LINGUIST List 32.2868

Thu Sep 09 2021

Diss: Cognitive Science; Historical Linguistics; Morphology; Syntax; Text/Corpus Linguistics: Anke Lensch: ''Onlookers, passers-by, sweeper-uppers, opt-outers and dumber-downerers. A corpus-based study of English -er nominalizations''

Editor for this issue: Sarah Robinson <>

Date: 31-Aug-2021
From: Anke Lensch <>
Subject: Onlookers, passers-by, sweeper-uppers, opt-outers and dumber-downerers. A corpus-based study of English -er nominalizations
E-mail this message to a friend

Institution: Johannes Gutenberg-Universität Mainz
Program: Department for English and Linguistics
Dissertation Status: Completed
Degree Date: 2020

Author: Anke Lensch

Dissertation Title: Onlookers, passers-by, sweeper-uppers, opt-outers and dumber-downerers. A corpus-based study of English -er nominalizations.

Linguistic Field(s): Cognitive Science
                            Historical Linguistics
                            Text/Corpus Linguistics

Dissertation Director:
Prof. Britta Mondorf
Prof. Graeme Trousdale
Prof. Damaris Nübling
Prof. Susanne Wagner
Prof. Jörg Meibauer

Dissertation Abstract:

English -er nominalizations of particle verbs (VPARTs) and prepositional verbs (VPREPs) are characterized by a high degree of formal and functional variation. This study is able to capture the extent of this variation by analysing more than 24.187 nominalizations that were extracted from diachronic and synchronic British and American English corpora, which together amount to roughly 3 billion tokens. Thus, the present analysis constitutes the most extensive quantitative and qualitative empirical analysis of English VPART/PREP -er nominalizations to date.

Formally, VPART/PREP -er nominalizations vary in the succession of the PART/PREP and the verb (V), the placement of the -er suffix and the number of times the -er suffix attaches to the complex base. It is possible to
distinguish five different patterns:
 pattern 1: onlooker PART/PREP Ver
 pattern 2: passer-by Ver PART/PREP
 pattern 3: sweeper-upper Ver PART/PREPer
 pattern 4: opt-outer V PARTer
 pattern 5: dumber-downerer Ver PARTerer

Each instantiation of pattern 2 breaches the Right-Hand Head Rule (cf. Williams 1981). Yet, in the Present-day English data set, pattern 2 is the most frequent and most productive of the five patterns. In addition,
instantiations of pattern 3 and 5 violate the Horror Aequi Principle (cf. Rohdenburg 2003) and thus they appear to be particularly unruly. However, upon close examination, the derivatives in this study reveal that multiple
marking with -er is quite systematic. What is more, derivatives featuring multiple -er suffixes are too frequently attested to simply dismiss them as mere outliers. In applying the framework of Construction Grammar, this study for the first time proposes that optional two-fold and multiple attachment of the -er suffix is based on an argument structure generalization.

By testing how far differences in form entail differences in meaning, moreover this study ascertains how far the five nominalization patterns can be regarded variant forms of the same construction. Both onlooker
(pattern 1) and looker-on (pattern 2) denote a ‘spectator.’ Consequently, the two derivatives are two variant forms of the same construction, which violates Goldberg’s Synonymy Principle (cf. Goldberg 1995: 67). In
contrast, comparing overtaker ‘someone outpacing someone else’ (pattern 1) and taker-over ‘the seizing of control over e.g. a company’ (pattern 2) leads to the conclusion that the two cannot be considered variants of
one another, as the two forms do not share equivalent descriptive meaning. The step-by-step analysis of more than 24.187 VPART/PREP -er nominalizations leads to the conclusion that most of them inherit their respective base’s descriptive meaning as well as the properties of the -er suffix construction. Hence, the semantic multifariousness
of VPART/PREPs can mostly be attributed to the notion of default inheritance. While this study gauges statistical tendencies among the highly heterogenous group of VPART/PREPs, importantly, the data analysis moreover unveils many exceptions to general statistical trends.

In one way or another, all VPART/PREP -er derivatives under scrutiny in this study are connected, either by at least one aspect of their form or function. In establishing the factors that govern the variation and by pointing out ambiguities and exceptions, this study argues that VPART/PREP -er nominalizations form a tightly knit and multi-facetted constructional cluster of several closely related constructions.

Page Updated: 09-Sep-2021