The role of accentedness in acceptability judgements in L3 Norwegian: an across-domain investigation

The role of accentedness in acceptability judgements in L3 Norwegian: an across-domain investigation The role of accentedness in acceptability judgements in L3 Norwegian: an across-domain investigation

Foreign-accented speech has been explored from various perspectives, yet the current study is one of the first to investigate the role of accentedness in relation to acceptability judgements in multilingual acquisition. We consider possible interactions of two domains, phonology and syntax, by presenting learners of L3 Norwegian with an Acceptability Judgement Task focusing on eight morphosyntactic properties. Stimuli were recorded with Polish, English and native Norwegian accents. The rationale of the study is based on activation competition; it is expected that a Polish or English accent may enhance activation of syntax of the relevant language and thus affect acceptability judgements. Test items were selected based on their grammatical similarities or dissimilarities between Norwegian and the two previously acquired languages, Polish and English. Prerecorded sentences were evaluated for grammaticality on a seven-point Likert scale. Participants included L1 Norwegian controls and L1 Polish-L2 English-L3/Ln learners of Norwegian divided into instructed and immersive groups. Results show significant effects of accent, group, and condition/grammaticality, as well as a significant group by condition/grammaticality interaction. Conditions not similar to Polish and English did not exhibit accent effects. Findings suggest that differences in ratings based on accentedness are linked to structural similarity and exposure conditions.

Introduction

Foreign-accented speech, due to its significant social and pedagogical implications, has been widely researched from various angles, featuring sociolinguistic, experimental, and theoretical perspectives. The concept of 'foreign accent' commonly refers to segmental and prosodic deviations from the native pronunciation norms exhibited by non-native language users. Generally,

it is assessed holistically by raters on the basis of such rating parameters as the degree of foreign accent, speech intelligibility and/or acceptability.

The ratings of perceived global foreign accent have been widely applied in second language acquisition research; yet accentedness is a complex construct and it is very difficult to assess it unidimensionally on a simple scale or to assign it to specific categories. The main results of foreign accentedness ratings tend to indicate that some degree of foreign accentedness does not necessarily preclude intelligibility. It is generally acknowledged that the degree of foreign accentedness may differ as a function of the characteristics of the subjects examined, including the L1 background, the L2 target and the amount of language experience.

Another variable which has been found to affect the level of perceived accentedness is the grammaticality of the speech samples; however, phonological and syntactic errors which frequently coincide in foreign language learners are not easy to disentangle. Recent studies suggest that native speakers exhibit reduced sensitivity to grammatical violations in foreign-accented speech in comparison to native speech. These conclusions based on foreign accentedness ratings studies have been further confirmed by electrophysiological investigations, which indicate that morphosyntactic violations tend to be ignored during real-time non-native speech processing.

Research findings point to the interplay of several factors affecting the perception and processing of foreign accents, yet many questions remain unanswered. To the best of our knowledge, the role of accentedness has not been investigated in relation to syntactic and semantic acceptability judgements in multilingual acquisition; thus this contribution aims to fill this research void.

Previous research has shown that speakers' accents can provide information about speaker type and identity, and that perceptions about different speakers may also modulate speech processing. The impact of speaker accent on syntactic priming seems particularly relevant in today's global society, since foreign language learners may be exposed to various speakers with different language backgrounds and accents. Phonetic cues can influence how we process language, perhaps especially at the syntactic level. Prior work on phonetic cue-based syntactic activation has been rather limited, with a notable exception of Chun and Kaan, who investigated to what extent L2 speakers' syntactic priming is influenced by different accents and familiarity with those accents. The study involved Korean learners of L2 English, whose task was to describe pictures of ditransitive events after having listened to sentences recorded by three speakers of English with different accents: a native American accent and two non-standard accents (Korean, English and Indian English). The results demonstrated that the L2 participants' structural priming for each construction changed over time and was modulated by the speakers' accents and the participants' familiarity with those accents. However, there is a scarcity of research into the effects of accentedness on L2 syntactic processing and to our knowledge, there is no such study from a multilingual perspective; thus, the present study contributes to the understanding of this phenomenon.

The current models of L3 acquisition, and consequently also most studies that have been carried out so far, have focused on the linguistic source of Cross-Linguistic Influence, i.e. whether it is from only one of the previously acquired languages,

and if so, which one, or whether it could be from both. While the L2 Status Factor and the Typological Primacy Model assume that Cross-Linguistic Influence is mainly or solely from one of the previously acquired languages at early or 'initial stages' of development, the Cumulative Enhancement Model, the Linguistic Proximity Model, the Scalpel Model and the Natural Growth Theory of Acquisition argue that both languages are active and in competition for Cross-Linguistic Influence. The decisive factors determining which language is selected or most active for Cross-Linguistic Influence are considered to be order and manner of acquisition (the L2 Status Factor) or lexical and structural similarity between the L3 and one of the previously acquired languages (the Typological Primacy Model, Cumulative Enhancement Model, Linguistic Proximity Model and Scalpel Model). Furthermore, most studies also focus on one domain, typically morphosyntax or phonology, and studies that consider several domains are the exception.

In the present study, we consider the possible interaction of two domains, phonology and syntax, by presenting learners of L3 Norwegian who have Polish and English as their previously acquired languages with an Acceptability Judgement Task with stimuli where a property of L3 Norwegian is either similar to Polish and different from English or vice versa. The stimuli are spoken with a Polish, an English, or a native Norwegian accent. The rationale of the study is based on the idea of activation competition as it is argued for in the LPM and the Scalpel Model, and thus, that a Polish or English accent may enhance the activation of the syntax of the relevant language and consequently affect the acceptability judgment given.

Specifically, we ask whether speakers will provide different judgements for the same morphosyntactic phenomenon depending on accent, so that a grammatical Norwegian structure that is similar to Polish (and different from English) may be more acceptable with a Polish accent and less acceptable with an English accent, assuming that the native accent is the baseline. Likewise, a grammatical Norwegian structure that is ungrammatical in English may be less acceptable with an English accent than with a Polish accent. The reason for this is that English morphosyntax may be more strongly activated with the English accent and correspondingly, Polish more activated with the Polish accent. Thus, ungrammatical Norwegian structures which are also ungrammatical in Polish should be even less acceptable with a Polish accent than an English accent, and vice versa for ungrammatical structures that are similar to English.

Table one presents the general predictions based on structural similarity between Norwegian and English or Polish, combined with accent-based expectations. For structures that are ungrammatical in Norwegian, the predictions hinge on whether the same structure is also ungrammatical in one of the other languages (Similarity). If an ungrammatical