/ instituut voor de Nederlandse taal / Taalportaal
bibliography glossary about
  • Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all
    Cliticization
    quickinfo

    In many languages, certain elements play an independent grammatical (syntactic) role, in which sense they are word-like, whereas they show dependent phonological behaviour, in which sense they are affix-like. Take the element er /ər/ he in the following sentence:

    1
    Dát hat er net sein
    thát has he not said
    That's not what he said

    Syntactically speaking, er is a full-fledged word: it occupies the canonical subject position and it bears the thematic role of agent. In this respect, there is no difference between er and a Noun Phrase (NP) like ús omke my uncle (lit.: our uncle).

    Phonologically speaking, however, things are quite different, for er forms one word with the finite verb which precedes it. This is clear from the realization of hat er /hat ər/ has he, viz. [hatr̩]. Syllabic sonorant consonants (see Syllabic sonorant consonants) only show up in phonological words, so the sequence hat er, which is not a syntactic constituent, behaves as one word phonologically.

    Linguistic elements with such a dual status are called clitics and the process of phonological amalgamation with other words in which they partake is called 'cliticization'. This section is devoted to clitics and cliticization in general, illustrated with examples from Frisian.

    readmore

    Anderson's (2011) article on clitics begins as follows: The notion of "clitic" derives from one of the oldest problems in the study of language: how to define the "word." Grammarians have long noted that a difficulty is posed in this area by the fact that certain elements in many languages seem to play an independent role in the grammatical structure of sentences, and thus to warrant the status of "grammatical words," but in terms of their sound structure form parts of unitary "words" (in a distinct, phonological sense) with other "grammatical words." Most elements with such a dual status are so-called 'little words'. They do not have their own, independent accent, but lean on an adjacent word in order to become part of an accentual unit of some sort. That is why such an element is called a 'clitic', a word deriving from Greek klitikosleaning, which in turn derives from klîneinto lean (Anderson (2011:2003, footnote 2)).

    According to Anderson (2011:2004), however, lack of stress cannot be the defining property of a clitic, since some clitics may occur in stressed positions, as long as this is in line with the overall prosodic organization of the language concerned. Anderson therefore proposes to analyze clitics as "prosodically deficient elements". The segmental content of a 'full word' can be assumed to be organized into syllables, feet and phonological words from the outset. The phonological organization of a clitic, on the other hand, does not go further than the syllable or the foot. A clitic thus does not have the status of a phonological word, which is why it must be incorporated into an adjacent one.

    Being 'little words', clitics are confined to specific categories, as stated clearly in Booij (1995:165): Clitics are function words such as pronouns, determiners, auxiliaries, particles, conjunctions, and prepositions which are phonologically dependent on a host word to which they attach, and with which they form a prosodic constituent.

    Most function words have both a 'strong' and a 'weak' form, that is, they can function as both a full word and a clitic. The strong form has an independent phonological status, viz. that of a phonological word, to which end it must contain a full vowel. Conversely, the weak form has a dependent phonological status, that of a syllable or a foot.

    Due to its dependent phonological status, a clitic is not able to stand on its own; this implies that it cannot form a one word-expression, which is exemplified below:

    2
    Examples of clitics not being able to form a one word-expression
    a. Wa wol dêrhinne? Ik! / Ikke! / *ek / *'k!
    [ɪk / ɪkə / *ək / *(ə)k]
    who wants there to I
    Who is willing to go there? Me!
    b. Wa tinksto dat ik op it each haw? Him! / *'em! / har! / *se! / mij! / *mi! / dij! / *di!
    [hɪm / *əm / har / *sə / mɛj / *mi / dɛj / *di]
    who think-you (SG, familiar) that I on the eye have him / her / me / you
    Who do you think that I have in mind? Him! / Her! / Me! / You!
    c. Wa wolle dêrhinne? Wij! / *Wi! / *We!
    [vɛj / *vi / *və]
    who want there to? We!
    Who are willing to go there? We!

    A phonological word must minimally have a full vowel, so it comes as no surprise that ik /ɪk/ I in its vowelless variant /k/ is out as a one word-expression (see (2a)). And since schwa cannot bear stress, it does not come as a surprise either that 'em /əm/, se /se/, and we /və/, which have schwa as their only vowel, are out here as well (see (2b) and (2c)). That these words cannot bear prominence is a direct consequence of their phonological make-up.

