm)], just like the plural form seagen /sɪəɣə+ən/ saws is syllabified as [(sɪə)(ɣən)].

    Final Devoicing, in the second place, also gives us a clue. Its effect on the realization of word-final plosives in the combination of a finite verb and a personal pronoun are exemplified below:

    5
    Examples of the effect of Final Devoicing on the realization of word-final plosives in the combination of a finite verb and a personal pronoun
    ried ek /riəd ək/ [riətək] [*riədək] drove I
    hold ek /ho:d ək/ [ho:tək] [*ho:dək] held I

    It seems to be clear that the final obstruents of the above verbs have undergone Final Devoicing. However, the syllabification of ried ek and hold ek − [(riə)(tək)] and [(ho:)(tək)], respectively − is such that verb-final /d/ ends up as the onset of the right-hand syllable, a position in which it cannot be devoiced. Devoicing thus must have taken effect here at a stage where /d/ was in coda position, i.e., when the personal pronoun ek /ək/ I had not yet been amalgamated with the verb. If the phonological word is the domain of syllabification, it is also the domain of resyllabification. Invoking resyllabification here seems to be the only way to reconcile /d/'s underlying voicedness, its devoicing in coda and its voicelessness in onset position. Devoiced /d/ in onset position thus also testifies to the the clitic behaviour of function words. See Booij (1995:174-176) for the same phenomenon in Dutch.

    extra

    bid ek /bɪd ək/ pray I is realized as [bɪtək], not as [*bɪdək]. Due to the short vowel of bid, bid ek is syllabified as [(bɪt)(tək)], with an ambisyllabic [t]. Before fusing with ek /ək/ I, the final /d/ of bid /bɪd/ pray must have undergone Final Devoicing. In putative [(bɪd)(dək)], /d/ is the coda of the left-most syllable; however, due to its link with the onset of the right-most syllable, it cannot undergo Final Devoicing.

    One and the same stem may be operated upon by inflection/derivation on the one and cliticization on the other hand; this may give rise to 'minimal pairs', examples of which are given below:

    6
    'Minimal pairs' as a result of the operation of cliticization and inflection/derivation upon the same stem
    a. Ik haw dy hoeden keard
    I have that one/those ones cautiously turned around
    I have turned that one/those ones around with caution
    huədn̩
    huədən
    a.' Ik haw dy hoed in kear stome litten
    I have that hat a time dry-clean let
    I've had that hat dry-cleaned once
    huətn̩
    huəd ən
    b. Ik haw dyn wurden seis kear ôfharke
    I have your words six times listened to
    I have listened to your words six times
    vødn̩
    vød+ən
    b.' Ik haw dyn wurd in seis kear ôfharke
    I have your word a six times listened to
    I have listened some six times to your word
    vøtn̩
    vød ən
    c. rieder
    advisor
    riədr̩
    riəd+ər
    c.' ried er
    drove he
    riətr̩
    riəd ər

    The primeless examples are realized with [d], the voicing value of which matches that of the /d/ of the underlying representations hoeden /huədən/ cautious, wurden /vød+ən/ words, and rieder /riəd+ər/ advisor. These words are syllabified as [(huə)(dn̩)], [(vød)(dn̩)], and [(riə)(dr̩)], respectively, in line with both the Rhyme Constraint and the onset demand for word-internal schwa syllables; the syllabic sonorant consonants, [n̩] and [r̩], derive from the sequences /ə{n/r}/ (see Evidence that syllabic consonants derive from /ə/ + consonant). Clearly, syllabification takes the underlying representations of these words as its point of departure. The primed examples consist of the same rows of segments as the primeless ones, but they accommodate different morpho-syntactic entities, viz. a sequence of words − like wurd in word a − versus one (complex) word − as with wurden words. If it is assumed that words are juxtaposed at the syntactic level after they have been operated upon by word-level phonology, among which Final Devoicing, it is explained why underlying /d/ is realized as [t] here (see Booij (1995:174-175)) for the same analysis of comparable Dutch cliticization facts). Apart from the difference between [d] and [t], the syllabification of the primed word sequences − [(huə)(tn̩)], [(vøt)(tn̩)], and [(riə)(tr̩)] − equals that of the primeless (complex) words, also with respect to the above demands. As with the examples in (5), devoiced /d/ in onset position is indicative of the the clitic behaviour of function words, in this case in /ən/ a(n) and er /ər/ he.

