Romancing with tone: On the outcomes of prosodic contact

ROMANCING WITH TONE: ON THE OUTCOMES OF PROSODIC CONTACT Guri Bordal Steien Kofi Yakpo Inland Norway University of Applied Sciences The University of Hong Kong This article presents a descriptive and theoretical framework for the analysis of prosodic systems that have emerged from contact between African tone and European intonation-only languages. A comparative study of the prosodic systems of two Romance contact varieties, Central African French and Equatorial Guinean Spanish, shows that they feature two-tone systems, fixed word-tone patterns, tonal minimal pairs, the arbitrary assignment of tone in function words, and tonal processes. Evidence from further contact varieties and creole languages shows that similar systems evolved in other Afro-European contact ecologies. We conclude that tone is imposed by default on contact varieties and creoles that take shape in ecologies characterized by sourcelanguage agentivity in tone languages. In doing so, we argue against claims that tone necessarily cedes to stress during language contact and creolization. Instead, contact varieties and creoles partake just like other languages in the convergence processes that lead to the areal clustering of prosodic systems.* Keywords: Central African French, Equatorial Guinean Spanish, creole, prosodic contact, tone, stress, source-language agentivity 1. Introduction. There is a common assumption that tone is prone to disappearing when it meets stress during language contact. In an influential article in Language, McWhorter states that ‘the elimination of such features [i.e. tone, among others] is prototypical of creole genesis in all contexts’ (1998:793), and that ‘in creoles, [there is] little or no use of tone to lexically contrast monosyllables’ (1998:798). Salmons suggests in yet more general terms that ‘the basic pattern of accentual change appears to begin with a tonal language, shifts to a pitch-accent language and finally becomes a stressaccent language’ (1992:185). Trudgill holds that phonological simplification in language contact includes ‘loss of tone’ (2010:309). Submissions on the vulnerability of tone in contact, even in ecologies where tone languages dominate, are also found in numerous earlier sources (Heine 1978:220, and the sources cited there). The notion that tone cedes to stress has also found its way into scholarly work beyond linguistics. A key textbook in human geography claims, ‘when tonal and non-tonal languages meet and hybridize, tone is soon a casualty’ (de Blij 2009:45). The assumption that tone is the loser seems linked to the view that it is somehow more marked than stress, and therefore more likely to be shed due to contact-induced ‘simplification’ (cf. Salmons 1992:63–65). Our skepticism toward the idea that tone is a loser per se during prosodic contact provided the impetus for this study. We perceived the need for a systematic approach to 1 * We are grateful to the Center for Multilingualism in Society across the Lifespan (MultiLing) of the University of Oslo for funding reciprocal research stays at the University of Hong Kong and at MultiLing, and for providing a fertile intellectual environment for this study. Part of the write-up was enabled by support from a Doris Zimmern HKU-Cambridge Hughes Hall Fellowship to Kofi Yakpo and a travel grant from the Inland Norway University of Applied Sciences to Guri Bordal Steien. This article benefited greatly from detailed comments and constructive criticism by three anonymous referees and Language editors Andries Coetzee and David Willis. A special note of thanks goes to Gjert Kristoffersen (University of Bergen) for his lucid and thorough comments in a crucial phase of revisions and to Bård Uri Jensen (Inland Norway University of Applied Sciences) for help with the statistics. We also wish to thank Salikoko Mufwene for fruitful conversations that helped shape our ideas on prosodic contact, as well as Carlos Gussenhoven and Danae Perez for comments on an earlier version of this article. We are indebted to the speakers of Central African French and Equatorial Guinean Spanish in the Central African Republic and in Equatorial Guinea without whose help this study could not have been conducted. The names of speakers are provided in earlier work where disclosure was possible (Yakpo 2009, 2019a). Printed with the permission of Guri Bordal Steien & Kofi Yakpo. © 2020. Bordal Steien, Guri & Kofi Yakpo. 2020. Romancing with tone: On the outcomes of prosodic contact. Language 96(1). 1–41. https://doi.org/10.1353/lan.2020.0000. prosodic contact, founded on empirical evidence, typologically informed, and embedded in social context. Only on such grounds can more general claims be made regarding the fate of prosodic features in language contact. Our own work on prosodic systems that have emerged via contact between African and European languages is the starting point of this study and sets the confines of our observations for now. Contact between prosodic systems has, on the whole, received little attention (cf. Downing 2015). The broad assumptions about the behavior of tone during language contact cited above therefore also rest on somewhat scanty empirical foundations. This circumstance provided a further motivation to explore the fate of tone in language contact. The first aim of this study is to show that the outcomes of prosodic contact in AfroEuropean contact scenarios are determined by the composition of linguistic ecologies1 and the nature of the input prosodic systems, not by typological characteristics of tone as such. Tone can be transferred during language contact, just like other structural features, if the ecology is dominated by speakers of tone languages. Our argument is based on a comparative acoustic and phonological study of the prosodic systems of Central African French (CAF) and Equatorial Guinean Spanish (EGS), two colonial contact varieties spoken in Africa that have diverged from Standard European French and Standard European Spanish, respectively (see §3). We show that CAF and EGS both have two-tone systems with a high (H) and a low (L) tone, tonal minimal pairs, and fixed word-tone patterns.2They therefore possess key diagnostic features of tone systems. At the same time, H tone is obligatory and culminative, meaning that most words (but not all) have one and only one H tone. Hence, CAF and EGS also exhibit two structural features associated with stress systems. In other words, their prosodic systems display a combination of characteristics inherited from the tone and the stress inputs in the ecology. The double entendre of our title alludes to the rather unexpected presence of tone in Romance, a linguistic grouping otherwise known to exclusively exhibit stress. At the same time, the retention of features identified with stress systems makes the relationship of CAF and EGS with tone appear somewhat flirtatious and not fully committed. The second aim of this study is to develop a descriptive and theoretical framework that accounts for the outcomes of prosodic contact in different types of ecologies. Our framework is based on the case studies of CAF and EGS, but we furnish additional evidence from other Afro-European contact varieties, especially West African Englishes, as well as the English-lexifier creoles of Africa and the Caribbean. Although we focus on ecologies in which tone languages are spoken by the majority of the population, our framework can potentially be extended to the inverse situation, in which speakers of 2 LANGUAGE, VOLUME 96, NUMBER 1 (2020) 1Linguistic ecology is a metaphor that goes back to Haugen (1972) to refer to the interaction between a language and the society in which it is spoken. We use it here in the sense of Mufwene (2001:xii) to characterize the ‘socioeconomic and ethnographic environment in which a language has evolved’ (external ecology), as well as the ‘systemic interaction of the linguistic codes in contact’ (internal ecology) (Ansaldo 2009:4). 