    However, despite their full vowels, the words mi /mi/ me ((2b)) and wi /vi/ we ((2c)) cannot be stressed either. Assuming that they do not count as full words, that is, do not constitute a phonological word, provides a formal, 'technical' explanation for the latter. But that does not explain why these clitics can have a full vowel at all. Wouldn't the vowel schwa, being an unambiguous indication of unstressability, be a more obvious alternative? The pronoun pairs dij ~ di you (object, sg., familiar), mij ~ mi me and wij ~ wi we show the alternation /ɛj ~ i/ (see Personal pronouns with /ɛj/ and their clitic allomorph /i/ in Klaaifrysk). It is not surprising that the alternants with /ɛj/ count as the full forms, for complex nuclei, like falling diphthongs, attract stress (prominence). Since it seems redundant to have two strong forms of one and the same function word, the one with the monophthong, though this is a full vowel, counts as the weak form. Such a form might be called a 'clitic by choice'.

    extra

    Having two weak forms seems to pose less of a problem, for alongside wi /vi/ we, there is the form we /və/; however, *me /mə/ and mi me and *de /də/ and di you (object form, singular, familiar) do not occur side by side.

    Unlike the members of the lexical categories (Noun (N), Verb (V), Adjective (A), Adposition (P)), function words can have schwa as their (only) vowel. Schwa being the weakest vowel of Frisian, this explains the dependent status of these words or, put differently, it explains why they cannot constitute a phonological word on their own and have to find support for their prosodification, that is, why they have to find a word which they can 'lean on' as it were (see above for the etymology of the word clitic). Such a supporting word is called a 'host word' (or simply a 'host'). When a clitic forms a prosodic constituent with a host word, it is said to 'cliticize onto its host word' or to undergo 'cliticization'.

    Function words which cliticize onto a host word to their right are called 'proclitics', while they are 'enclitics' in case the host word is on their left-hand side. The very shape of the function word may decide on the direction of cliticization. If, for instance, a clitic begins with schwa, it can only be an enclitic, since schwa is excluded in word-initial position in Frisian. A function word with a full vowel, on the other hand, is able to act as both an enclitic, a proclitic and a freestanding word.

    extra

    The dialect of Hylpen has a cliticization pattern which deviates from the one normally found in Frisian, see De Boer (1950:122-123). The personal pronoun wi /vi/ we has the clitic form we /və/. In case the latter follows a finite verb in the past tense, it may either follow the verb form as a whole, as is customary, or it may show up immediately before the plural inflectional ending -en ( /-ən/). This gives rise to pairs like the following (De Boer's transcriptions):

    i
    [hēən wə ~ hēwən] had we
    [su:odən wə ~ su:odwən] should we
    [tɔ:ətən wə ~ tɔ:ətwən] thought we
    [fitstən wə ~ fitstwən] cycled we
    [lɪbən wə ~ lɪbwən] lived we

    The pattern of the left-hand row may be taken as basic, from which the 'deviant' pattern derives, for instance: [lɪbən wə] → [lɪbwəən] → [lɪbwən] (the intermediate sequence [-əə-] in [lɪbwəən] is subject to degemination). In this specific use, we acts as a 'mesoclitic'.

    Zwicky (1977) makes a distinction between 'special clitics' and 'simple clitics'. A simple clitic a) occupies the same syntactic position as its non-clitic (independent, strong) counterpart, b) is a reduced form of this counterpart, c) behaves as a word with respect to conjunction reduction, just like this counterpart.

    Syntactically speaking, simple clitics are more word-like than affix-like, for they have the same content, meaning, and function as their non-clitic counterparts. Some personal pronouns, however, have a clitic form which can be used in a way different from the use of its independent counterpart. Examples are given below:

    3
    Examples of clitic forms in a special use
    Se /sə/ sizze, dat ...
    *Hja /ja/ sizze, dat ...
    *Sy/sij /s{i/ɛj}/ sizze, dat ...
    they say, that ... They say, that ..., it is said, that ..., rumour has it, that ...
    [se has a generic interpretation here, which hja/sy/sij cannot have]
    Do bist 'em /əm/
    *Do bist him /hɪm/
    you are him You are it
    Hy hie 'em /əm/ om
    *Hy hie him /hɪm/ om
    he had him around He had had a few (too many)
    Dat wurdt 'em /əm/ net
    *Dat wurdt him /hɪm/ net
    that becomes him not That won't work
    Dêr leit it 'em /əm/ net oan
    *Dêr leit it him /hɪm/ net oan
    there lies it him not on That's not the reason why

