This paper looks at the perception of prosodic prominence and the interpretation of focus position in the unrelated languages, Samoan and English. In many languages, prosodic prominence is a key marker of focus, so it is expected that prosodic prominence would affect judgments of focus position. However, it is shown that focus position, in turn, influences the perception of prosodic prominence according to language-specific expectations about the alignment between focus position and nuclear accent placement. Two sets of parallel perception experiments in Samoan and English are reported. In the first, participants judged the most prosodically prominent word in sentences which varied in syntactic construction (cleft/canonical) and intended stress position (subject/object). In both languages, participants were more likely to choose the intended stressed word if it was in the focus position. However, this effect was much larger in Samoan, which fits with its relatively lower functionality of prosodic prominence. In the second experiment, participants were asked to choose which question had been asked, consistent with subject or object focus. It was found that in both languages, participants weighted syntactic and prosodic cues to focus in line with expectations from their language. These findings have implications for how we conceive the role of prosodic prominence in speech processing across languages.

<p>Uptalk, or the ‘innovative use of rising intonation on declarative utterances’ (Warren, 2016, p. xiii), is a common intonational contour in New Zealand English that has been assigned an uncommonly large range of meanings. Negative interpretations of uptalk include uncertainty, tentativeness, and lack of expert knowledge (Barr, 2003; Conley et al, 1978; Guy & Vonwiller, 1984; Spindler, 2003; Tomlinson Jr. & Fox Tree, 2010; Warren, 2016), and positive interpretations of uptalk include helpfulness, solidarity, and inclusivity (Borgen, 2000; Britain, 1992; Guy & Vonwiller, 1984; Meyerhoff, 1991; Warren, 2016). There is evidence that a listener’s interpretation of uptalk as positive or negative may depend on their age (Di Gioacchino & Crook Jessop, 2010), and there is also conflicting evidence over whether uptalk users are perceived as more or less suitable for highly-skilled employment (Borgen, 2000; Gorelik, 2016; Guy & Vonwiller, 1984; Steele, 1995). In an earlier study (Wollum, 2016), I found that older female listeners (aged 60-70) were significantly more likely than younger female listeners (aged 18-28) to assess a speaker as less competent and less trustworthy when the speaker was using uptalk, in an IT support context. In a further exploration of the age-based perception differences revealed in my 2016 research, this thesis reports a more extensive study, in which a new group of younger (aged 18-30) and older (aged 60-72) female listeners as well as a group of younger and older male listeners assessed recordings from four different young (aged 22-30) female speakers representing four different highly-skilled professions: IT support person, doctor, lawyer, and librarian. Listeners in this study were asked to assess not only the competence and trustworthiness of the speaker, but also the speaker’s education level. For both competence and trustworthiness, significant interactions were found between the listener’s age group, the profession of the speaker, and the speaker’s use or non-use of uptalk, with the older listeners ranking all four professions significantly lower for competence and trustworthiness in the uptalk condition, and a particularly strong effect of these lower ratings for the speaker representing a lawyer. There was also a recurring significant interaction between listener age group and gender, with younger male participants providing significantly higher ratings than all other demographics for both competence and trustworthiness, and significantly higher ratings than all but the older female group for education level. For education level, there was also a significant simple effect of speaker profession. There was no effect of uptalk on perceptions of education level. As all speakers used in the study were young (aged 22-30) females, part of the demographic that most often uses uptalk in New Zealand (Britain, 1992; Warren & Britain, 2000), the trend of lower competence and trustworthiness ratings from the older listener group is indicative of an out-group effect regarding positive versus negative perceptions of uptalk (House, 2006). In addition, the significant interactions between speaker profession and presence or absence of uptalk suggest that uptalk, considered predominantly a marker of female speech in New Zealand, is perceived more negatively by older listener groups in professions that have been historically male-typed. For competence and trustworthiness, the profession least negatively affected by the use of uptalk was the librarian, a profession that has previously been viewed as predominantly female (Morrissey & Case, 1988; Panek, Rush, & Greenawalt, 1977), and the profession most negatively affected by the use of uptalk was the lawyer, a profession that has previously been described as ‘aggressively male’ (Bolton & Muzio, 2007, p.56), and rewarding of women who adopt more masculine characteristics (Sommerlad & Sanderson, 1998). This research shows that older listeners are less likely than younger listeners to accept uptalk as indicative of competence and trustworthiness, and that these effects are particularly strong for professions in which women have previously been underrepresented. It also shows that a speaker’s perceived profession is more important than the presence or absence of uptalk for a listener assessing the speaker’s education level.</p>