2For reference, the following abbreviations are used for language and place names throughout the article: CAF: Central African French, CAR: Central African Republic, EGS: Equatorial Guinean Spanish, SEF: Standard European French, SES: Standard European Spanish. The following tonal notations and abbreviations are used in glosses and elsewhere: – : morpheme boundary, = : clitic morpheme boundary; %: boundary tone, ó: high tone, ò: low tone, o: mid tone; acc: accusative case, adj: adjective, ads: adstrate, comp: complementizer, con: contact variety, cop: copula, dat: dative case, def: definite article, excl: exclusive, f: feminine gender, H: high tone, incl: inclusive, indf: indefinite article, indp: independent person form, ipfv: imperfective aspect, L: low tone, loc: locative, M: mid tone, m: masculine gender, nom: nominative case, pfv: perfective aspect, pl: plural number, poss: possessive case, pron: pronoun, prs: present tense, pst: past tense, refl: reflexive, rel: relative pronoun, sbj: subject case, sg: singular number, sub: subordinator (in glosses), sub: substrate (in text), sup: superstrate. stress languages dominate, and yet other contact constellations (see §2.2). We argue that three mechanisms are at work in the creation of contact prosodic systems in tone-dominated ecologies. The most important one is stress-to-tone mapping: speakers of tone languages create the foundations of a tone system by exploiting the perceptual analogies between the phonetic realizations of stress and tone (§4.1). Paradigmatization, the second mechanism, involves the systematic attribution of tone patterns to words and syntactic categories (§4.2). The third mechanism we term idiosyncratization: phonological reanalysis gives rise to tonal irregularities within paradigms and to idiosyncratic tonal processes in the contact varieties that often have no direct correspondence in the input languages (§4.3). At several junctures, we address the linguistic and social factors at play in the emergence of contact prosodic systems (§2.1, §3.1, and §5) and thereby underline the importance of linguistic ecologies in shaping the outcomes of prosodic contact. Section 6 concludes this article and includes some observations about the typological nature of Afro-European contact prosodic systems. 2. Prosody and language contact. We now address social and linguistic factors that play a role in determining the outcomes of prosodic contact (§2.1). We also provide definitions of the two prosodic types of tone and intonation-only systems, identify four contact constellations between these types, and limit the scope of this study to one of these constellations (§2.2). 2.1. Social and linguistic factors in language contact. We are concerned with how a structural category, namely prosody, is affected by language contact in a midto long-term perspective. To investigate this, we take a synchronic look at the prosodic systems of CAF and EGS. We also place the emergence of these systems in a diachronic context by developing hypotheses about general mechanisms that determine the outcomes of prosodic contact. These outcomes are mediated by the interaction of social and linguistic factors in a linguistic ecology (see Mufwene 2010:Ch. 6, Lim & Ansaldo 2015:118–27). Social factors. The two most relevant social factors at work in the Afro-European contact ecologies covered here are linguistic demography—that is, the relative proportions of speakers of the input languages—and the socioeconomic stratification of speaker populations and the languages they speak (see also Yakpo 2020). CAF, EGS, and other Afro-European contact varieties and languages all emerged in highly stratified linguistic ecologies during European colonial rule in Africa and the Americas. In these ecologies, the European speakers of European colonial languages were in the minority, while speakers of African languages constituted the overwhelming majority. The social dominance relations that typified these ecologies can be expressed in shorthand via the terms employed for sociolinguistic strata. In the colonial ecologies, the European varieties constituted socially superordinate superstrates, which were restructured by the African majority into varieties like CAF, EGS, and the English-lexifier creoles. In the following, restructured varieties like CAF and EGS are collectively referred to as contact varieties, while creoles are referred to as contact languages. Following practice in historical and comparative linguistics (e.g. Witzel 1999), the term substrate refers to a socially subordinate language that is eventually lost through shift to the contact variety or language. The African languages that fed into the development of the creole languages of the Caribbean, for example, but are no longer spoken in these ecologies are therefore substrates. Adstrates, in turn, coexist with the contact varieties. Adstrates therefore continue to provide features in a situation of long-term language maintenance and societal multilingualism, as is the case in the African contact ecologies treated here. On the outcomes of prosodic contact 3 Linguistic factors. All African adstrates and substrates (that were) spoken in these linguistic ecologies are exclusively tone languages, while the European superstrates are, without exception, languages that employ stress. Socioeconomic stratification meant that, save for a small African elite, the African majority had little to no exposure to the European colonial languages, acquired them nonnatively during adolescence or adulthood if at all, and employed these languages in functionally restricted, mostly formal domains. The combination of social and linguistic factors set the stage for structural influence from the natively acquired tone systems of the African languages on the emerging contact varieties. With respect to linguistic factors, we draw on Van Coetsem’s (2000) psycholinguistic metaphor of agentivity. The language that has agentivity is the language that provides features to the (emerging) contact variety or language. To account for the differing roles of input languages in contact-induced change, Van Coetsem makes the distinction between the recipient language (RL) and the source language (SL). In the case of RL agentivity, features from a weaker, nonnative SL are borrowed into the speaker’s dominant or native RL. SL agentivity represents the opposite case. The speaker uses a nonnative, weaker RL, and features from her dominant or native SL are transferred to the RL by imposition (referred to as ‘interference’ in earlier work; see Weinreich 1953, Thomason & Kaufman 1988). SL agentivity therefore manifests itself as substrate (in cases of shift