Factors that determine the omission of the Afrikaans complementiser dat "that"

Bertus Van Rooy, Haidee Kruger

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Verb complement clauses in written Afrikaans have two formal variants. One form has an overt complementiser dat "that", followed by dependent word order (dat+SXV), while the other form has no complementiser and independent word order (Ø+SVX). Previous research on Afrikaans has not yielded conclusive findings about the factors that influence the choice between the variants, although the factors conditioning the alternation have been studied extensively in English and German. Based on that research, a number of potential conditioning factors are identified, which relate to the syntactic complexity of the main clause (its subject, modification of tense and modality features of the main verb, passivisation, negation or additional modifiers between the main clause and the complement clause); the semantics, lexical choice and frequency of the verb of the main clause; and the formality of the register. This article adopts corpus linguistics as method to determine which factors best predict the choice between the variants. The data are drawn from the Taalkommissiekorpus ("Corpus of the Language Commission"), which is a 57 million word corpus of contemporary written Afrikaans. A sample of 10 084 instances of the declarative complement clause were extracted from the corpus by using 104 different verb lemmas (see Appendix A) as extraction terms. These were classified manually for their complementation pattern, and all the potential features identified as potential conditioning factors in the literature review were annotated for every valid instance in the sample. The classified data were then subjected to decision tree modelling, to yield a classification tree that identifies the most important factors related to the choice between the two variants of the declarative complement construction in Afrikaans. The results indicate that the verb of the main clause is the most important factor. A few very general verbs with high frequency, such as sê "say", weet "know" and dink "think" are strongly associated with the variant Ø+SVX. Verbs that are more specific in their meaning and less frequent tend to take the variant dat+SXV. In addition, register is an important conditioning factor for verbs that allow combination with both variants. More formal registers, such as academic writing, study guides and published popular books are more likely to make use of the variant dat+SXV, while the less formal registers of newspapers, magazines and fiction make more frequent use of the variant without a complementiser (Ø+SVX). The findings are interpreted as support for the view that thematic prominence is the underlying force behind the variability. When the main clause is very general and low in informativity, the most prominent information is contained in the complement clause, which is therefore presented without an overt complementiser and with main clause word order (Ø+SVX). However, when the main clause is thematically prominent, carries a higher information load, and demands the attention of the reader, the complement clause is overtly marked as subordinate by means of an overt complementiser and the use of dependent word order (dat+SXV).

Translated title of the contributionFactors that determine the omission of the Afrikaans complementiser dat "that"
LanguageAfrikaans
Pages102-116
Number of pages15
JournalTydskrif vir Geesteswetenskappe
Volume56
Issue number1
DOIs
Publication statusPublished - 1 Mar 2016

Fingerprint

conditioning
magazine
newspaper
semantics
linguistics
Omission
Afrikaans Language
Complementizer
language
Main Clause
Verbs
Complement Clause
Conditioning

Keywords

  • Afrikaans
  • syntax
  • word order
  • variation
  • complement clause
  • subordinator
  • complementiser
  • dependent word order
  • register
  • verb lemma
  • informativity
  • corpus linguistics