    extra

    er /ər/, the clitic counterpart of hy /hi/ or hij /hɛj/ he, only occurs in the position following a finite verb or a subordinating conjunction. If the host word ends in a vowel or a sonorant consonant, er has the variant form der /dər/, as in the following examples:

    i
    Dat koe er/der net
    that could he not
    He couldn't do that
    Dêr kaam er/der oan
    there came he on
    There he came

    The use of der is obligatory in case the host word ends in /r/. This is nicely illustrated by the verb-subject sequence waard er became he, realized as either [va:tr̩] or [va:dr̩]. The realization with [t] derives from /va:d ər/ through Final Devoicing, cliticization of er onto [va:t], and resyllabification of [t] from the coda of waard to the onset of er (note that the r of waard is purely orthographic). The verb wurde /vød+ə/ to become has the variant form wurre, with the present tense stem stem wur /vør/ and the past tense stem waar /va:r/. The past tense stem, which is not inflected, can only be followed by der, not by er. Since /r/ is deleted when preceding /d/, waar der is realized as [va:dr̩].

    extra

    Frisian has three voiced-voiceless plosive pairs, viz. /b/ ~ /p/, /d/ ~ /t/, and /ɡ/ ~ /k/. It is, however, only the pair /d/ ~ /t/ which can provide evidence here. The bilabial plosive /b/ has a low rate of occurrence, and there are no verbs of the appropriate classes − weak I and strong/irregular (see Inflectional classes) − with a stem ending in /b/; the pair /b/ ~ /p/ doesn't count, though this is purely accidental. This is different for the pair /ɡ/ ~ /k/; since the velar plosive /ɡ/ only occurs in syllable-initial position (see The dorsal obstruents), the absence of verbs with stem-final /ɡ/ can be accounted for on independent grounds.

    extra

    The voicing behaviour of the /d/ of /d/-words after a fricative is treated in Progressive Voice Assimilation: function words beginning with /d/). The pattern appears to be fairly clear: a voiceless cluster only occurs if a) the host word ends in a voiceless fricative preceded by a short vowel and b) the /d/-word has the vowel schwa. If the /d/-word has a full vowel, the outcome is a voiced cluster. This also holds in case the host word ends in a voiced fricative preceded by a long vowel, no matter whether the /d/-word has schwa or a full vowel.

    In the third place, word-level phonological processes are expected to apply within the combination of a host word and a clitic. The occurrence of syllabic sonorant consonants is a case in point (see Syllabic sonorant consonants). This is exemplified below (see also the cases in (6) above):

    7
    Examples of the occurrence of syllabic sonorant consonants in combinations of a host word and a clitic
    a. Dat hat er sein
    /hat ər/
    [hatr̩]
    that has he said
    He said that
    b. Hja rôp 'em deryn
    /rɔ:p əm/
    [rɔ:pm̩]
    she called him therein
    She called him in
    c. op in kear
    /op ən/
    [opm̩]
    on a time
    one day

    The above combinations of host word and clitic are syllabified as [(hat)(tr̩)], [(rɔ:)(pm̩)] and /(op)σ(pm̩)σ/: the final segment of hat /hat/ has, rôp /rɔ:p/ called, and op /op/ on has undergone resyllabification ((8b)) or ambisyllabification ((8a,c)), in line with a) the Rhyme Constraint and b) the demand that a word-internal schwa syllable have an onset (note again that a schwa syllable underlies the syllabic sonorant consonants). The occurrence of the syllabic sonorant consonants therefore also testifies to the the clitic behaviour of function words.

