• Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all
Univerbation
quickinfo

Univerbation (also called coalescence (Haspelmath 2002); often called Zusammenrückung or Zusammensetzung in German; conglomorés in French – see Lehmann 2018) is the historical process in which phrases arbitrarily change into (complex) words. In as such, it is epitomised in Givón's aphorism that "today's morphology is yesterday's syntax" (Hopper and Traugott 2003:25 ). Acknowledging that different theoretical viewpoints exist (see discussion below), the general point of departure here is that univerbation is not strictly speaking a word formation (i.e. morphological) process, but rather one of lexicalisation (most often with orthographical implications). Put simply, if an existing, productive morphological process cannot sanction the formation of a given word, one might consider univerbation as the underlying force of such a lexical item.

The process is an important source in the diachronic development of, among others, conjunctions, prepositions, interjections, and adverbs (especially linking adverbs, also sometimes called discourse particles; and modal adverbs, also sometimes called modal particles – see Van der Wouden 2002). Examples are:

1
Conjunction
deur_dat
through_that
because
2
Preposition
onder_deur
under_through
underneath
3
Interjection
dankie < dank u
thank.you
thank you
4
Adverb
voort_aan
[vor'tan]
forth_on
from now on, in the future
[Note that syllabification in univerbations need not respect the original word boundaries.]
5
Linking adverb (a.k.a discourse particle)
kort_om
short_around
in brief, in a word
6
Modal adverb (a.k.a. modal particle)
so_waar
so_true
truly
extra
Note on glossing notations

On the Taalportaal website, we use the middle dot ( · ) to indicate affix boundaries (following Bauer 2003), and the plus sign ( + ) for word boundaries in compounds. For the sake of clarity and differentiation with compounds, the underscore ( _ ) is used to indicate word boundaries in univerbations on this page only (elsewhere we use the plus sign again). Since the hyphen ( - ) plays an important role in the orthography of Afrikaans, it is always treated as a grapheme; in some cases, for the sake of brevity, we also use the hyphen as a demarcation symbol, e.g. the orthographic form pa-hulle might be glossed as dad-3PL instead of pa·-·hulle dad·LK·3PL.

readmore
[+]Univerbation as a diachronic process

Since univerbation is seen as "the merging of two (or more) words due to their frequent co-occurrence in discourse" (Bauer, Lieber and Plag 2013:442), it is often not easy to distinguish univerbation from compounding (Lehmann 2018; cf. also Bauer 2017:49-51; Brinton and Traugott 2005:48). As a matter of fact, compounding in Indo-European languages is aptly explained if considered a result of univerbation in Indo-European (see Kastovsky 2009:328-31 for a lengthy discussion). Univerbation therefore also illustrates the interface between syntax, morphology, and orthography eloquently: frequently collocating syntactic words can become a complex morphological word, which is written/styled as a single orthographic word in a language like Afrikaans. Hence, it is on the one hand possible that a particular compounding construction could develop from a set of similar, lexicalised univerbations (e.g. the case of ADJ+N attributive compounds in Afrikaans), or on the other hand that a univerbation targets a result already represented by a type of complex words of the system (e.g. complex conjunctions in Afrikaans) (Lehmann 2018:29).

extra

Hopper and Traugott (2003:135) have a somewhat more restricted use of the term univerbation: "Morphologization involves the creation of a bound morpheme (i.e. an affix) out of an independent word by way of cliticization. The final stage of this process, the uniting of the affix with its stem, is referred to as 'univerbation'." According to them, a case in point in the Germanic languages would be the development of the past tense dental suffix from a form of doen to do. Hickey (2003:94), on the other hand, used the term univerbation to refer to "a more general structural shift in the language by which phrases are reduced to single words", giving English examples such as they overnighted for they stayed overnight, to minibus for to take the minibus, and separated (father) for (father) who is separated from his wife.

Similarly, many nominal compounds in Germanic languages have a phrasal origin. For instance, the Afrikaans compound koning·s·huis king·LK·house royal family was originally a phrase in Dutch, e.g. 't Koning·s Huis op 't Loothe king·GEN house on the LooHet Loo Palace. Through semantic extension the literal meaning palace was extended to now mainly referring to royal house; royalty, and was hence increasingly written as a single word. The now obsolete genitive case ending -s (which is still productive in Dutch) became "trapped" inside the word, and was reinterpreted as a (semantically empty) linking element. In this way, univerbation was also responsible for a large number of linking elements that became part of the compounding in Afrikaans (and Dutch) (see Booij 2009).