from the SL to the RL) or adstrate (in cases of maintenance of the SL and the RL) influence on the RL. The Afro-European contact scenarios treated here all constitute cases of SL agentivity. Linguistic categories have different degrees of entrenchment in the minds of speakers and hence enjoy varying degrees of stability with respect to the likelihood of transfer (see Matras 2009:193–233 for an overview). The more stable the structural category is, the less likely it is to undergo transfer. Phonology is the most stable category of a natively acquired linguistic system (Van Coetsem 2000). Within phonology, prosody (defined as stress, intonation, and tone by Matras (2009:231–33)) has been noted as particularly stable. Under RL agentivity, prosody is therefore the least likely category to be borrowed from an SL into an RL. With SL agentivity, it is the opposite. Prosody, including tone, is the most likely category to be imposed on the RL. When an initially nonnative RL becomes nativized to some degree in a situation of SL agentivity, as has been the case in the scenarios investigated here, features that have been imposed from the SL can be conventionalized as the RL develops into a distinct contact variety (in the case of CAF and EGS) or language (in the case of Afro-European creoles). A final aspect of Van Coetsem’s model is that SL and RL agentivity are not mutually exclusive. They can operate at the same time, as in areas characterized by long-term ‘small-scale’ (Lüpke 2016) or ‘egalitarian’ (François 2012) multilingualism. This makes allowance for the convergence and fusion outcomes between tone and stress systems that have also been documented (e.g. in the case of Papiamentu; see §4.1). In the ecologies of the Central African Republic and Equatorial Guinea, demographic imbalances and social stratification, as well as SL agentivity in tone languages, therefore ushered in the speciation of CAF and EGS as Romance varieties with tone systems, distinct from their European siblings Standard European French and Standard European Spanish, which employ stress. A complete understanding of prosodic contact would require a detailed survey of outcomes under SL and RL agentivity. The focus of this study on SL agentivity in Afro-European contact scenarios is a first step in that direction, from which we can eventually arrive at conclusions about prosodic contact of a more general nature. 4 LANGUAGE, VOLUME 96, NUMBER 1 (2020) 2.2. Prosodic types and contact constellations. Two types of prosodic systems are distinguished with respect to the phonological use of pitch (Gussenhoven 2004): (i) tone systems and (ii) intonation-only systems. Tone systems. In tone systems, ‘both pitch phonemes and segmental phonemes enter into the composition of at least some morphemes’ (Welmers 1959:2). Pitch is therefore tone when it is part of the lexical realization of a morpheme. Tone manifests itself in ‘the output of the lexical phonology’ by being inherent to the morpheme, or by assignment through a rule (Hyman 2006:230). There are substantial differences between languages in the types and numbers of tones they employ (see Gussenhoven 2004:26–28). Equally, in most, if not all, tone languages, the placement of tones and hence the number of possible tone patterns per word (i.e. word-tone patterns) is conditioned to some degree by metrical and segmental structure (for Africa, see e.g. Downing 2010). Sango, the major adstrate of CAF, for example, allows its three tones (high, mid, and low) to occur in many different combinations over a prosodic word, which gives rise to a large number of word-tone patterns (see §3.3). Other languages may allow only a single type of tone per prosodic word (e.g. a high tone) and fewer permutations of tones over a word than theoretically possible, which leads to a lower number of word-tone patterns. This is the case in many Bantu languages (Downing 2010:410–16), as well as in CAF and EGS, as will be shown (see §3.6 and §3.7). Intonation-only systems. Intonation-only systems make use of stress. A language with stress is one in which there is an indication of word-level metrical structure meeting the following two central criteria: (1) obligatoriness (i.e. every lexical word has at least one syllable marked for the highest degree of metrical prominence (primary stress)), and (2) culminativity (i.e. every lexical word has at most one syllable marked for the highest degree of metrical prominence). (Hyman 2006:231) The acoustic correlates of stress are language-specific but usually involve a combination of the cues of length, loudness, vowel quality, and pitch variations over the stressbearing syllable. Pitch contours of utterances in intonation-only systems are composed of intonational pitch accents anchored to stressed syllables, and boundary tones associated with the edges of phrases and boundaries. Contrary to tone, as defined above, the term intonational pitch accent refers to anchors in the resulting intonational tunes. Intonational pitch accents are therefore also instantiated by pitch, but in contrast to tone, they are assigned above the morpheme level. It is important to underline two aspects of this typology. First, pitch is used in intonation-only and tone systems for the purposes of intonation, that is, for marking phrase boundaries and expressing pragmatic functions (see e.g. Downing & Rialland 2016a). The difference between the two prosodic types is that intonation-only systems use pitch distinctively only in this way (hence ‘intonation-only’) and pitch is not part of the composition of any morpheme. Second, languages can combine intonation-only and tone features. Languages with tone can therefore also have stress and intonational pitch accents (Hyman 2006). In other words, tone languages can have all features that characterize intonation-only languages, but not vice versa. Prosodic contact constellations. Based on the two types of systems, intonationonly (IN) and tone (TO), the following four prosodic contact constellations and outcomes are possible in the situation of SL agentivity that this study is concerned with. The subscript typifies the sociolinguistic stratum that either of the two prosodic systems is associated with in the four constellations, hence superstrate (sup), substrate (sub), adOn the outcomes of prosodic contact 5 strate (ads), and contact variety or language (con). The plus sign designates contact between two strata, and the equals sign the outcome of such contact. ii(i) INSUP + TOSUB/ADS = TOCON i(ii) TOSUP + INSUB/ADS = INCON (iii) TOSUP + TOSUB/ADS = TOCON (iv) INSUP + INSUB/ADS = INCON We are exclusively concerned with constellation (i) (INSUP + TOSUB/ADS = TOCON). The inverse constellation (ii) (TOSUP + INSUB/ADS = INCON) is outside the scope of this study because the outcome is an intonation-only contact variety in an ecology involving a tonal superstrate and SL agentivity in intonation-only substrates or adstrates. Constellation (ii) is attested, for example, in northern Norway (Jahr 1984, Bull 1995) and southern Finland (Bruce 2004), where the tonal superstrates Norwegian and Swedish have undergone extensive substratal and adstratal transfer from the intonation-only languages Sami and Finnish, and Finnish, respectively. The resulting contact varieties of Norwegian and Swedish feature intonation-only systems. Constellations (iii) and (iv) are not relevant either. For the present purpose, ‘prosodic contact’ only designates contact between intonation-only and tone systems. The outcomes of contact between different types of tone systems (constellation (iii)) or intonation-only systems (constellation (iv)) are therefore not addressed (for tone-tone contact, see e.g. Vydrine 2004 for Africa, Matisoff 2001 for Asia). Constellation (iv) is probable, for example, in the formation of the Portuguese-lexifier creole of Cape Verde. Lang (2009) argues that Wolof was one of the major substrates of Cape Verdean Creole. Wolof belongs to the Atlantic subfamily of the Atlantic-Congo family, one of the few of Sub-Saharan Africa to feature intonationonly systems (Downing 2010:381). The four initial constellations above are, of course, dynamic in nature and susceptible to diachronic change, for example through areal contact and convergence (see §6). Changes in prevailing social and linguistic factors can lead to changes in agentivity, and consequently to shifts from tone to intonation-only systems in contact varieties and languages over time, even in ecologies that are initially dominated by tonal substrates or adstrates. This is very likely to have occurred in the Caribbean creoles (see §4.1). 3. Prosodic systems of central african french and equatorial guinean spanish. We now present a case study of the prosodic systems of CAF and EGS. A description of the ecologies in which they emerged and are spoken (§3.1) is followed by an overview of the prosodic systems of their superstrates (§3.2) and adstrates (§3.3). A presentation of the research methods and the data (§3.4) precedes the phonetic (§3.5) and phonological analysis of CAF (§3.6) and EGS (§3.7). Finally, we provide a comparison of the prosodic systems of the contact varieties, their adstrates, and their superstrates (§3.8). The somewhat detailed analysis of CAF and EGS in this section prepares the ground for determining the place of their prosodic systems in the broader context of Afro-European contact and allows us to identify specific mechanisms in the emergence of contact prosodic systems (§4). 3.1. Ecologies. We proposed in §2.1 that both social and linguistic factors determine the outcomes of prosodic contact. The case studies therefore begin with an overview of the complex linguistic ecologies of the Central African Republic and Equatorial Guinea, focusing on adstrates and superstrates and their relation to each other (see n. 1 for a definition of ‘linguistic ecology’). Adstrates. Both ecologies are highly multilingual. Seventy-two languages are estimated to be spoken in the Central African Republic (CAR), and most of these belong to 6 LANGUAGE, VOLUME 96, NUMBER 1 (2020) the Ubangian subbranch of the Atlantic-Congo family (Hammarström, Forkel, & Haspelmath 2017). At least fourteen languages are spoken by the peoples of Equatorial Guinea besides Spanish (Hammarström et al. 2017). Two of these are European-lexifier creoles: the Portuguese-lexifier creole Fa d’Ambô (Zamora 2010, Post 2013) and Pichi, an Afro-Caribbean English-lexifier Creole (Yakpo 2010, 2019a). All other languages belong to the Bantu subfamily of Atlantic-Congo, among which Fang and Bube have the largest speaker numbers. Moreover, both ecologies feature lingua francas: Sango in the CAR and Pichi in Equatorial Guinea. These two languages function as the principal adstrates to the varieties of CAF and EGS described in this study, namely those spoken in Bangui and Malabo, the capitals of the CAR and Equatorial Guinea, respectively. Sango is an African-lexifier creole that emerged as a trade language in the area around the Ubangui River toward the end of the nineteenth century (Pasch 1993, Samarin 2000). The bulk of its lexicon is derived from the Ubangian language Ngbandi. The sources of the grammar are more eclectic, with Ubangian languages like Ngbandi, Gbanziri, and Banda among the inputs (Samarin 2013). Today, Sango is spoken by the vast majority of Central Africans, and it has become the principal language of the capital Bangui where it tends to be the first language acquired by children (Thornell 1997). Moreover, Sango is co-official with French and is used in formal contexts, such as in courts and religious ceremonies. Although Sango is presently the dominant adstrate of CAF in Bangui, it was still consolidating itself as the primary lingua franca of the CAR in the late nineteenth and early twentieth centuries (Samarin 2013:13–14). Allowance should therefore also be made for the transfer to early CAF of tonal features from other languages of the CAR, and particularly Ngbandi, the lexifier of Sango (e.g. in the case of idiosyncratization; see §4.3). Pichi is an offshoot of Krio, a language of the family of Afro-Caribbean Englishlexifier Creoles brought to Bioko (the insular part of Equatorial Guinea) in the early nineteenth century by African settlers from Sierra Leone (Martín del Molino 1993). Pichi is also the principal adstrate to Bube, a continuum of Bantu varieties spoken by the autochthonous population of Bioko. In the course of language shift, Pichi has become the main home language and lingua franca of the Bube ethnic majority in the capital Malabo (Bolekia Boleká 2007, Morgades Besari 2011). It also functions as an urban lingua franca and a youth sociolect for people of all ethnicities and nationalities in Malabo (Yakpo 2016). EGS has undergone influence from other adstrates besides Pichi, in particular from Bube (see Yakpo 2013:287–90 for Bube influence on the segmental phonology of Pichi). Superstrates. French and Spanish were introduced into the ecologies by European colonizers. The area that today constitutes the CAR was colonized by France at the end of the nineteenth century, while Equatorial Guinea was subjected to Spanish colonial rule from the middle of the nineteenth century. Little is known about the exact nature of the linguistic input by colonial administrators, soldiers, educators, missionaries, and settlers, who disseminated French and Spanish in the early colonial period. It is, however, certain that CAF and EGS began crystalizing at the end of the nineteenth century from varieties of French and Spanish spoken in Europe. We also know with sufficient certainty that only a minute proportion of Central Africans and Equatoguineans used French or Spanish as a vernacular language in their daily affairs during the colonial period (Diki-Kidiri & Caprile 1982, Castillo-Rodríguez 2013). All adstrates of the two ecologies are tone languages, and the speakers of these adstrates at all times vastly outnumber(ed) any past or present population with a command of the European superstrate varieties of French or Spanish. The CAR and Equatorial On the outcomes of prosodic contact 7 Guinea therefore provide a typical setting for SL agentivity, in which the imposition of tone from the adstrates on idiolects and on an emergent contact variety is plausible from the very beginning of the implantation of the two colonial languages. Social status, acquisition, and practices. French and Spanish have remained the de facto and de jure official languages of the CAR and Equatorial Guinea from colonial times until today. The social functions of French and Spanish are therefore very similar in the two ecologies. French and Spanish are the languages of public