Rob Goedemans

Acoustic Characteristics of Infant- directed Speech as a Function of Prosodic Typology

When interacting with infants, caregivers, regardless of sex, age, or social status, modify the prosodic and phonetic features of their speech by using infantdirected speech (IDS; e.g., Fernald 1993, Papoušek & Papoušek 1981, Snow 1977, Soderstrom 2007). This is a speech style that infants prefer to adultdirected speech (ADS; Cooper & Aslin 1990, Fernald 1985). During the past forty years, a rather extensive body of research has investigated the acoustic differences between IDS and ADS in various languages. Early results suggest that both suprasegmental and segmental acoustic cues differ between the two registers crosslinguistically (e.g. Bernstein Ratner & Luberoff 1984, Fernald et al. 1989, McMurray, KovackLesh, Goodwin & McEchron 2013, Soderstrom 2007). The afore mentioned and other studies have identified distinctive characteristics of IDS, such as slower tempo, higher pitch, wider pitch range, shorter utterances, and larger vowel spaces. Indeed, in a recent systematic review of this literature, 85 percent of studies reported slower tempo and/ or longer vowel duration; 92 percent found higher pitch (typically measured through fundamental frequency (F0); we will use pitch and F0 interchangeably); and 82 percent documented larger vowel spaces (Cristia 2013). However, these studies have not explored whether IDS and ADS differ to the same extent or along the same acoustic dimensions in typologically diverse languages; nor have they examined the related question of whether the contrasts along the relevant prosodic dimensions (stress, tone, and pitch accent) are maintained in IDS. These questions are important since they may inform us about why IDS in general displays exaggerated acoustic characteristics as compared to ADS in the first place. They may also explain the similarities and differences that we see in various languages. A further benefit of this research is that it provides a muchneeded insight into our understanding of the input to infants. Thus, it is worthwhile to foster this line of research by providing a review of IDS acoustic

The separation of accent and rhythm: evidence from StressTyp

This chapter offers a demonstration of various appl ications and uses of the StressTyp dataset (Goedemans and van der Hulst 2009). Drawing on previous studies, the first part of this chapter presents overviews of the major typ es of stress systems as these are represented in StressTyp, both in tabular form and plotted in maps. In the second part of this chapter we focus on the use of StressTyp in pr oviding support for a particular theoretical claim, namely the separation of (primar y) stress and rhythm. Many languages display stress patterns that involve a distinction between one primary stress and one or more non-primary stresses (or rhythmic beats). Appr oaches to the formal analysis of stress patterns differ in various ways, one being w hether primary stress and non-primary stress are derived in terms of a single algorithm o r tw separate algorithms. Van der Hulst (1996, 2009, 2010, 2012, this volume b), for exampl e, resents a formal theory of word stress that separates the representation of primary stress (called the accent) from the representation of syllables that are rhythmically s trong. The basic idea is that accent is calculated first, while rhythmic beats are accounte d for independently (at a later derivational stage), although ‘with reference’ to t he accent location. 1 This approach, which dates back to van der Hulst (1984), has been called ‘a (primary) accent first theory’ ([P]AF), e.g. in van der Hulst (1996, 2009). The th ory was developed as an alternative to standard metrical phonology (Liberman and Prince 19 77, Vergnaud and Halle 1978; Hayes 1980, 1995; Halle and Vergnaud 1987; Idsardi 1992) which develops a unified account of primary and rhythmic stress. In previous work, we have supplied several argumen ts that support ‘the separation theory’. In this article we review these arguments, this time with the specific goal of providing quantitative evidence for them with data from StressTyp, a typological database containing information about word accent a nd rhythm in 511 languages (Goedemans and van der Hulst 2009). Some of the Str essTyp query results reported here have been taken from Goedemans and van der Hulst (2 009) and Goedemans (2010a). 2

