7 - Melodic Complexity and Lexical Stress in Singapore English: An Experimental Study

Summary

0 Introduction

Emerging Englishes, New Englishes or outer-circle Englishes such as Singapore English (SgE) may sound markedly different from native or inner-circle varieties. One of the main perceptive features that distinguishes SgE from inner-circle Englishes is claimed to be its staccato, machine-gun-like rhythm due to its distinctive rhythmic properties, namely syllable timing instead of stress timing. Stressed syllables may not be substantially longer nor pronounced with higher pitch; moreover, vowel reduction in unstressed syllables may not be systematic either (Levis 2005; Bao 1998). The absence of quantity raises an interesting question: while inner-circle Englishes all organise their stress algorithm partly around quantity sensitivity, at least for the non-derived part of the lexicon, how do SgE speakers assign lexical stress (accent) if their system lacks quantity?

This chapter looks at lexical stress patterns in standard SgE. The evidence comes from two types of complementary corpora: (i) recordings following the PAC protocol (informal and formal conversations, text reading), and (ii) findings from a nonce experiment testing SgE native speakers’ intuition about the stress of disyllables, which we carried out in 2016. Our nonce experiment follows the protocol of Turcsan and Herment (2015), inspired by Krämer (2009) and Bárkányi (2002). The experiment involves reading tasks with embedded nonce words displaying different phonological and morphological structures forced by the spelling. The exact duplication of the protocol allows us to compare SgE speakers with inner-circle speakers in their stress placement and to shed light to SgE speakers’ internalised system with respect to lexical prominence.

In section 1 we briefly discuss some necessary background elements for evaluating the results of the nonce-word test. In section 1.1, we present our object of study by locating the variety of SgE we have access to through our PAC recordings. Section 1.2 is concerned with lexical prominence: we compare findings in the literature with those from our PAC recordings. Section 2 gives the rationale for our nonce-word test by explaining our methodology. Section 3 describes the results from various points of view: first with respect to other statistical approaches based on the lexicon in 3.1, then by looking at speakers’ preferences in 3.2. The link to quantity sensitivity is established in 3.3, constraints on unstressed syllables are explained in 3.4 and finally, possible analogical patterns are discussed in 3.5. Section 4 contains our conclusions.