In this grammar of Afrikaans, we consider univerbation to be a historical process, a process in language change. A clear case is Afrikaans complex adverbs that were historically prepositional phrases, with remnants of the obsolete case system still visible. Compare for example the complex adverb te_gelyk·er_tyd [[te](PREP)[[gelyk](ADJ)[er](DAT)](ADJ)[tyd](N)](ADV) at_same·DAT_time simultaneously, with the adverbial phrase te enig·er tyd at any·DAT time at any moment/time. While the former is written as a single morphological word, the latter is still written as a phrase (i.e. as three separate words), illustrating the arbitrariness of univerbation as a historical process.

Another prominent group of lexicalised, complex words that can be motivated in terms of univerbation, is so-called pronominal adverbs(Van Riemsdijk 1978), like daar_op PN_on on it in example (7a). These pronominal adverbs are formed by replacing a preposition (op) followed by a pronoun (dit it with a locative adverb (the pronoun – either daar there, or hier here) joined with a prepositional adverb (the preposition). In example (7b) we see the alternative construction (i.e. with a prepositional phrase instead of a pronominal adverb), which is still considered ungrammatical by most prescriptivist sources (e.g. Carstens 2018:72; Spies and Combrink 1994:36).

7
a. Identifiseer jou prioriteite en fokus daar_op.
identify your priorities and focus PN_on
Identify your priorities and focus on it.
VivA-KPO
b. ?Fokus op dit en gee ons eerder 'n behoorlike tegniese verslag.
focus on it and give us rather a proper technical report
Focus on that, and rather give us a proper technical report.
VivA-KPO

Complex referential pronouns (see the topic on R-pronouns) is another group of lexicalised, complex words that can be motivated in terms of univerbation. Similar to pronominal adverbs, complex R-pronouns consist of a relativising pronominal element waar where joined with a preposition (which, since it follows the pronoun, is termed a postposition in such a construction), as in example (8).

8
... die plan van aksie waar_op daar ooreengekom is ...
... the plan of action REL_on there agreed be.AUX.PASS.PST ...
... the plan of action which was agreed upon ...
VivA-KPO

Similar to separable complex verbs (see discussion below), these complex pronominal adverbs and R-pronouns are separable by other syntactic elements, which illustrates the phrasal nature of such univerbations. Compare examples (9) and (10).

9
"daarop" split by "bo"
... wat jy so maklik kaalvoet kon uitklim tot daar bo op die top.
... that you so easy barefoot can.PST climb to there above on the top
... that you could easily climb barefooted to the very top.
VivA-KPO
10
"waarop" split by "sy"
... en alles waar sy op sit, sal onrein wees.
... and everything where she on sit shall impure be
... and everything whereupon she sits, will be impure.
VivA-KPO
[+]Univerbation vs. productive morphological processes

The syntactic basis of univerbations is clearly illustrated by the separability of separable complex verbs (called skeibare saamgestelde werkwoorde or samekoppelings in Afrikaans grammars). A prototypical example is the combination of an adverb (e.g. weg away) with a verb that it frequently collocates with (e.g. sleep to drag). While adverbs and verbs normally co-occur in predicates as separate, individual words (like agterna behind and sleep drag in (11)), it might happen that they become a single morphological word through univerbation, as illustrated in (12). A key feature of such a separable complex verb is that it can be separated by other words in main clauses in the present tense (as in (13)), and that the past participle form of the verb is formed on the verbal stem only, and not on the combined form (as in (14)). Due to this frequently occurring process of univerbation, the formation of separable complex verbs has become a productive morphological process in Afrikaans, and such verbs are synchronically best analysed as compounds (e.g. weg+sleep).

11
... repe velle wat agterna sleep
... strips hide that behind drag
... strips of hide that drag behind
VivA-KPO
12
... mense wat sakke weg_sleep
... people who bags away_drag
... people who drag bags away
VivA-KPO
13
Hulle sleep ou ysters weg.
they drag old irons away
They drag old iron away.
VivA-KPO, adjusted
14
a. Die kleinstes van die kudde sal weg_ge·sleep word.
the smallest of the herd will.AUX away_PST·drag become.AUX.PASS.PRS
The smallest members of the herd will be dragged away.
VivA-KPO, adjusted
a.' *ge·weg_sleep
extra

Several inseparable complex verbs, such as om·singel around·moat to surround; to besiege, derive from separable complex verbs, but are no longer separable. For example, the following type of usage (given in the WNT) was possible in Middle Dutch So quam die gloriose Here Jezus om·ge·cingel·t met grote scharen der englen thus came the glorious Lord Jesus around·PTCP.PST·moat·PTCP.PST with great multitude the.GEN angels thus came the glorious Lord Jesus encircled by great multitudes of angels. In Afrikaans the past participle of omsingel is ø·omsingel, instead of the more regularly formed *om·ge·singel.