administration, and school instruction is mainly dispensed in the colonial languages. In both the CAR and Equatorial Guinea, proficiency in French and Spanish indexes prestige and social status, although they are spoken by a far smaller section of the population than the autochthonous languages of these two countries. Today, they are mainly transmitted by people who themselves acquired them in the local ecologies. The linguistic models of those who learn French or Spanish, often the idiolects of teachers, are thus no longer the European superstrate varieties of before, but the local contact varieties. Further, speakers of the two contact varieties deploy them as part of multilingual repertoires characterized by extensive code-switching between Sango and CAF in the CAR, and EGS and Pichi in Equatorial Guinea (for the latter language pair, see Yakpo 2018). The prosodic characteristics described in the following have nevertheless coalesced into a rather coherent norm for the majority of speakers. These characteristics are subject to little variation and are one of the most consistent indicators of the ‘nativization’ (Schneider 2003) of CAF and EGS and the evolution of an endocentric standard. In CAF, for instance, there is far more intraand interspeaker variation in segmental phonology (Bordal 2012a) and grammar (Bordal Steien, Boutin, & Beyom 2016) as an effect of its acquisition as an additional language for the vast majority of Central Africans than there is in prosody. 3.2. Superstrate prosodic systems. This section gives a brief overview of the prosodic systems of Standard European French (SEF) and Standard European Spanish (SES), the historical superstrates of CAF and EGS. We employ the metrical-autosegmental model (Pierrehumbert 1980), which has been used for phonological descriptions of SEF, SES, and other intonation-only languages (see Ladd 2008) and therefore facilitates comparison. The description is synchronic. As mentioned in §3, not much is known about the exact nature of the prosodic systems from which CAF and EGS developed. We therefore extrapolate from the prosodic systems of SEF and SES as they are described by contemporary phonological models. Standard european french. SEF is one of the best-described languages in the world. There is therefore a corresponding level of factionalism with respect to the analysis of the entire spectrum of SEF prosodic features. Because most of the fine theoretical details of these analyses do not matter for the key aspects covered in this study, we do not delve into them here. Instead, the focus is placed on aspects of the system that are relevant for the comparison with CAF. SEF has phrasal stress that primarily occurs on the last syllable of the accentual phrase, a prosodic constituent that can consist of one or more content words and dependent function words, for example, des gentilles filles ‘nice girls’ (Avanzi, Bordal, & Nimbona 2014). The accentual phrase also has an optional secondary stress on one of its first syllables whose realization depends on factors like speech rate, syntactic constituency, emphasis and focus marking, style, and idiolectal particularities, for example, des gentilles filles. In addition to pitch variations, the main acoustic correlate of stress in SEF is vowel length. 8 LANGUAGE, VOLUME 96, NUMBER 1 (2020) The final stressed syllable of accentual phrases is in most cases aligned with a high intonational pitch accent, notated as H*. If the accentual phrase has secondary stress, the initial stressed syllable is also associated with a high intonational pitch accent, notated Hi (Jun & Fougeron 2000, 2002, Delais-Roussarie et al. 2015). Moreover, SEF has two optional low intonational pitch accents, notated as aL and L (Delais-Roussarie et al. 2015). The first target, aL, can occur at the beginning of an accentual phrase, and the second, L, somewhere between Hi and H*. The default and most common pitch contour of the SEF accentual phrase is thus a rising-falling-rising pattern, which can be represented by the template /(aL)(Hi)(L)H*/, in which optional targets are in parentheses (cf. Jun & Fougeron 2002, Delais-Roussarie et al. 2015). Depending on various factors, the accentual phrase is realized with different pitch contours, such as [aLH*], [aLHiH*], [aLHiL*], and so forth. Further, SEF makes use of boundary tones over the final syllable(s) of prosodic constituents larger than the accentual phrase, such as utterance-final H% and L%. Different intonational contours are constituted by the interplay of utterance-internal intonational pitch accents and boundary tones. These contours serve pragmatic functions and differentiate declaratives from yes/no questions, for instance (Delais-Roussarie et al. 2015). Standard european spanish. While SEF features phrasal stress, SES makes use of word stress. Stress occurs on one of the three last syllables of the prosodic word. Stress placement is largely conditioned by phonotactics and membership in word classes (e.g. nouns vs. verbs) and by syntactic categories (content vs. function words). The position of stress is therefore both phonologically derived and morphologically determined (Roca 2005). About 80% of SES content words bear penultimate stress, and about 18% have stress on the final syllable. Altogether, 95% of nouns and adjectives carry stress on the final vowel of the base, for example african-o, universidad(Quilis 1978, Hualde 2012). The stressed syllable of utterance-internal words can be aligned with different kinds of intonational pitch accents depending on the realized speech act, focus conditions, and so forth. Stress is realized by a culmination of acoustic cues, mainly length and loudness, in addition to pitch variations. The most common intonational pitch accent of utterance-internal prosodic words is rising pitch, referred to as /L+H*/ or /L+<H*/ (Hualde & Prieto 2015). The ‘<’ of the latter template indicates that the pitch peak is delayed, hence realized after the stressed syllable. Such a contour occurs in prenuclear position of broad-focus declarative utterances, while a pitch peak on the stressed syllable itself conveys contrastive focus or emphasis (for examples of other intonational pitch accents and their pragmatic meanings, see Hualde & Prieto 2015). SES also makes use of boundary tones over the final syllable(s) of prosodic constituents larger than the prosodic word, such as utterance-final H% and L%. As in SEF, utterance-internal intonational pitch accents and boundary tones constitute intonational contours that mark different question and statement types (Hualde & Prieto 2015). Prosodic typology. The intonational pitch accents in SEF and SES are different from word-level tones, since they are not a part of the composition of morphemes (see §2.2). They instead constitute the building blocks of intonational contours that express pragmatic functions (e.g. emphasis, counter-expectation, yes/no questions) as a whole. This is why SEF and SES are classified as intonation-only languages (for SEF, see Gussenhoven 2004:253). 