Cite this

Van Rooy, Bertus ; Kruger, Haidee. / Faktore wat die weglating van die Afrikaanse onderskikker dat bepaal. In: Tydskrif vir Geesteswetenskappe. 2016 ; Vol. 56, No. 1. pp. 102-116.
@article{697f5f4ed9374f9c924807ee012d15c4,
title = "Faktore wat die weglating van die Afrikaanse onderskikker dat bepaal",
abstract = "Verb complement clauses in written Afrikaans have two formal variants. One form has an overt complementiser dat {"}that{"}, followed by dependent word order (dat+SXV), while the other form has no complementiser and independent word order ({\O}+SVX). Previous research on Afrikaans has not yielded conclusive findings about the factors that influence the choice between the variants, although the factors conditioning the alternation have been studied extensively in English and German. Based on that research, a number of potential conditioning factors are identified, which relate to the syntactic complexity of the main clause (its subject, modification of tense and modality features of the main verb, passivisation, negation or additional modifiers between the main clause and the complement clause); the semantics, lexical choice and frequency of the verb of the main clause; and the formality of the register. This article adopts corpus linguistics as method to determine which factors best predict the choice between the variants. The data are drawn from the Taalkommissiekorpus ({"}Corpus of the Language Commission{"}), which is a 57 million word corpus of contemporary written Afrikaans. A sample of 10 084 instances of the declarative complement clause were extracted from the corpus by using 104 different verb lemmas (see Appendix A) as extraction terms. These were classified manually for their complementation pattern, and all the potential features identified as potential conditioning factors in the literature review were annotated for every valid instance in the sample. The classified data were then subjected to decision tree modelling, to yield a classification tree that identifies the most important factors related to the choice between the two variants of the declarative complement construction in Afrikaans. The results indicate that the verb of the main clause is the most important factor. A few very general verbs with high frequency, such as s{\^e} {"}say{"}, weet {"}know{"} and dink {"}think{"} are strongly associated with the variant {\O}+SVX. Verbs that are more specific in their meaning and less frequent tend to take the variant dat+SXV. In addition, register is an important conditioning factor for verbs that allow combination with both variants. More formal registers, such as academic writing, study guides and published popular books are more likely to make use of the variant dat+SXV, while the less formal registers of newspapers, magazines and fiction make more frequent use of the variant without a complementiser ({\O}+SVX). The findings are interpreted as support for the view that thematic prominence is the underlying force behind the variability. When the main clause is very general and low in informativity, the most prominent information is contained in the complement clause, which is therefore presented without an overt complementiser and with main clause word order ({\O}+SVX). However, when the main clause is thematically prominent, carries a higher information load, and demands the attention of the reader, the complement clause is overtly marked as subordinate by means of an overt complementiser and the use of dependent word order (dat+SXV).",
keywords = "Afrikaans, syntax, word order, variation, complement clause, subordinator, complementiser, dependent word order, register, verb lemma, informativity, corpus linguistics",
author = "{Van Rooy}, Bertus and Haidee Kruger",
year = "2016",
month = "3",
day = "1",
doi = "10.17159/2224-7912/2016/v56n1a7",
language = "Afrikaans",
volume = "56",
pages = "102--116",
journal = "Tydskrif vir Geesteswetenskappe",
issn = "0041-4751",
publisher = "South African Academy for Science and the Arts",
number = "1",

}

Faktore wat die weglating van die Afrikaanse onderskikker dat bepaal. / Van Rooy, Bertus; Kruger, Haidee.

In: Tydskrif vir Geesteswetenskappe, Vol. 56, No. 1, 01.03.2016, p. 102-116.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Faktore wat die weglating van die Afrikaanse onderskikker dat bepaal

AU - Van Rooy, Bertus

AU - Kruger, Haidee

PY - 2016/3/1

Y1 - 2016/3/1

N2 - Verb complement clauses in written Afrikaans have two formal variants. One form has an overt complementiser dat "that", followed by dependent word order (dat+SXV), while the other form has no complementiser and independent word order (Ø+SVX). Previous research on Afrikaans has not yielded conclusive findings about the factors that influence the choice between the variants, although the factors conditioning the alternation have been studied extensively in English and German. Based on that research, a number of potential conditioning factors are identified, which relate to the syntactic complexity of the main clause (its subject, modification of tense and modality features of the main verb, passivisation, negation or additional modifiers between the main clause and the complement clause); the semantics, lexical choice and frequency of the verb of the main clause; and the formality of the register. This article adopts corpus linguistics as method to determine which factors best predict the choice between the variants. The data are drawn from the Taalkommissiekorpus ("Corpus of the Language Commission"), which is a 57 million word corpus of contemporary written Afrikaans. A sample of 10 084 instances of the declarative complement clause were extracted from the corpus by using 104 different verb lemmas (see Appendix A) as extraction terms. These were classified manually for their complementation pattern, and all the potential features identified as potential conditioning factors in the literature review were annotated for every valid instance in the sample. The classified data were then subjected to decision tree modelling, to yield a classification tree that identifies the most important factors related to the choice between the two variants of the declarative complement construction in Afrikaans. The results indicate that the verb of the main clause is the most important factor. A few very general verbs with high frequency, such as sê "say", weet "know" and dink "think" are strongly associated with the variant Ø+SVX. Verbs that are more specific in their meaning and less frequent tend to take the variant dat+SXV. In addition, register is an important conditioning factor for verbs that allow combination with both variants. More formal registers, such as academic writing, study guides and published popular books are more likely to make use of the variant dat+SXV, while the less formal registers of newspapers, magazines and fiction make more frequent use of the variant without a complementiser (Ø+SVX). The findings are interpreted as support for the view that thematic prominence is the underlying force behind the variability. When the main clause is very general and low in informativity, the most prominent information is contained in the complement clause, which is therefore presented without an overt complementiser and with main clause word order (Ø+SVX). However, when the main clause is thematically prominent, carries a higher information load, and demands the attention of the reader, the complement clause is overtly marked as subordinate by means of an overt complementiser and the use of dependent word order (dat+SXV).

AB - Verb complement clauses in written Afrikaans have two formal variants. One form has an overt complementiser dat "that", followed by dependent word order (dat+SXV), while the other form has no complementiser and independent word order (Ø+SVX). Previous research on Afrikaans has not yielded conclusive findings about the factors that influence the choice between the variants, although the factors conditioning the alternation have been studied extensively in English and German. Based on that research, a number of potential conditioning factors are identified, which relate to the syntactic complexity of the main clause (its subject, modification of tense and modality features of the main verb, passivisation, negation or additional modifiers between the main clause and the complement clause); the semantics, lexical choice and frequency of the verb of the main clause; and the formality of the register. This article adopts corpus linguistics as method to determine which factors best predict the choice between the variants. The data are drawn from the Taalkommissiekorpus ("Corpus of the Language Commission"), which is a 57 million word corpus of contemporary written Afrikaans. A sample of 10 084 instances of the declarative complement clause were extracted from the corpus by using 104 different verb lemmas (see Appendix A) as extraction terms. These were classified manually for their complementation pattern, and all the potential features identified as potential conditioning factors in the literature review were annotated for every valid instance in the sample. The classified data were then subjected to decision tree modelling, to yield a classification tree that identifies the most important factors related to the choice between the two variants of the declarative complement construction in Afrikaans. The results indicate that the verb of the main clause is the most important factor. A few very general verbs with high frequency, such as sê "say", weet "know" and dink "think" are strongly associated with the variant Ø+SVX. Verbs that are more specific in their meaning and less frequent tend to take the variant dat+SXV. In addition, register is an important conditioning factor for verbs that allow combination with both variants. More formal registers, such as academic writing, study guides and published popular books are more likely to make use of the variant dat+SXV, while the less formal registers of newspapers, magazines and fiction make more frequent use of the variant without a complementiser (Ø+SVX). The findings are interpreted as support for the view that thematic prominence is the underlying force behind the variability. When the main clause is very general and low in informativity, the most prominent information is contained in the complement clause, which is therefore presented without an overt complementiser and with main clause word order (Ø+SVX). However, when the main clause is thematically prominent, carries a higher information load, and demands the attention of the reader, the complement clause is overtly marked as subordinate by means of an overt complementiser and the use of dependent word order (dat+SXV).

KW - Afrikaans

KW - syntax

KW - word order

KW - variation

KW - complement clause

KW - subordinator

KW - complementiser

KW - dependent word order

KW - register

KW - verb lemma

KW - informativity

KW - corpus linguistics

UR - http://www.scopus.com/inward/record.url?scp=84963811321&partnerID=8YFLogxK

U2 - 10.17159/2224-7912/2016/v56n1a7

DO - 10.17159/2224-7912/2016/v56n1a7

M3 - Article

VL - 56

SP - 102

EP - 116

JO - Tydskrif vir Geesteswetenskappe

T2 - Tydskrif vir Geesteswetenskappe

JF - Tydskrif vir Geesteswetenskappe

SN - 0041-4751

IS - 1

ER -