Analogous to the formation of separable complex verbs, is the formation of infinitival nominals (or bare-inf nominalisations), based on a verb phrases consisting of a noun and a bare infinitive verb (without a determiner), and which can function both as a noun and verb. Numerous examples of the kind [[x](N)[y](V)](N|V) can be found in Afrikaans corpus data, e.g.:

  • TV-kyk TV-watch TV watching; to watch TV
  • vleis_braai meat_grill barbecue; to barbeque
  • perd_ry horse_ride horseback riding; to ride horses
  • kar_was car_wash car washing; to wash a car
  • boek_lees book_read reading books; to read a book (see example (15 below)
Cases like these would traditionally best be analysed as univerbations, since verbal compounding is in theory not productive in Germanic languages. However, since this construction is assumedly productive in Afrikaans, we could say that univerbation motivated the formation of such constructions, with the result that this kind of verbal compounding ( [[x](N)[y](V)](V), and subsequent conversion to a noun) is increasingly productive in Afrikaans. It would therefore be apt to analyse novel cases as compounds, rather than as univerbations.

15
... nét die plek vir mediteer of boek_lees.
... just the place for meditate or book_read
... just the place for meditation or reading books.
VivA-KPO
extra

Note that novel compounds of the kind [[x](N)[y](V)](V) correlates strongly with (a) the frequency of collocation; and (b) the length of the components, which suggests that the development from univerbation to compounding is perhaps not yet fully completed. Compare for instance the compound vrou+wees woman+be womanhood; to be a woman in example (16) below, with the infrequent combinations of getuie witness with wees to be, and verhaal story with vertel to tell in example (17). The latter cases are not compounded to form verbs or infinitival nominals (contrary to some of the examples listed above), due to either their infrequent occurrence and/or the length of the components.

i
Sy het onder meer oor vrou_wees gesels ...
she have.AUX under more about woman_be talk
She has talked about womanhood, among other things ...
ii
Getuie wees is meer as net 'n verhaal vertel.
witness be is more than only a story tell
Being a witness is more than only telling a story.
VivA-KPO

The best case in point of univerbation developing into a productive morphological process, is the difference between [(NP) ADJ N] phrases and [[x](ADJ)[y](N)](N) compounds. AWS-11 (rule 15.32) spells out that such phrases can become compounds due to semantic narrowing, as illustrated by the minimal pairs in (18) and (19).

16
a. sagte vrugte
soft fruit (in contrast to firm fruit)
[(NP) sagte vrugte]
a.' sagte+vrugte
soft fruit (like strawberries and blueberries, in contrast to top fruit, like apples and pears)
[[sagte](ADJ)[vrugte](N)](N)
17
a. wilde bees
wild/untamed bovine
[(NP) wilde bees]
a.' wilde+bees
gnu
[[wilde](ADJ)[bees(N)](N)

Demarcation is a recurring problem in the study of univerbations: when can we say that a fixed phrase has attained word status? Although spelling cannot be taken to be the decisive factor (since spelling is arbitrary to a certain extent), spelling could reflect how users perceive or interpret such combinations. Compare for example minimal pairs where a spelling difference (with or without a white space) corresponds arbitrarily to a meaning difference, as in agter mekaar in (20), vs. agtermekaar in (21):

18
... omdat die fietsryers nie in 'n streep agter mekaar ry nie.
... because the cyclists not in a row behind each.other ride PTCL.NEG
... because the cyclists don't ride in single file.
VivA-KPO
19
My vakansie+planne is agter_mekaar.
my holiday+plans are behind_each.other
My holiday plans are organised.
VivA-KPO

Various output types of univerbation can be distinguished in Afrikaans. The following lists are not exhaustive.

[+]Output type: Conjunction

  • Subordinate conjunctions
    • hoe_wel how_well although
    • of_skoon or_though although
    • soos < Dutch zo_alssuch_iflike
    • as_of as_if as if
    • as_ook as_also as well as
    • in_die·n in_that·DAT if
    • na_mate to_measure in so far as
  • Subordinate conjunctions with a verbal base
  • Subordinate conjunctions ending in "dat"
    (which appears to be a case of regularisation cum specialisation)
    • deur_dat through_that in that
    • na_dat after_that after
    • om_dat to_that because
    • voor_dat before_that before
    • op_dat up_that so that
    • so_dat so_that so that
      • Note in contrast that combinations like sonder dat without that without (example (22)), and as dat as that than (example (23)) are still considered to be more phrasal than lexicalised.
        20
        Amptenare word bevorder sonder dat hulle oor die nodige ondervinding en kwalifikasies beskik.
        officials become.AUX.PASS.PRS promote without that they over the necessary experience and qualifications dispose
        Officials are promoted without having the necessary experience and qualifications.
        VivA-KPO
        21
        Dit is beter dat een mens sterf as dat die hele mensdom vergaan.
        it is better that one person die as that the whole mankind
        It is better for one man to die than for all mankind to perish.
        VivA-KPO
    extra