    Also, the clitic form of the personal pronouns seems to be called for in specific syntactic configurations, as is the case with reflexive constructions, exemplified in the example below:

    4
    The obligatory use of object clitics in reflexive constructions
    a. Skamje di / *dij!
    [di / *dɛj]
    be ashamed of you (object)
    Be ashamed of yourself!
    b. Ik skamme mi / *mij dead
    [mi / *mɛj]
    I was ashamed dead
    I was so embarrassed, it was so embarrassing
    c. Hy hat 'em / *him ferriden
    [əm / *hɪm]
    he has himself misdriven
    He took the wrong turning
    extra

    Of the verbs in (5) above, skamje be ashamed is inherently reflexive, whereas ferride take the wrong turning is not. As far as the use of the clitic form of the personal pronouns is concerned, this does not seem to make any difference.

    extra

    se, the clitic allomorph of hja/sij she, cannot be used as a reflexive pronoun. This cannot be ascribed to the fact that it has schwa as its only vowel, for 'em /əm/ him in reflexive use is fine (see the final example is (5) above)). The latter also holds for the Klaaifrysk clitic allomorphs mi /mi/ me and di /di/ you (singular, familiar), with /i/ instead of /ɛj/ (see Personal pronouns with /ɛj/ and their clitic allomorph with /i/ in Klaaifrysk). The impossibility of se to function as a reflexive pronoun leads Hoekstra (1994:54) to the assumption that se must bear structural case. His argument runs as follows: reflexive constructions form a maximal A-chain, which is assumed to allow for only one link fully specified for grammatical features, like structural case. That is why se is out as a reflexive pronoun; har/harren her/them is fine, since it is assumed to be licensable by inherent case, which se is not.

    extra

    Embedded in a verb phrase headed by an infinitive or a participle, the clitic allomorph of the object form of personal and reflexive pronouns may occur in sentence-initial position, as in the following examples:

    i
    a. Dij / di skamje, dat wie wolris in goed ding
    [dɛj / di]
    yourself feel ashamed, that was all right a time a good thing
    It would certainly be good for you to be ashamed of yourself for once
    b. Mij / mi skamje doch ik net
    [mɛj / mi]
    myself feel ashamed do I not
    I certainly am not ashamed of myself
    c. Him / 'em skamje docht er net
    [hɪm / əm]
    himself feel ashamed does he not
    He certainly is not ashamed of himself
    d. Him / 'em sjoen haw ik net
    [hɪm / əm]
    him seen have I not
    I have not seen him
    e. Him / 'em sjende, tocht ik fuort ...
    [hɪm / əm]
    him seeing thought I at once ...
    Seeing him, I at once thought ...

    Here, both the clitic and the independent form − dij [dɛj], mij [mɛj], him /hɪm/ and di /di/, mi [mi], 'em /əm/, respectively − are allowed, with a slight preference though for the former. Sentence-initial position and object clitics do not seem to go together well, in all likelihood due to the fact that a preposed element always bears some (extra) prominence.

    Conversely, there are syntactic configurations where the clitic form of the personal pronouns seems forbidden, as in the complement position of prepositions, see the example below for examples:

    5
    The ban on object clitics as the complement of prepositions
    a. Fan dij / *di / him / *'em
    [dɛj / *di / hɪm / *əm]
    Of you / him
    b. Mei dij / *di / har / *se
    [dɛj / *di / har / *sə]
    With you / her/them
    extra

    An idiosyncratic property of the preposition neffens /nɛfəns/ according to is that it does not allow for a clitic as a complement, irrespective of the syntactic context. Booij (1995:168) notes the same for Dutch volgens /ʋɔlɣəns/ according to. He says that volgensrequires prominence of its complement-noun phrase. However, the latter seems to hold, in principle, for the complement of any preposition.

    Now, neffens − and volgens as well − ends in a schwa syllable. This might make for a phonologically motivated reason for the impossibility of clitics in the complement position of this preposition. Cliticization would add another weak syllable, yielding a rhythmic lapse, which is a disfavoured configuration. There are more bisyllabic prepositions, viz. binnen /bɪnən/ inside, boppe /bopə/ above, bûten /butən/ outside, efter /ɛftər/ behind, njonken /njoŋkən/ next to, sûnder /sundər/ without, tusken /tøskən/ between, and ûnder /undər/ under. They are more reluctant in allowing for a complement DP headed by the clitic allomorph e /ə/ of the definite article de /də/ the than are monosyllabic prepositions, like fan /fɔn/ of, oer /uər/ over, and tsjin /tsjɪn/ against (see The clitic allomorph 'e /ə/ of the definite article de /də/ the).