    extra

    The impersonal pronoun men /mən/ one never shows up with syllabic /n/, as the following examples illustrate:

    i
    Dy hiemen dêr
    /hiəm+ən/
    [hiəmm̩]
    those properties/premises there
    Those properties/premises over there
    Dy hie men dêr net
    /hiə mən/
    [hiəmən/*hiəmm̩]
    those had one there not
    Those ones weren't available there
    Krekt as immen dy't ...
    /ɪmən/
    [ɪmm̩]
    just like someone who ...
    Just like someone who ...
    Krekt as is men dy kwyt
    /ɪs mən/
    [ɪmən/*ɪmm̩]
    just like is one that one/those ones lost
    As if one has lost that one/those ones

    Like men, the adverb der /dər/ there also has the vowel schwa and an initial consonant, but unlike men it does allow for a syllabic realization of its final sonorant consonant. The above behaviour thus seems typical of men. The fact that men is able to bear stress, despite its vowel, may be responsible for this.

    Finally, a phonological word constitutes an uninterruptable unit. This also holds of the combination of a host word and a clitic, as is illustrated below:

    8
    Examples of the uninterruptability of a host word plus clitic combination
    a. Dat kin ... hy/*er wol dwaan
    that can ... he all right do
    Let me think ... he can do that
    b. Set de ljedder mar tsjin ... de/*'e muorre
    put the ladder just against ... the wall
    Let me think, just put the ladder against the wall
    c. Dat koe, tinkt my alteast, sij/*se sa sein hawwe
    that could, thinks me at least, she so said have
    She might have said it like that, at least that's what I think
    d. Soks noasket, soe ik tinke, jim/*jem net
    such a thing pleases, should I think, you (object form, plural, familiar and polite) not
    You won't be pleased by such a thing, I should think

    If a content word and a function word are separated by a pause ((10a,b)) or a parenthesis ((10c,d)), only the full form of the function word is allowed. As to the above cases this means that hy /hi/ he, de /də/ the, sij /sɛj/ she, and jim /jɪm/ you (object form, plural, familiar and polite) are allowed, to the exclusion of their clitic counterparts er /ər/, 'e /ə/, se /sə/, and jem /jəm/. Cases like those in (10) therefore lend support to the assertion that a clitic and its host word form one phonological word or, the other way around, they testify to the the clitic behaviour of the above function words.

    extra

    An apposition is a phrase which indicates how a preceding phrase should be interpreted in the context given. As an independent intonational unit, it always has a certain prominence. This may be the reason that a clitic cannot be the 'antecedent' of an apposition, as exemplified below:

    i
    Examples of clitics not being able to be the 'antecedent' of an adposition
    Dat koe hij/*er, de heit, him net begripe
    Dat koe hja/*se, de mem, har net begripe
    Ik seach him/*'em, myn buorman, dêr wol stean
    Dat kinne wij/*wi, de erfgenamten, sels wol regelje

    hij /hɛj/ he, hja /ja/ she, him /hɪm/ him, and wij /vɛj/ we are the independent forms of the personal pronouns concerned, whereas er /ər/, se /sə/, 'em /əm/, and wi /vi/ are the clitic counterparts. The former can act as the antecedent of an apposition, the latter cannot.

    References
    • Booij, Geert1995The phonology of DutchOxfordOxford University Press
    • Booij, Geert1995The phonology of DutchOxfordOxford University Press
    • Gussenhoven, Carlos1985Over de fonologie van Nederlandse cliticaSpektator15180-200
    • Gussenhoven, Carlos1985Over de fonologie van Nederlandse cliticaSpektator15180-200
    About this article
    Author(s): Willem Visser
    Category: Frisian Phonology
    Keywords: cliticization
    Version: ,
    Version history:
    version editor(s) date remarks
    printreport errorcite
    Jump to:
    X