    • The WNT suggests that the conjunction voordat before may have developed from the preposition voor for; before via an intermediate step voor dien dat before that which with a cataphoric die·n that·DAT/ACC.
    • The word behalwe except, besides, which functions both as a preposition and a conjunct (Komen 1994:Ch. 6), derives from a prepositional phrase (Etymologiebank).
    • The conjunct maar but has a phrasal origin, viz. ne warenot be(-it) (Etymologiebank).

[+]Output type: Preposition

Afrikaans has many complex prepositions, e.g., teen_oor against_over across, and onder_deur under_through underneath, to name but a few. The distinction between compounding and univerbation may seem arbitrary – even impossible – here. Following the assumed argumentation of Kempen (1984:11), we could attempt to distinguish between these two processes in the following way:

  • Compounding: The right-hand component is the semantic head, so that bo-op above-on on top of always also at least means on, and hence that the complex preposition (bo-op) can be replaced by the right-hand component (op). Compare for instance (24a) with (24b):
    22
    a. … die jong leeus wat bo-op die berg saamgetrek het ...
    the young lions that above-on the mountain gather.PST have.AUX
    … the young lions that have gathered on (top of) the mountain ...
    VivA-KPO
    b. … die spesie wat op die berg aangetref is ...
    the specie that on the mountain find.PST be.AUX.PASS.PST
    the specie that was found on (top of) the mountain ...
    VivA-KPO
  • Univerbation: The right-hand component cannot be clearly identified as the semantic head, so that bo_aan above_on at the top of does not always also means on, and hence that the complex preposition (bo_aan) cannot always be replaced by the right-hand component (aan). Compare for instance (25a) with the ungrammatical constructed example in (25b):
    23
    a. Iran is bo_aan die lys.
    Iran is above_on the list
    Iran is at the top of the list.
    VivA-KPO
    b. *Iran is aan die lys.
    Iran is on the list
extra

Also see Bauer, Lieber and Plag (2013:453) with reference to English examples, and Neef (2009) regarding German. Kempen (1984:6) provided a comprehensive list of simplex and complex prepositions in Afrikaans; Ponelis (1979:171) presented a smaller list.

[+]Output type: Adverb

In addition to pronominal adverbs like daar_op there_on on that discussed above, Afrikaans has numerous proper adverbs (all inherited from Dutch) that are the result of univerbation. Consider the following examples (categorised in different main semantic categories):

  • Modal adverbs
    (a.k.a. stance adverbs, or modal particles; see Van der Wouden (2002) with reference to Dutch)
    • des_al_nie_te_min the.GEN.M_already_not_too_less nevertheless
    • des_nood·s the.GEN.M_need·GEN if necessary
    • in_der_daad in_the.DAT.F_deed indeed
    • sommer < so_maar so_but merely
    • so_waar so_true truly
    • wel_is_waar well_be_true admittedly
  • Temporal adverbs
    (a.k.a. adverbs of time)
    • agter_af after_off afterwards
    • al_weer all_again once again
    • laas_jaar last_year last year
    • laas_week last_week last week
    • nog_maal·s more_time·GEN again
    • onder_tussen under_between meanwhile
    • van_dag of_day today
    • van_dees_maand of_this_month this month
    • van_jaar of_year this year
    • wel_een·s well_once·GEN once; sometimes
  • Spatial adverbs
    (a.k.a. adverbs of place)
    • agter_op behind_on in/on the back
    • dié_kant this_side this side
    • een_kant one_side alone
    • noorde_kant north_side northern side
    • onder_weg under_way on the way; en route
    • voor_uit front_out forward, ahead

Also note these examples of linking adverbs (a.k.a. discourse particles; see Van der Wouden (2002) with reference to Dutch):

  • as_ook as_also as well as
  • by_voorbeeld by_example for example
  • daar_enteen then again < Middle Dutch daar_entegenthere_against
  • des_nie_teen_staande the.GEN.M_not_against_standing notwithstanding
  • en_so_voorts and_so_forth etcetera
  • kort_om short_around in brief, to summarise
  • ten_minste to.the.DAT.M_least at least

extra

In a number of word groups functioning as adverbs, the semantics are non-transparent and the distribution is idiosyncratic, which can be taken as arguments in favour of word status. Nonetheless, examples like the following remain phrases (rather than conjunctively written words):