    There is a difference between the use of an Adposition Phrase (PP) as a bare phrase − for instance, in answering a question − and a PP embedded in a sentence. If a PP stands on its own, as in (7) above, the complement position of the preposition cannot be occupied by the clitic allomorph of the object form of personal pronouns. If, on the other hand, a PP is embedded within the larger whole of a sentence, a clitic does not seem to be excluded rightaway, which is exemplified in the example below:

    6
    Object clitics as the complement of prepositions do occur
    a. Ik wol dat wol fan dij / di / him / 'em / har / se oernimme
    [dɛj / di / hɪm / əm / har / sə]
    I want that all right from you / him / her / them take over
    It's all right with me to take that over from you / him / her / them
    b. Wy sille dat mei dij / di / him / 'em / har / se opnimme
    [dɛj / di / hɪm / əm / har / sə]
    we will that with you, him, her/them take up
    We will contact you / him / her / them about that

    In (8), both the independent and the clitic form of the pronouns are allowed, provided the pronouns do not receive (extra) prominence.

    extra

    It should be noted, though, that fan se from her/them and mei se with her/them sound worse than the combinations with the other object clitics. This cannot be accounted for with an appeal to the fact that se has (unstressable) schwa as its (only) vowel; the latter also holds of 'em /əm/, and fan 'em from him and mei 'em with him do not sound markedly worse than fan him and mei him, with the independent form him /hɪm/. As noted, se cannot be used as a reflexive pronoun, whereas the other object clitics can.

    extra

    Being vowel-initial, 'em may make for a smoother prosodic integration with its host. It only enforces the resyllabification (ambisyllabification) of the final segment of the preposition in order to provide the schwa syllable with an onset, as in the following examples:

    i
    fan 'em [(fɔn)(əm) → (fɔn)(nəm)] of him
    mei 'em [(maj)(əm) → (maj)(jəm)] with him

    The initial /s/ of se, on the other hand, brings much more phonological machinery in its wake: 1) following a preposition ending in a voiced segment it undergoes voicing: fan[z]e, mei [z]e, 2) in case of the /n/-final preposition fan /fɔn/, the initial /s/ of se triggers Vowel nasalization: fan se /fɔn zə/ [fɔ̃ zə].

    If the PP is preposed as a whole, only the independent form is allowed, see the example below:

    7
    PPs do not occur in preposed positions with a clitic as a complement
    Fan dij / *di / him / *'em / har / *se wol ik dat wol oernimme
    [dɛj / *di / hɪm / *əm / har / *sə]
    from you / him / her/them want I that all right take over
    It's all right with me to take that over from you / him / her/them

    In case the pronoun is preposed and the preposition remains in its basic position (Preposition Stranding), that is if the pronoun ends up in sentence-initial position, where only the independent form is allowed, as exemplified below:

    8
    Clitics do not occur in combination with a stranded preposition
    Dij / *di / him / *'em / har / *se wol ik dat wol fan oernimme
    [dɛj / *di / hɪm / *əm / har / *sə]
    you / him / her/them want I that all right from take over
    It's all right with me to take that over from you / him / her/them

    As noted before, preposing and clitics do not go together well.

    The same does not seem to hold for clitics and extraposing, see the example in (12):

    9
    Object clitics as the complement of prepositions in extraposed position
    a. Dat wol ik wol fan dij / di / him / 'em / har / se oernimme
    [dɛj / di / hɪm / əm / har / sə]
    that want I all right from you / him / her/them take over
    It's all right with me to take that over from you / him / her/them
    b. Dat wol ik wol oernimme fan dij / di / him / 'em / har / se
    [dɛj / di / hɪm / əm / har / sə]
    that want I all right take over from you / him / her/them
    It's all right with me to take that over from you / him / her/them

    In a non-extraposed PP both the full and the clitic form are allowed (see (12a)). And though the full form sounds (somewhat) better in an extraposed PP, the clitic form is by no means forbidden there (see (12b)). A preposed constituent then seems to have a higher degree of prominence than an extraposed one.