3.3. Adstrate prosodic systems. In the cities of Bangui and Malabo, where we collected our data, Sango and Pichi dominate in everyday conversations, and thus function On the outcomes of prosodic contact 9 Polysyllabic function words. Polysyllabic function words behave the same way as polysyllabic content words. They have L tones on the first syllable(s) and an H tone on the last syllable, and hence feature an /(L)H/ pattern. Prosody beyond the word level. CAF has boundary tones that mark the right edge of utterances. Unlike the word-level tone patterns covered above, boundary tones instantiate the pragmatic functions of intonation and are therefore assigned to utterance and phrase boundaries. CAF has a least two boundary tones, L% and H%, which express declarative and continuative intonation, respectively. When an utterance is marked by an L% tone, the pitch of the utterance-final syllable starts at a higher point than that of the preceding L-toned syllable and then falls. This indicates that the H tone of the final syllable is not deleted by the L% boundary tone. Both tones are therefore realized together as an HL% contour over the same utterance-final syllable. When an utterance is marked by an H% boundary tone, the pitch of the utterance-final syllable culminates on a higher level than a preceding H tone within the same utterance. No evidence was found for utterance-level downstep in CAF. Illustration. The tonal system of CAF can be illustrated by an example consisting of the two utterances displayed in Figure 2. The figure presents the spectrogram and pitch contour of Je suis née à Mobaye. Mon père fut un instituteur ‘I was born in Mobaye. My father was a school teacher’ (SP7).8 16 LANGUAGE, VOLUME 96, NUMBER 1 (2020) 8In the pictures we use to illustrate our analysis, the first tier under the spectrogram represents our phonological annotation of tones. The second tier shows the segmentation into syllables and the SAMPA phonemic transcription. The last tier contains the segmentation into orthographically transcribed words. The pictures type words Singular indefinite determiners un ‘indf.sg.m’, une ‘indf.sg.f’ Demonstratives ce/cette/ces ‘this/that/these/those’ (adj) celle/ceux ‘this/that/these/those’ (pron) Some subject pronouns tu ‘2sg.sbj’, on ‘3sg.sbj/1pl.sbj’ nous ‘1pl.sbj’, vous ‘2pl.sbj’ ils ‘3pl.sbj.m’, elles ‘3pl.sbj.f’ Some adnominal possessives mon ‘1sg.poss.m’, ton ‘2sg.poss.m’, son ‘3sg.poss.m’ notre ‘1pl.poss’, votre ‘2pl.poss’ Some prepositions pour ‘for’, dans ‘in’, par ‘by’, en ‘at’ Plural indefinite determiner des ‘indf.pl’ Definite determiners le ‘def.sg.m’, la ‘def.sg.f’, les ‘def.pl’ Some subject pronouns je ‘1sg.sbj’, il ‘3sg.sbj.m’, elle ‘3sg.sbj.f’, ce ‘3sg.sbj’ All object pronouns me ‘1sg.obj’, te ‘2sg.obj’ le ‘3sg.acc.m’,la ‘3sg.acc.f’, les ‘3pl.acc’ se ‘3sg/pl.obj.refl’ nous ‘1pl.obj’, vous ‘2pl.obj’ lui ‘3sg.dat’, leur ‘3pl.dat’ Some prepositions à ‘at’, de ‘of’ Table 4. Tone in CAF monosyllabic function words found in the corpus. word tone gloss word tone gloss ce H ‘this/that’ (adj) ce L ‘it/that’ (pron) nous H ‘1pl.sbj’ nous L ‘1pl.obj’ vous H ‘2pl.sbj’ vous L ‘2pl.obj’ ils H ‘3pl.sbj.m’ il L ‘3sg.sbj.m’ elles H ‘3pl.sbj.f’ elle L ‘3sg.sbj.f’ Table 5. Tonal minimal pairs in the CAF corpus. H tone L tone The monosyllabic content words in Fig. 2 (suis ‘am (= was)’, née ‘born’, fut ‘was’, père ‘father’) invariably bear an H tone. The rightmost syllables of the polysyllabic content words Mobaye [mò.báj] (a place name) and instituteur ‘school teacher’ also bear an H tone. Further, the pitch contour falls sharply toward the first L-toned syllable of instituteur from the preceding H tone over un ‘indf’. Rises from L to H are equally steep, as shown by the transition between the penultimate and the final syllables of instituteur. The syllables without H tones thus bear L tones. Figure 2 also shows two utterance-final declarative L% boundary tones, which mark the end of each utterance. They occur on the last syllables of Mobaye [báj] and instituteur [tœ´] and induce sudden drops over the final syllables of each of these words. 3.7. Equatorial guinean spanish. The first indication that the prosodic system of EGS may be tonal was provided by Yakpo 2009. The pitch contours of utterances involving code-switched EGS words in mixed Pichi-EGS discourse led Yakpo to conclude that in EGS, ‘Spanish primary stress [gets] reinterpreted as a high tone’ (2009: 13). Lipski’s later analysis of a number of EGS utterances, although framed in the methodology and terminology proper to the study of intonation-only systems, prompted him to conclude that EGS words carry a single ‘pitch accent’ on the syllables that bear stress in SES (Lipski 2015:247). In other words, EGS speakers place the highest pitch on the stressed syllable and not on the syllable after the stressed one, as is usually the case in SES declaratives (see §3.2). Lipski further shows via pitch measurements and a statistical summary that EGS words ‘approach tonal invariance’. We interpret this to mean that EGS word-tone patterns are fixed. Lipski’s study does not provide an acoustic or phonological analysis of the pitch patterns of individual EGS words or an overview of the distribution of wordtone patterns. Nor does it mention the presence of tonal minimal pairs nor analyze tonal processes and the functions of boundary tones. Lipski also concludes (mistakenly, as we show in due course) that EGS words occur in ‘citation forms’ in connected speech (in our understanding, this implies that all utterance-medial words feature a declarative L%) and that there is no H-tone spreading in EGS (Lipski 2015:248). In contrast to the two sources cited above, this section therefore contains the first fuller account of the main features of the prosodic system of EGS. On the outcomes of prosodic contact 17 were generated by the Praat script create_pictures v.4.5 developed by the Laboratory of Phonetics at the University of Barcelona (http://stel.ub.edu/labfon/en/praat-scripts). The sound files for the spectograms in Figs. 2 and 3 can be accessed online at http://muse.jhu.edu/resolve/91. Figure 2. Pitch contour of two CAF utterances. Tones and word-tone patterns. Like CAF, the tonal system of EGS is characterized by the presence of two level tones, H and L. Equally, tone patterns apply to the prosodic word inclusive of affixes (e.g. matricul-amos /LLLHL/ ‘matriculate-1pl.prs’). Four possible tone patterns were identified for content and function words, and these are provided in Table 6. The word-tone patterns of EGS words are again listed separately for the two syntactic categories, but identical patterns bear the same number, irrespective of the category. 