  • en so meer and so more etcetera
  • let wel heed/attend well note
  • met name with name namely
  • ook maar also mere also; even
  • selfs maar even mere even
  • vir seker for sure surely, absolutely

Also compare the following two contrastive sentences. In (26a) adverbial phrases are used to express the same meaning as the adverbs in (26b), all of which came into being through univerbation.

i
a. Hulle is al twee al weer op pad huis toe.
they are all two all again on road house towards
b. Hulle is al_bei weer_eens onder_weg huis_waarts.
they are all_both again_once under_way home_turn.ADVZ
They are both once again on their way home

  • Note that the root -bei- comes from Proto-Germanic *bai or *bhoiboth. When combined with the distributive quantifier al(le) all, the combination is written as one word (al_bei both). However, when al co-occurs with a standard cardinal, it is written as two words, as in al twee all two both, the two of them. The prefix aller- is the old genitive plural of alle that has developed into a category-neutral intensifier prefix that combines with superlative forms of adjectives, as in aller·moeilik·ste CN·difficult·SUPL most difficult.
  • Also note that al weer all again once again is written conjunctively in Dutch as alweer. While WAT distinguishes in meaning between al weer and alweer, AWS-11 only recognises al weer.
  • The adverbialising suffix -waarts has its origins in Middle Dutch wertturn; become, plus an adverbialising -s. In 17th and 18th century Cape Dutch (Afrikaans) we find several adverbial phrases where -waarts is still used as an independent word, e.g. Batavia waartstowards Batavia, Caap waartstowards the Cape, lande waartstowards the land, beneeden waartsdownwards, etc. (VivA-KPO).
)

One should be careful not to assume univerbation in all cases that seem complex and "historic".

  • An adverb like nogtans nonetheless comes from the Middle Dutch compound noch+danneither+than, which was written mostly as nochtan(ne). By means of the adverbialiser -s, the word was derived as an adverb.
  • Similarly, the adverb al_tans at least comes from Dutch al_thans < al te hand·sall to hand·ADVZ, which should be seen as a parasynthetic compound, rather than a univerbation.
  • The adjective/adverb agterbaks underhand, secret(ly) is also a parasynthetic compound: from a prepositional phrase and the adverbialiser -s (Etymologiebank).

References
  • Bauer, Laurie, Lieber, Rochelle & Plag, Ingo2013The Oxford Reference Guide to English MorphologyOxford University Press
  • Bauer, Laurie, Lieber, Rochelle & Plag, Ingo2013The Oxford Reference Guide to English MorphologyOxford University Press
  • Booij, Geert2009Morphological analysisHeine, Bernd & Narrog, Heiko (eds.)The Oxford handbook of linguistic analysisOxfordOxford University Press
  • Brinton, Laurel J. & Traugott, Elizabeth Closs2005Lexicalization and language changeCambridge UKCambridge University Press
  • Hickey, Raymond2003Tracking lexical change in present-day EnglishAndrew Wilson, Paul Rayson and Tony McEnery (ed.)Corpus Linguistics by the Lune. A Festschrift for Geoffrey LeechFrankfurtPeter Lang93-105
  • Hopper, Paul J. & Traugott, Elisabeth C2003GrammaticalizationCambridge U.K.Cambridge University Press
  • Hopper, Paul J. & Traugott, Elisabeth C2003GrammaticalizationCambridge U.K.Cambridge University Press
  • Kastovsky, Dieter2009Diachronic perspectivesLieber, Rochelle & Stekauer, Pavol (eds.)The Oxford handbook of compoundingOxfordOxford University Press323-340
  • Komen, J.A.M1994Over de ontwikkeling van absolute constructiesUniversity of AmsterdamThesis
  • Ponelis, Frits A1979Afrikaanse sintaksisPretoriaJ.L. van Schaik
  • Riemsdijk, Henk C. van1978A case study in syntactic markedness: the binding nature of prepositional phrasesPeter de Ridder Press
  • Wouden, Ton van der2002Partikels: naar een partikelwoordenboek voor het NederlandsNederlandse Taalkunde720-43
  • Wouden, Ton van der2002Partikels: naar een partikelwoordenboek voor het NederlandsNederlandse Taalkunde720-43
  • Wouden, Ton van der2002Partikels: naar een partikelwoordenboek voor het NederlandsNederlandse Taalkunde720-43
printreport errorcite