    However, matters may be more complicated than this. If, in a sentence like (12a), the sentence adverb wol immediately precedes the verb oernimme, the clitic form is not allowed to occupy the complement position of the preposition:

    10
    Dat wol ik fan dij / *di / him / *'em / har / *se wol oernimme
    [dɛj / *di / hɪm / *əm / har / *sə]
    that want I from you / him / her/them all right take over
    It's all right with me to take that over from you / him / her/them

    In (13), the PP seems to have undergone preposing as well, be it within the Verb Phrase (VP). This, however, is enough to render the occurrence of the clitic form ill-formed.

    Clitics have the same meaning and function as their independent counterparts. They are therefore expected to occur in the same syntactic positions (provided this is not rendered impossible for independent phonological and/or prosodic reasons). This squares with the facts. There is one exception, though. Personal pronouns with the function of direct object occupy the most leftward position within the VP, whereas their NP-counterparts are closer to the verb, see (14):

    11
    Examples of the special syntactic position of direct object personal pronouns
    Dat ik myn buorman [it kopke kofje] oanlange that I my neighbour the cup of coffee passed That I passed the cup of coffee to my neighbour *Dat ik [it kopke kofje] myn buorman oanlange
    Dat ik [it] myn buorman oanlange that I it my neighbour passed That I passed it to my neighbour *Dat ik myn buorman [it] oanlange
    Dat ik dêr [myn buorman] sjoen haw that I there my next.door neighbour seen have That I saw my next door neighbour there Dat ik [myn buorman] dêr sjoen haw
    Dat ik [him] dêr sjoen haw that I him there seen have That I saw him there *Dat ik dêr [him] sjoen haw

    The above distribution also holds for the clitics 't [ət] in the case of the a-example and 'em [əm] in the case of the b-example. The syntactic position of these direct object clitics, is thus not a specific, independent property, since it links up with that of their independent counterparts it and him. This shows once more that Frisian has a system of simple clitics.

    extra

    The same pattern is found in Dutch, see Berendsen (1983).

    Syntactically speaking then clitics are more word-like than affix-like. Phonologically speaking, however, the reverse holds. Enclitics with schwa as their vowel are more affix-like than word-like, that is, they have much in common with schwa-suffixes. The latter also have a dependent phonological status, which is why they act as 'cohering' suffixes. This is why Klavans (1985) considers cliticization as 'phrasal affixation'. Thus the defining feature of clitics is their phonological (prosodic) behaviour.

    extra

    Most clitics are monosyllabic ('little words'). As far as enclitics are concerned, this may facilitate their integration into prosodic structure, the unmarked metrical structure in Frisian being the trochee.

    extra

    See De Haan (1997) and Hoekstra (1997) for a syntactic analysis of cliticization facts concerning inflected subordinating conjunctions and pro-drop in Frisian.

    Clitics are both more word-like than affix-like (from a syntactic point of view) and more affix-like than word-like (from a phonological (prosodic) point of view). Cliticization therefore may result in a non-isomorphy between syntactic and prosodic structures, of which the sentences below are examples:

    12
    Examples of the mismatch between syntactic and prosodic structures in cliticization
    Dat ik 'em sjoen haw [əm] that I him seen have That I have seen him
    It skip dat dêr fart the ship that there sails The ship which sails over there
    Op dy dyk wurdt meastal te hurd riden on that road is usually too fast driven People use to drive too fast on that road

    In (15a), the clitic 'em has the function of direct object, so it is part of the same syntactic constituent as the verb on its right-hand side, the VP 'em sjoen haw him seen have. In contrast, it forms a phonological word with the subject ik I on its left-hand side. This is clear from the fact that ik 'em can be realized with a syllabic sonorant consonant − [ɪkm̩] (< /ɪk əm/) −, since the syllabic realization of sonorant consonants is a phenomenon confined to the phonological word (see Syllabic sonorant consonants).

    extra

    Due to its very shape, 'em /əm/ has to cliticize onto a host word to its left.

    In (15b), the relative pronoun dat that has the function of subject of the relative clause dat dêr fart which sails over there. However, it can form a phonological word with the noun preceding it. This is clear from the (possible) realization [skɪptɔt] (< /skɪp dɔt/); the progressive voice assimilation which can be observed here, is an intra-word phenomenon (see Progressive Voice Assimilation: function words beginning with /d/). The same holds for (15c), where op dy on that can be realized as [opti] (< /op di/). Syntactically, the demonstrative pronoun dy that is part of the NP (Determiner Phrase (DP)) dy dyk that road, but phonologically it can form one word with the preceding preposition op on.

    extra

    Having a full vowel, dat /dɔt/((15b)) and dy /di/((15c)) can form a phonological word on their own. This implies that they need not cliticize or, put differently, that they are 'accidental clitics'. If they form an independent phonological word, they induce regressive voice assimilation: [skɪbdɔt] (< /skɪp dɔt/) and [obdi] (< /op di/) (see Regressive Voice Assimilation: type 1).