18 LANGUAGE, VOLUME 96, NUMBER 1 (2020) syntactic tone patterns examples categories 1 /(L)H/ pan /H/ ‘bread’ edad /LH/ ‘age’ educación /LLLH/ ‘education’ content words 2 /(L)HL/ clase /HL/ ‘class’ picante /LHL/ ‘spicy’ matricul-amos /LLLHL/ ‘matriculate-1pl.prs’ 3 /(L)HLL/ jóvenes /HLL/ ‘youths’ película /LHLL/ ‘film’ 1 /(L)H/ yó /H/ ‘1sg.sbj’ porqué /LH/ ‘why?’ 2 /(L)HL/ esta /HL/ ‘this’ function words nosotros /LHL/ ‘1pl.sbj’ 4 /(L)L/ me /L/ ‘1sg.obj’, en /L/ ‘in’, a/L/ ‘at, to’ desde /LL/ ‘since’ excepto /LLL/ ‘except’ Table 6. Word-tone patterns in EGS. Content words. As in CAF, all content words bear at least one and at most one H tone. However, H tones are distributed in a more arbitrary fashion in EGS and are found in three word positions, namely on the final, penultimate, and antepenultimate syllable. The most common tone pattern over EGS polysyllabic words in the corpus features an H on the penultimate syllable (103/171 = 60.2%), followed by an H on the final (52/171 = 30.4%) and the antepenultimate syllables (16/171 = 9.4%). As in the CAF corpus, there are sharp falls and rises between adjacent H and L tones from one syllable to the next rather than gradual falls between nonadjacent H tones over the span of several syllables. Further, nowhere in the EGS data is there variation in the relative pitch of the tones that form the four word-tone patterns of individual content and function words identified in Table 6, even in the most emphatic of contexts. We deduce from this that in EGS, too, a syllable without an H tone bears an L tone. The only exception is in the clearly circumscribed context of H-tone spreading, a tonal process with the following characteristics. H-tone spreading. In some utterances, the L-toned syllable(s) right-adjacent to an H-toned syllable is/are also realized with high pitch. This therefore applies only to polysyllabic words without a final H, hence with tone patterns 2 /(L)HL/ and 3 /(L)HLL/. These pitch realizations are instances of H-tone spreading. The domain of H-tone spreading is limited to the prosodic word (stem + affixes and host + clitic), and spreading exclusively occurs rightward. Spreading is common but not systematic in the corpus. There is variation within the same speakers and within syntactic categories (i.e. nouns, verbs, and adjectives). Tone spreading may therefore be an incipient process in EGS that is yet to consolidate itself. We considered, and eventually discarded, the alternative explanation that the rightmost H in words that otherwise have an /(L)HL/ pattern might instead be a boundary tone of some prosodic constituent, such as an accentual or intermediate phrase. However, H-tone spreading is not limited to the final syllable. Instead, it occurs rightward up to the word boundary, irrespective of the position of the lexical H tone. Hence in words with an antepenultimate H tone such as jóvenes ‘youths’, we find /HLL/ → [HHH]. Function words. As in CAF, the lexical tone of monosyllabic function words is arbitrary and can be either H or L. Some polysyllabic function words bear one and only one H tone, like polysyllabic content words, while others bear only L tones. There is therefore a small group of polysyllabic function words that have only L tones. An exhaustive list of function words found in the corpus with their tone patterns is given in Table 7. On the outcomes of prosodic contact 19 Minimal pairs. Like CAF, EGS also has a few tonally distinguished minimal pairs. In contrast to CAF, where minimal pairs are encountered only among monosyllabic function words, EGS minimal pairs are found among both monosyllabic and disyllabic function words. All minimal pairs in the corpus are listed in Table 8. tone type words patterns 1 /(L)H/ Indefinite determiners and demonstratives un /H/ ‘indf.sg.m’ Subject and independent pronouns yo /H/ ‘1sg.sbj’, mí /H/ ‘1sg.indp’ tú /H/ ‘2sg.sbj’, él /H/ ‘3sg.sbj.m’ Question words qué ‘what?’, quién /H/ ‘who?’ Conjunction después /LH/ ‘after’ 2 /(L)HL/ Demonstratives esta /HL/ ‘this’, eso /HL/ ‘that’ Indefinite determiners and demonstratives unos /HL/ ‘indf.pl.m’ Subject and independent pronouns ella /HL/ ‘3sg.sbj.f’ nosotros /LHL/ ‘1pl.sbj’, vosotros /LHL/ ‘2pl.sbj’ ellos /HL/ ‘3pl.sbj.m’, ellas /HL/ ‘3pl.sbj.f’ Question words cómo /HL/ ‘how?’, cuál(es) /HL/ ‘which (ones)?’ 4 /(L)L/ All definite determiners el /L/ ‘def.sg.m’, la /L/ ‘def.sg.f’ los /L/ ‘def.pl.m’, las /L/ ‘def.pl.f’ Adnominal possessives mi /L/ ‘1sg.poss’, tu /L/ ‘2sg.poss’ su /L/ ‘3sg.poss’, mis /L/ ‘1pl.poss’ Object pronouns me /L/ ‘1sg.obj’, te /L/ ‘2sg.obj’ lo/la /L/ ‘3sg.acc.m/f’, le /L/ ‘3sg.dat’ se /L/ ‘3sg/pl.obj.refl’ nos /L/ ‘1pl.obj’, os /L/ ‘2pl.obj’ los/las /L/ ‘3pl.acc.m/f’, les /L/ ‘3pl.dat’ Conjunctions & relativizers que /L/ ‘sub’ como /LL/ ‘how, as (well as)’, pero /LL/ ‘but’ porque /LL/ ‘because’, para /LL/ ‘in order to’ Some prepositions a/L/ ‘to’, de /L/ ‘of’, por /L/ ‘by’, con /L/ ‘with’ para /LL/ ‘for’, como /LL/ ‘like’ hasta /LL/ ‘until’, mientras /LL/ ‘while’ excepto /LLL/ ‘except’ Table 7. Tone in EGS function words found in the corpus. word tones gloss word tones gloss qué H ‘what?’ que L ‘that’ (sub) cómo HL ‘how?’ como LL ‘like’; ‘how’ (rel) esta HL ‘this’ está LH ‘cop.loc.3sg.prs’ él H ‘he’ (3sg.sbj.m) el (que) L ‘he (who)’ (3sg.sbj.rel.m) tú H ‘you’ (2sg.sbj) tu L ‘your’ (2sg.poss) Table 8. Tonal minimal pairs in the EGS corpus. Prosody beyond the word level. Just like CAF, EGS makes use of boundary tones on the right edge of utterances. We have, for now, identified an L% and an LH% (con- tour) boundary tone. The L% occurs in pragmatically neutral declarative utterances, and its realization begins at the pitch level of the tone of the utterance-final syllable (L or H). The L% then provokes a steep fall over the utterance-final syllable, with a drop to a level lower than the pitch of a preceding L tone. LH% is a contour boundary tone that occurs over the utterance-final syllable in yes/no questions. The H tone that forms part of the LH contour may rise above the height of a preceding H tone in the utterance-final word, which then conveys pragmatic nuances like emphasis or counter-expectation in addition to polarity. Further, EGS utterances systematically display downstep: noninitial H tones are lower than preceding ones, and the slope thus falls gradually from the beginning to the end of the utterance. Illustration. The following spectrogram and pitch contours illustrate most of the pitch characteristics described for EGS above. Figure 3 shows the two EGS utterances Que no abandonéis los estudios. Tanto chicos como chicas ‘You should not give up your studies. Neither boys nor girls’ (SP12). 20 LANGUAGE, VOLUME 96, NUMBER 1 (2020) Figure 3. Pitch contour of two EGS utterances. The two utterances in Fig. 3 contain four polysyllabic content words that reflect the most frequent tone patterns of EGS content words. The word abandon-éis ‘abandon-2pl.prs’ has the pattern /(L)H/, while estudio-s ‘study-pl’, chico-s ‘boy-pl’, and chica-s ‘girl-pl’ have the pattern /(L)HL/. Figure 3 also contains the L-toned monosyllabic function word que /L/ ‘sub’ and the polysyllabic, yet L-toned, adverbial/relative clause introducer como /LL/ ‘how, as (well as)’. Sharp falls and rises between H and L tones rather than interpolation between H tones over the span of several syllables are illustrated by the transition between the H tone of no and the L tone on the first syllable of abandon-éis ‘abandon-2pl.prs’. Further, the behavior of pitch over chico-s ‘boy-pl’ indicates H-tone spreading: the H tone of the first syllable [tʃí] spreads to the second—and last—phonologically L-toned syllable [kò] (which is therefore annotated as L). The two utterances also illustrate prosodic characteristics of EGS beyond the word level. Both feature utterance-level downstep. This is reflected in the gradual downward slope of the entire utterance. Finally, the presence of declarative L% tones over the final syllable of estudios [djɔ`s] ‘studies’ and the final syllable of chicas ‘girls’ [kàs] produces final falls in the pitch register of each of the two utterances. 3.8. Prosodic systems compared. We now provide a comparison of the central aspects of the prosodic systems of the contact varieties CAF and EGS with those of the superstrates SES and SEF as well as their adstrates Sango and Pichi. A comparison is useful for identifying the contact-related aspects of the CAF and EGS systems and, in a second step, situating them in the context of the emergence of contact prosodic systems, which follows in §4. The prosodic characteristics discussed in the previous sections are listed as features in Table 9. A plus sign (+) indicates the presence of a particular feature, a minus sign (−) its absence. Both signs (+/−) are used when the variety has the feature in some parts of the lexicon but not in others. The same abbreviations used in §2.2 apply here: superstrate (sup), substrate (sub), adstrate (ads), contact variety (con). On the outcomes of prosodic contact 21 9A more complete picture of the situation in Pichi with respect to features 3 and 4 is the following (also see §3.3): we find the characteristic obligatoriness and culminativity of H in most content words of English origin. However, monosyllabic English-derived function words bear an arbitrary L or H. In addition, Pichi has a stock of African-derived words with more diverse tone patterns, among them multisyllabic words bearing only L or H tones (for Pichi, see Yakpo 2019a:45–46; for Krio, see Hancock 1971, Finney 2004). The layered nature of the lexicon of Pichi is reflected by (+/−) in the corresponding cells of features 3 and 4 in Table 9. features central african republic equatorial guinea con ads sup con ads sup CAF Sango SEF EGS Pichi SES 1. Tones ++−++− 2. Tonal minimal pairs ++−++− 3. Obligatoriness of word-level + −− + +/−+ H or stress in content words 4. Culminativity of word-level + −− + +/−+ H or stress in content words 5. H-tone spreading −−− ++− 6. Downstep −−− ++− 7. Intonational pitch accents −−+−−+ 8. Boundary tones +++ +++ Table 9. Comparison of prosodic features. Tones (feature 1). Every syllable in CAF and EGS bears a low (L) or high (H) tone, and words have fixed tone patterns, indicating that tones are part of the composition of morphemes (see §2.2). Feature 1 is thus exclusively found in CAF, EGS, and their adstrates Sango and Pichi, but not in SEF and SES. Tonal minimal pairs (feature 2). CAF and EGS have tonally distinguished minimal pairs in the category of function words (see Tables 5 and 8). This is also a feature they share with their adstrates, although in the latter, tonal minimal pairs are also found among content words. The superstrate SES has minimal pairs distinguished by stress placement and the presence vs. absence of stress. Both EGS and SES therefore have prosodic minimal pairs, but the difference between EGS and its superstrate SES is the prosodic feature that distinguishes segmental homophones. SEF, in contrast, does not have minimal pairs that are distinguished prosodically. This shows that CAF has developed in this regard an idiosyncrasy vis-à-vis its superstrate that EGS does not have visà-vis its own superstrate (also see §4.3). Obligatoriness (feature 3) and culminativity (feature 4). The presence of at least one (feature 3) and at most one (feature 4) H or primary stress characterizes content words in CAF, EGS, Pichi, and SES.9These four varieties/languages differ from Sango and SEF in this regard but for very different reasons: while SEF does not have word stress, Sango has few restrictions on tone combinations in content and function words. As a result, Sango has many words with more than one H tone or with L tones alone (see §3.3). Turning to EGS, tone patterns in content words faithfully reproduce those of stress patterns in SES. For instance, words with the pattern /(L)H/ in EGS have stress on the last syllable in SES, and so forth. In contrast, CAF H tone placement less directly reflects SEF stress placement. In SEF, the smallest prosodically marked constituent is the accentual phase, which is often larger than the word (see §3.2). Hence in natural discourse, not every content word necessarily bears stress in SEF. Nevertheless, the most common position for primary stress in SEF is the final syllable of the accentual phrase, which is very often also the final syllable of a content word. In that sense, the position of H tone in CAF content words mirrors the most frequent position of stress at the word level in SEF. On the whole, the binary contrast between stressed and unstressed syllables in SEF and SES has been converted into a binary H vs. L contrast in CAF and EGS. This translates into culminative and obligatory H tone placement in CAF and EGS content words. Due to its canonical placement on the final syllable, the position of H tone in CAF content words is entirely predictable. In EGS, however, the position of H is arbitrary from a synchronic point of view, leaving aside knowledge of the stress pattern of the SES etymon. H-tone spreading (feature 5) and downstep (feature 6). The processes listed under features 5 and 6 are typical of tone systems and hence not found in either of the superstrates. They are, nevertheless, not reported for Sango, and no evidence was found for them in CAF. EGS, by contrast, has both H-tone spreading and downstep, and both processes are also found in Pichi. Intonational pitch accents (feature 7). This feature is attested only in SEF and SES. There is no evidence in our data that word-level tones in CAF and EGS may be altered by intonational pitch accents that form part of intonational tunes spanning the utterance. In CAF and EGS, utterance-final boundary tones fulfill the pragmatic functions associated with intonational pitch accents and intonational tunes in SEF and SES. Intonational pitch accents are not found in the adstrates Pichi and Sango either. Boundary tones (feature 8). Boundary tones fulfilling the functions of intonation are common both to intonation-only and to tone systems, and they are present in the superstrates, the adstrates, and the contact varieties. Summary. The comparison shows that, on the one hand, CAF and EGS possess acoustic and phonological features associated with tone systems, namely tone inherent to morphemes, word-tone patterns, minimal pairs, tonal processes, and the absence of intonational pitch accents. On the other hand, both contact varieties also have two structural features generally identified with intonation-only systems, that is, obligatoriness and culminativity. The relation between tone and intonation-only features in the contact varieties is discussed further in the following section, where we suggest three mechanisms that drive the emergence of contact prosodic systems. 4. Mechanisms in the emergence of contact prosodic systems. Prosodic contact was defined in general terms in §2.2 as contact between tone and intonation-only systems. The synchronic description of the resulting contact prosodic systems CAF and EGS led to the conclusion in §3.8 that they possess defining features of tone systems. At the same time, they share structural features with intonation-only systems: H tone in the tonal contact varieties and primary stress in the intonation-only superstrates are obligatory and culminative. These characteristics of contact prosodic systems therefore also raise a question of a diachronic nature. How do contact prosodic systems emerge and consolidate them22 LANGUAGE, VOLUME 96, NUMBER 1 (2020) selves? It is well known that the distribution of prosodic systems across the globe is largely areal (Matisoff 2001, Gussenhoven 2004:42–45, