Comparing explanatory principles of complement selection statistically: a case study based on Canadian English

Juho Ruohonen, Juhani Rudanko

Research output: Contribution to journalArticleScientificpeer-review

Abstract

Several factors have been identified in the recent literature to explain variation in the selection of sentential complements in recent English, and the article begins with a survey of such factors. The article then offers a case study of the impact of such factors on non-finite complements of the adjective afraid on the basis of the Strathy Corpus of Canadian English. Attention is paid for instance to the Extraction and Choice Principles, passive lower predicates, and text type. Multivariate analysis is applied to compare and to shed light on such different explanatory principles. The Choice Principle proves to be by far the most significant predictor of the alternation, while the heavily correlated syntactic feature of Voice appears non-significant. Fiction, as opposed to the informative registers, shows a notable preference for to infinitives, though this finding needs to be replicated in datasets where controlling for author idiolect is possible. Theoretically plausible odds ratios are observed on the Extraction Principle and negation of the predicate, but they are not statistically significant. In the former case, this may well be due to the variable’s collinearity with the Choice Principle and its low overall frequency, resulting in a low effective sample size.
Original languageEnglish
JournalStudia Neophilologica
Number of pages18
ISSN0039-3274
DOIs
Publication statusE-pub ahead of print - 23 Jun 2019
MoE publication typeA1 Journal article-refereed

Fields of Science

  • Complementation
  • syntax
  • corpus linguistics
  • multivariate analysis
  • non-finite clause
  • GRAMMAR
  • 6121 Languages

Cite this

@article{8d644fefe4044efc9b1309d4f35e0b39,
title = "Comparing explanatory principles of complement selection statistically: a case study based on Canadian English",
abstract = "Several factors have been identified in the recent literature to explain variation in the selection of sentential complements in recent English, and the article begins with a survey of such factors. The article then offers a case study of the impact of such factors on non-finite complements of the adjective afraid on the basis of the Strathy Corpus of Canadian English. Attention is paid for instance to the Extraction and Choice Principles, passive lower predicates, and text type. Multivariate analysis is applied to compare and to shed light on such different explanatory principles. The Choice Principle proves to be by far the most significant predictor of the alternation, while the heavily correlated syntactic feature of Voice appears non-significant. Fiction, as opposed to the informative registers, shows a notable preference for to infinitives, though this finding needs to be replicated in datasets where controlling for author idiolect is possible. Theoretically plausible odds ratios are observed on the Extraction Principle and negation of the predicate, but they are not statistically significant. In the former case, this may well be due to the variable’s collinearity with the Choice Principle and its low overall frequency, resulting in a low effective sample size.",
keywords = "Complementation, syntax, corpus linguistics, multivariate analysis, non-finite clause, GRAMMAR, 6121 Languages",
author = "Juho Ruohonen and Juhani Rudanko",
year = "2019",
month = "6",
day = "23",
doi = "10.1080/00393274.2019.1616215",
language = "English",
journal = "Studia Neophilologica",
issn = "0039-3274",
publisher = "Taylor & Francis",

}

Comparing explanatory principles of complement selection statistically : a case study based on Canadian English. / Ruohonen, Juho; Rudanko, Juhani.

In: Studia Neophilologica, 23.06.2019.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - Comparing explanatory principles of complement selection statistically

T2 - a case study based on Canadian English

AU - Ruohonen, Juho

AU - Rudanko, Juhani

PY - 2019/6/23

Y1 - 2019/6/23

N2 - Several factors have been identified in the recent literature to explain variation in the selection of sentential complements in recent English, and the article begins with a survey of such factors. The article then offers a case study of the impact of such factors on non-finite complements of the adjective afraid on the basis of the Strathy Corpus of Canadian English. Attention is paid for instance to the Extraction and Choice Principles, passive lower predicates, and text type. Multivariate analysis is applied to compare and to shed light on such different explanatory principles. The Choice Principle proves to be by far the most significant predictor of the alternation, while the heavily correlated syntactic feature of Voice appears non-significant. Fiction, as opposed to the informative registers, shows a notable preference for to infinitives, though this finding needs to be replicated in datasets where controlling for author idiolect is possible. Theoretically plausible odds ratios are observed on the Extraction Principle and negation of the predicate, but they are not statistically significant. In the former case, this may well be due to the variable’s collinearity with the Choice Principle and its low overall frequency, resulting in a low effective sample size.

AB - Several factors have been identified in the recent literature to explain variation in the selection of sentential complements in recent English, and the article begins with a survey of such factors. The article then offers a case study of the impact of such factors on non-finite complements of the adjective afraid on the basis of the Strathy Corpus of Canadian English. Attention is paid for instance to the Extraction and Choice Principles, passive lower predicates, and text type. Multivariate analysis is applied to compare and to shed light on such different explanatory principles. The Choice Principle proves to be by far the most significant predictor of the alternation, while the heavily correlated syntactic feature of Voice appears non-significant. Fiction, as opposed to the informative registers, shows a notable preference for to infinitives, though this finding needs to be replicated in datasets where controlling for author idiolect is possible. Theoretically plausible odds ratios are observed on the Extraction Principle and negation of the predicate, but they are not statistically significant. In the former case, this may well be due to the variable’s collinearity with the Choice Principle and its low overall frequency, resulting in a low effective sample size.

KW - Complementation

KW - syntax

KW - corpus linguistics

KW - multivariate analysis

KW - non-finite clause

KW - GRAMMAR

KW - 6121 Languages

U2 - 10.1080/00393274.2019.1616215

DO - 10.1080/00393274.2019.1616215

M3 - Article

JO - Studia Neophilologica

JF - Studia Neophilologica

SN - 0039-3274

ER -