Usage-Based Approaches to Language Acquisition and Language Teaching

Jacqueline Evers-Vermeul, Elena Tribushinina, Jacqueline Evers-Vermeul, Elena Tribushinina

• Publisher: De Gruyter Mouton
• Year: 2017
• Print ISBN: 9781501517525
• eBook ISBN: 9781501505423
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Teaching Language Arts
• Index: Languages & Linguistics

Although usage-based approaches have been successfully applied to the study of both first and second language acquisition, to monolingual and bilingual development, and to naturalistic and instructed settings, it is not common to consider these different kinds of acquisition in tandem. The present volume takes an integrative approach and shows that usage-based theories provide a much needed unified framework for the study of first, second and foreign language acquisition, in monolingual and bilingual contexts. The contributions target the acquisition of a wide range of linguistic phenomena and critically assess the applicability and explanatory power of the usage-based paradigm. The book also systematically examines a range of cognitive and linguistic factors involved in the process of language development and relates relevant findings to language teaching. Finally, this volume contributes to the assessment and refinement of empirical methods currently employed in usage-based acquisition research. This book is of interest to scholars of language acquisition, language pedagogy, developmental psychology, as well as Cognitive Linguistics and Construction Grammar.

IIDriving forces of language development

Rasmus Steinkrauss

5L1 acquisition beyond input frequency

Rasmus Steinkrauss, University of Groningen

1Introduction

In recent years, the frequency of linguistic forms in spoken language has increasingly been recognized as one of the main factors shaping language development. This applies to language development both in the language community and in the individual speaker (e.g. Bybee 2006; Croft 2000). For example, in the area of first language (L1) acquisition, the frequency of linguistic structures in a child’s ambient speech has repeatedly been shown to possess strong explanatory power for the linguistic development of that child (Lieven 2010).

Due to an emphasis on input frequency in some lines of L1 research, attention has shifted away from more qualitative factors in acquisition, although these are believed to influence L1 development as well (Lieven 2010; Lieven and Tomasello 2008). The present chapter therefore focuses explicitly on these factors and looks at L1 acquisition beyond input frequency. It will be shown that the explanatory power of input frequency is limited, and that previously learnt linguistic knowledge, communicative interest, and the conditions of use of linguistic structures may override the impact of input frequency, even in the early stages of L1 acquisition.

1.1The role of frequency

A major work highlighting the role of frequency in language development is Bybee’s (1995) book on morphology and linguistic change. In this book, she draws a distinction between the token frequency and the type frequency of linguistic items. Token frequency denotes the frequency of a specific linguistic item, while type frequency refers to the frequency of a linguistic pattern. For example, a speaker might use three different combinations of what and an auxiliary verb (e.g. what is, what was and what does) in a speech sample. The type frequency of the what+AUX pattern is in this case three. The token frequency on the other hand is the number of times each individual combination occurs in the sample – for example, it would be twelve for what does if that combination occurred twelve times.

The frequency of linguistic items and patterns influences their respective representation in the minds of speakers. Generally speaking, it is assumed that items that frequently occur together are eventually stored as one linguistic unit, while patterns that occur frequently are stored as a linguistic construction with open slots (Bybee 2006). Taking the above example, if one of the three what+AUX combinations would be extremely frequent in the speech, i.e. show a high token frequency, a speaker would store that two-word combination as a single unit, much in the way as if it was one word. If none of the combinations would be frequent and the what+AUX pattern instead exhibits a high type frequency, with a lot of different auxiliary verbs in the second place and only what staying constant, a speaker would come to represent this pattern as a what+AUX construction. That construction would incorporate a fixed item, what, and an open slot in which to insert auxiliary verbs. Importantly, these two types of representation are not mutually exclusive. A speaker might store a few frequent what+AUX combinations as fixed units, ready for use, and at the same time possess a schematic what+AUX construction with a slot, enabling her to assemble also less frequent what+AUX combinations.