    Simple clitics are said to behave as words with respect to conjunction reduction − a syntactic phenomenon −, just like their independent counterparts. This is illustrated below on the basis of we [və] we, the reduced form of wy [vi]:

    13
    Examples of conjunction reduction with wy and we 'we'
    a. Dat kin de Direkteur-Bestjoerder dwaan en dat mei de Direkteur-Bestjoerder dwaan
    Dat kin en mei de Direkteur-Bestjoerder dwaan
    that can (the managing director) and (that) may the managing director do
    The managing director is able (to do that) and (the manager director is) allowed to do that
    a. Dat kinne wy dwaan en dat meie wy dwaanDat kinne en meie wy dwaan
    that can (we do) and (that) may we do
    We are able (to do that) and (we are) allowed to do that
    a.' Dat kinne we dwaan en dat meie we dwaanDat kinne en meie we dwaan
    that can (we do) and (that) may we do
    We are able (to do that) and (we are) allowed to do that

    Special clitics − see Zwicky (1977) − have properties opposite to those of simple clitics, for they a) occupy a syntactic position which is different from that of their independent counterparts, b) stand in a relation of suppletion to these counterparts, c) behave as affixes with respect to conjunction reduction, unlike these counterparts. The last property is illustrated below, with an example from French:

    14
    An example, taken from French, of conjunction reduction with a special clitic
    a. Je connais Jean et je crains Jean
    Je connais et crains Jean
    I know (Jean) and (I) fear Jean
    I know (Jean) and (I) fear Jean
    b. Je le connais et je le crains
    *Je le connais et crains
    I him know and I him fear
    I know him and I fear him

    The full noun (proper name) Jean can be left out in conjoined structures (see (17a)), but the pronoun le him − the syntactic position of which also differs from that of Jean − cannot (see (17b)), which is indicative of le being a special clitic. As shown below, le acts as an affix in this respect:

    15
    An example of the impossibility of conjunction reduction with an affix
    Readich /rɪəd+əɣ/ en grienich /ɡriən+əɣ/ *Read en grienich
    Reddish and greenish

    The sequence read and grienich as such is well-formed. However, it can only mean red and greenish, and not reddish and greenish, so the suffix -ich cannot be left out from readich in readich en grienich under preservation of the full meaning of the sequence.

    extra

    The behaviour of a phonologically dependent suffix like -ich ( /-əɣ/) is different from that of suffixes with a full vowel, like -eftich ( /-ɛftəɣ/), see the example below:

    i
    Grieneftich en readeftich
    Grien en readeftich
    Green- and red-like

    Though grien and readeftich can also mean green and red-like, the meaning green-like and red-like is possible as well. As to conjunction reduction full vowel-suffixes thus behave like full words, as do prefixes, whether or not the latter have schwa as their vowel.

    extra

    The subject form of the personal pronoun 3rd person singular masculine, hy /hi/ or hij /hɛj/ he, has the clitic counterpart er /ər/. hy/hij and er are not phonologically related or, put differently, they stand in a relation of allomorphy to each other. That is why er is expected to count as a special clitic. But this is not the case, as the coordination facts in the example below make clear:

    i
    Dat kin de Direkteur-Bestjoerder dwaan en dat mei de Direkteur-Bestjoerder dwaan
    Dat kin en mei de Direkteur-Bestjoerder dwaan
    that can (the managing director) and (that) may the managing director do
    The managing director is able (to do that) and (the manager director is) allowed to do that
    Dat kin hy/er dwaan en dat mei hy/er dwaan
    Dat kin en mei hy/er dwaan
    that can (he) en (that) may he do
    He is able (to do that) and (he is) allowed to do that

    These facts are a strong indication that Frisian only has simple clitics.