Word Prominence and Areal Linguistics

The goal of this chapter is to present an overview of the consequences of language contact, with the understanding that linguistic areas arise through contact (Hickey 2010), focusing on word prominence (i.e. stress and pitchaccent). Section 7.2 deals with preliminary issues which are relevant to the study of contact-induced change. In Section 7.3 we briefly present some cases of convergence involving changes in word prominence attributable to contact. Section 7.4 makes the point that language contact often leads to hybrid systems. Realizing that language contact can lead to both convergence (linguistic areas) and divergence, Section 7.5 focuses on divergence of an ancestral system, partly due to language contact, into a variety of closely related systems, taking Basque as a case study. Given the rather limited availability of systematic studies of contact-induced change in word prominence, in Sections 7.6 and 7.7 we present two detailed original case studies. Section 7.6 deals with the languages of North America where, drawing on Rice (2010, 2014), but based on a larger sample of languages, various instances of language contact are studied, focusing on areal distributions that cut across language families. Section 7.7, drawing on Goedemans (2010a), offers a case study concerning the aboriginal languages of Australia, in whichwe find variation between initial stress and penultimate stress in a geographically concise contact area. Section 7.8 presents some conclusions and directions for further research.

Iquito: The Prosodic Colon and Evaluation of OT Stress Accounts

der Hulst, Matt Gordon, Brett Hyde and Irene Vogel, as well as to my colleagues at the University of Leipzig, for their comments and input. I am particularly grateful to three anonymous reviewers for their careful reports and invaluable comments or questions, that I have for the most part tried to address. Finally, thanks are due to the editors of the present volume and especially to Rob Goedemans for his helpful editorial comments. The usual disclaimers apply. Iquito: the prosodic colon and evaluation of OT stress accounts* Nina Topintzi Universität Leipzig nina.topintzi@uni-leipzig.de

Covert Representations, Contrast, and the Acquisition of Lexical Accent

Lexical accent presents in a particularly sharp form some basic problems that learning models must overcome. Dresher & Kaye (1990) attempt to address these problems in the context of a learning model for a parametric metrical phonology for languages with predictable stress systems. I will show how this model, as modified by Dresher (1994), can be extended to learn stress systems that incorporate lexical accent, at least in the case of a constructed simple language inspired by Russian. Many of the problems in acquiring lexical accent have to do with hidden structure. Hidden structure has been discussed in the context of ‘overt’ and ’covert’ structures. I argue that these categories are fluid, and not fixed, as might sometimes appear from other learning models. Though I assume a derivational parametric metrical theory, the basic strategies have some affinity with ideas proposed by Tesar and his colleagues (Tesar et al. 2003; Tesar 2006; 2014) for learning an Optimality-Theoretic grammar. The notion of contrast plays an important role in the acquisition of lexical representations, as suggested by Tesar (2006; 2014). I will extend the approach of Dresher (2009), which considers contrast in the context of segmental features, to lexical accent. Recognizing the role of establishing contrasts in the lexical inventory will also help to streamline the system of cues proposed by Dresher & Kaye (1990). In particular, I show how the learning model in Dresher and Kaye (1990), as modified by Dresher (1994), can be revised and extended to learn the lexical accent system of a constructed simple language inspired by Russian. Importantly, this model compares multiple forms in order to establish contrasts in the lexicon. While the model is sufficiently specified to understand how it works on the simple constructed example, it remains to be implemented computationally. The plan of this chapter is as follows. In Section 8.2, I briefly review what the grammar of stress and metrical representations look like in the Simplified Bracketed Grid theory. Problems attending to the acquisition of metrical representations and the grammar of stress are reviewed in Section 8.3, where I present the learning model of Dresher & Kaye (1990). Section 8.4 takes up stress systems characterized by lexical accent. In Section 8.4.1 I briefly review the PAKA world model of Tesar and his associates, and in Section 8.4.2 I outline a learning model for lexical accent that builds upon the proposals of Dresher & Kaye (1990) and Dresher (1994). The role of contrast in the acquisition of segmental and metrical representations is discussed in Section 8.5. Section 8.6 is a brief conclusion.