<p>Uptalk, or the ‘innovative use of rising intonation on declarative utterances’ (Warren, 2016, p. xiii), is a common intonational contour in New Zealand English that has been assigned an uncommonly large range of meanings. Negative interpretations of uptalk include uncertainty, tentativeness, and lack of expert knowledge (Barr, 2003; Conley et al, 1978; Guy & Vonwiller, 1984; Spindler, 2003; Tomlinson Jr. & Fox Tree, 2010; Warren, 2016), and positive interpretations of uptalk include helpfulness, solidarity, and inclusivity (Borgen, 2000; Britain, 1992; Guy & Vonwiller, 1984; Meyerhoff, 1991; Warren, 2016). There is evidence that a listener’s interpretation of uptalk as positive or negative may depend on their age (Di Gioacchino & Crook Jessop, 2010), and there is also conflicting evidence over whether uptalk users are perceived as more or less suitable for highly-skilled employment (Borgen, 2000; Gorelik, 2016; Guy & Vonwiller, 1984; Steele, 1995). In an earlier study (Wollum, 2016), I found that older female listeners (aged 60-70) were significantly more likely than younger female listeners (aged 18-28) to assess a speaker as less competent and less trustworthy when the speaker was using uptalk, in an IT support context. In a further exploration of the age-based perception differences revealed in my 2016 research, this thesis reports a more extensive study, in which a new group of younger (aged 18-30) and older (aged 60-72) female listeners as well as a group of younger and older male listeners assessed recordings from four different young (aged 22-30) female speakers representing four different highly-skilled professions: IT support person, doctor, lawyer, and librarian. Listeners in this study were asked to assess not only the competence and trustworthiness of the speaker, but also the speaker’s education level. For both competence and trustworthiness, significant interactions were found between the listener’s age group, the profession of the speaker, and the speaker’s use or non-use of uptalk, with the older listeners ranking all four professions significantly lower for competence and trustworthiness in the uptalk condition, and a particularly strong effect of these lower ratings for the speaker representing a lawyer. There was also a recurring significant interaction between listener age group and gender, with younger male participants providing significantly higher ratings than all other demographics for both competence and trustworthiness, and significantly higher ratings than all but the older female group for education level. For education level, there was also a significant simple effect of speaker profession. There was no effect of uptalk on perceptions of education level. As all speakers used in the study were young (aged 22-30) females, part of the demographic that most often uses uptalk in New Zealand (Britain, 1992; Warren & Britain, 2000), the trend of lower competence and trustworthiness ratings from the older listener group is indicative of an out-group effect regarding positive versus negative perceptions of uptalk (House, 2006). In addition, the significant interactions between speaker profession and presence or absence of uptalk suggest that uptalk, considered predominantly a marker of female speech in New Zealand, is perceived more negatively by older listener groups in professions that have been historically male-typed. For competence and trustworthiness, the profession least negatively affected by the use of uptalk was the librarian, a profession that has previously been viewed as predominantly female (Morrissey & Case, 1988; Panek, Rush, & Greenawalt, 1977), and the profession most negatively affected by the use of uptalk was the lawyer, a profession that has previously been described as ‘aggressively male’ (Bolton & Muzio, 2007, p.56), and rewarding of women who adopt more masculine characteristics (Sommerlad & Sanderson, 1998). This research shows that older listeners are less likely than younger listeners to accept uptalk as indicative of competence and trustworthiness, and that these effects are particularly strong for professions in which women have previously been underrepresented. It also shows that a speaker’s perceived profession is more important than the presence or absence of uptalk for a listener assessing the speaker’s education level.</p>