Because of these different types of representation, the frequency of linguistic items and patterns in the speech has certain effects for both language change and language acquisition. For example, in processes of language change, because high-frequency items and structures are repeated so frequently and acquire a special representational status compared to less frequent structures, they are more likely to undergo phonetic reduction, resist structural change, and lose transparent semantic connections to similar constructions (Bybee 2006). In L1 acquisition, input frequency has a measurable effect on what children produce in their early phases of development. The following section discusses various findings in this respect.

1.2Wh-question acquisition

A well-researched area with respect to the influence of input frequency is the L1 acquisition of wh-questions. In English, wh-questions usually start with a wh-pronoun followed by a form of an auxiliary verb.13 Because the number of wh-pronouns and auxiliary verbs is restricted and there are only a few different inflected forms of auxiliary verbs, the number of possible WH+AUX combinations is limited. In addition, some of the combinations are very frequent. If input frequency is an important factor for acquisition, one may expect children who acquire English as L1 to quickly develop unit-like representations of the frequent WH+AUX combinations and subsequently produce these combinations from early on. This is exactly what has been found. In their wh-question development, English-learning children start out with fixed phrases such as what’s this or specific WH+AUX combinations such as where is... (Dabrowska 2000; Dabrowska and Lieven 2005; Klima and Bellugi 1966), and the combinations they produce first are those that tend to be highly frequent in their ambient speech (Rowland et al. 2003). The WH+AUX combinations they use seem to be fixed, because children do not freely combine any wh-pronoun with any form of auxiliary verb even if either of the words already occurs in their speech in other combinations. In addition, children’s questions starting with the frequent WH+AUX combinations show fewer errors than wh-questions starting with other combinations (see also Rowland 2007; Rowland and Pine 2000, 2003). Taken together, the findings suggest that children initially operate with specific WH+AUX combinations that they have extracted from their ambient speech, possibly aided by the fact that the WH+AUX combinations occur in a salient position at the beginning of wh-questions. Input frequency thus seems to have a direct impact on children’s processing and production of language.

After an initial phase of wh-question production based on specific linguistic items (for item-based learning, see, e.g. Lieven and Tomasello 2008; Tomasello 1992), English-learning children gradually abstract more schematic representations of wh-questions. For example, the fixed string What’s+Mommy+doing that the child observed by Dabrowska (2000) produced around the age of 2;0 had step by step developed into a more abstract What+is+NP+V construction by the age of 3;0.14 Such a development from specific to abstract constructions is assumed to be an effect of children’s accumulated linguistic experience. As children hear more and more wh-questions, the variation in the different parts of the construction, e.g. the wh-pronouns and the verb forms, grows. This variation is not random: wh-pronouns will always be in the first position, auxiliary verbs in second position, and so on. Children will gradually detect these patterns and develop open slots for the positions in the construction. In such a way, children get increasingly more creative over time and depart more and more from the specific question structures they have heard in the input (Dabrowska and Lieven 2005).

The finding that variation in the input fosters the development of abstract linguistic knowledge is supported by a case study of a boy learning German. As German shows more variation in the verb types and verb forms allowed after wh-pronouns, the type frequency of verbs after wh-pronouns is higher than in English. As a consequence, the boy in this case study developed an abstract WH+V construction quicker than typical English-learning children and relied on specific WH+V combinations only to a very limited extent (Steinkrauss 2009).

To sum up, the frequency of linguistic structures in the input has found to be a powerful explanation for children’s course of language acquisition, not only in the area of wh-questions, but also in other fields of L1 development (Cameron-Faulkner, Lieven, and Tomasello 2003; Theakston et al. 2001, 2002). The success of input frequency a...

Table of contents

• Cover
• Title Page
• Copyright
• Acknowledgments
• Table of contents
• List of contributors
• Language acquisition and language teaching in the usage-based framework
• I Theoretical and methodological foundations of language acquisition from a usage-based perspective
• II Driving forces of language development
• III Implications for language teaching and translation
• Index