    Another characteristic of clitics having to do with coordination is that they cannot show up in conjoined phrases, either in one or in both conjuncts, see the examples below:

    16
    Examples of the impossibility of clitics to show up in conjoined phrases
    a. Ik seach de man en de frou/him en har
    *Ik seach 'em en se/'em en har/him en se
    I saw the man and the woman/him and her
    I saw the man and the woman/him and her
    b. Ik seach de man noch de frou/him noch har
    *Ik seach 'em noch se/'em noch har/him noch se
    I saw neither the man nor the woman/neither him nor her
    I saw neither the man nor the woman/neither him nor her

    The nouns man man and frou woman can show up in conjoined phrases. This also holds for the personal pronouns him /hɪm/ him and har /har/ her which may 'replace' them. The clitic forms of these pronouns, 'em /əm/ and se /sə/, however, do not have this possibility. It seems a likely assumption that each of two conjoined phrases must have a certain degree of prominence, which does not match with clitics.

    extra

    Booij (1995:169) observes the same for Dutch. He mentions the coordination of two full pronouns (hem en haar him and her) and of their clitic counterparts (əm and ər him and her), of which the former is well-formed and the latter is not. It can safely be assumed that the combination of a full pronoun and a clitic, in either order, is also disallowed in Dutch, just as it is in Frisian.

    With respect to their phonological behaviour, clitics come in three types:

    • Due to their phonological shape, function words may have a dependent phonological status. This is the case when they have schwa as their only vowel (especially when this is their initial vowel). Such function words may be called 'inherent clitics'. Examples are 'em /əm/ him and jem /jəm/ you (subject and object form, plural, familiar and polite).
    • Function words may have a full vowel. Though this means that they can act as a free-standing, independent phonological word, some of them always act as clitics, that is, they display dependent phonological behaviour. Such function words may be called 'clitics by choice'. Examples are mi /mi/ me and wi /vi/ we.
    • Function words − only those with a full vowel − may display both dependent and independent phonological behaviour. Such function words may be called 'accidental clitics'. Examples are the demonstrative pronoun dat /dɔt/ that and the relative pronoun dy't /dit/ that, which.

    References
    • Anderson, Stephen R2011Cliticsvan Oostendorp, Marc and Ewen, Colin J and Hume, Elizabeth and Rice, Keren (ed.)The Blackwell Companion to Phonology4: Phonological InterfacesWiley-Blackwell2002-2018
    • Anderson, Stephen R2011Cliticsvan Oostendorp, Marc and Ewen, Colin J and Hume, Elizabeth and Rice, Keren (ed.)The Blackwell Companion to Phonology4: Phonological InterfacesWiley-Blackwell2002-2018
    • Anderson, Stephen R2011Cliticsvan Oostendorp, Marc and Ewen, Colin J and Hume, Elizabeth and Rice, Keren (ed.)The Blackwell Companion to Phonology4: Phonological InterfacesWiley-Blackwell2002-2018
    • Berendsen, Egon1983Objectsclitica in het NederlandsDe nieuwe taalgids76209-224
    • Boer, Bernardus de1950Studie over het dialect van HindeloopenAssenVan Gorcum & Comp.
    • Booij, Geert1995The phonology of DutchOxfordOxford University Press
    • Booij, Geert1995The phonology of DutchOxfordOxford University Press
    • Booij, Geert1995The phonology of DutchOxfordOxford University Press
    • Haan, Germen J. de1997Voegwoordcongruentie in het FriesHoekstra, E. & Smits, C. (eds.)Vervoegde voegwoorden: lezingen gehouden tijdens het Dialectsymposion 1994Cahiers van het P.J. Meertens-Instituut 9Amsterdam50-67
    • Hoekstra, Jarich1994Pronouns and Case. On the distribution of Frisian harren and se 'them'Leuvense bijdragen8347-65
    • Hoekstra, Jarich1997Pro-drop, clitisering en voegwoordcongruentie in het WestgermaansHoekstra, E. & Smits, C. (eds.)Vervoegde voegwoorden: lezingen gehouden tijdens het Dialectsymposion 1994Amsterdam68-86
    • Klavans, Judith L1985The independence of syntax and phonology in cliticizationLanguage6195-120
    • Zwicky, Arnold M1977On CliticsBloomington
    • Zwicky, Arnold M1977On CliticsBloomington
    About this article
    Author(s): Willem Visser
    Category: Frisian Phonology
    Keywords: clitic
    Version: 1.0, December 2015
    Version history:
    version editor(s) date remarks
    1.0 Willem Visser December 2015
    printreport errorcite
    Jump to:
    X