10. Word accent systems in the languages of Asia

This chapter surveys accentual systems in the languages of Asia. Our objective has been to provide information on as many languages as we could lay our hands on, given the inevitable limitations on the time for this project and on access to sources in this period. This survey, then, does not claim anything near completeness. In a way, it presents an agenda for further studies especially witnessed by the many sections where little or no information on accentual systems is presented. We are not defending these limitations by claiming that the basic descriptive work has not been done in all these cases, although for many languages this is probably true. We have no doubt, however, that a lot of useful information on word accent has been gathered and is present in the countless language descriptions that we have been unable to consult. There may even be typological surveys that focus on word accent that we have overlooked. With all these limitations and shortcomings, we hope that this chapter still o¤ers a useful inventory which will stimulate further typological and theoretical research. In section 2, we outline the contents of this chapter, motivating its organization and introducing the conventions which guide the presentation of accentual data. Section 3 explains on which materials this survey is based. Sections 4, 5, 6 and 7 present data on word accent systems in

11. Word accent systems in the Middle East

This chapter deals with accentual systems in the languages spoken in the Middle East and includes a discussion of a number of isolate ancient languages and some Afro-Asiatic languages, notably Egyptian and Semitic languages. The notion ‘Middle East’ is not well-defined linguistically, and some languages that could have been included here are dealt with in other chapters. All Altaic languages, Indo-Iranian languages (such as Kurdish and Avestan) are discussed in Schiering and van der Hulst (this volume). All Caucasian languages as well as Indo-European languages such as Hittite (and related Anatolian languages) and Armenian are treated in van der Hulst (this volume). Afro-asiatic languages spoken in North Africa can be found in Downing (this volume). This survey, incomplete as it is, was included because the relevant area was not included in the areas covered in the other chapters of this volume. The organization of each section (or subsection) with accentual data is as follows: a. Genetic structure of the (sub)family; these are based on sources such as Ruhlen (1991), Comrie et al (2003), the Ethnologue (15th edition) and several others of the many sources that o¤er classifications. We have tried to strike a compromise in cases of conflicting groupings and no claim is made here that the resulting groupings are superior to those o¤ered in other sources. In each case (sub)family names are presented in capitals. We have not included information of the numbers of languages per (sub)family and in most cases list only a (sometimes rather arbitrary) subset of the languages in each (sub)family. Languages that are included in StressTyp (see section 3) are indicated in bold. Languages that are not in StressTyp about which this chapter provides information are underlined.

9. Word accent systems in the languages of Europe

This chapter deals with accentual systems in the languages of Europe. This work, however, does not claim completeness. Rather, it presents an agenda for further research especially witnessed by several sections where little or no information on accentual systems is presented. I am not defending these limitations by claiming that the basic descriptive work has not been done in all these cases, although for a number of languages this is most likely true. I have no doubt that a lot of useful information on word accent not reported here has been done and is present in the countless language descriptions that, for various reasons, I have been unable to consult. With these limitations, I hope that this chapter still o¤ers a useful survey which will stimulate further typological and theoretical work. In section 2 I outline the contents of this chapter, motivating its organization. Section 3 explains the materials on which this chapter is based. Sections 4 present the accentual data. Finally, in section 5 I will o¤er some general observations and conclusions.

5. Accent in the native languages of North America

The goal of the volume under review is to present in survey fashion what is known about the accentual properties of the world’s languages. For this purpose the editors have assembled a distinguished list of authors, including themselves, all deeply interested in accentual systems, several having broad areal expertise. The result is an invaluable handbook that phonologists, typologists, and arealists will want to often consult, if not own. The task of conceptualizing, let alone producing such a work is an enormous one, as researchers in this area face several obstacles which are appropriately evoked in various chapters. The first is to establish a reasonable amount of agreement on what kind of properties count as “accentual”. Assuming that this issue has been resolved (it hasn’t), a second obstacle concerns the available literature: Beyond the fact that many languages have not been described at all, descriptions which do exist often provide only scanty coverage of accentual phenomena—or even fail to mention whether a given language has stress-accent or not. Finally, where multiple descriptions exist, one often finds conflicting views on whether a given language has stress-accent or not, and if so, on how the stress system works. Unless otherwise noted, in what follows I will use the words accent and accentual as cover terms for the broad range of phenomena covered in this survey, which mostly concern stress-accent (SA), the more elusive concept of pitch-accent (PA), and occasionally other properties to be discussed below.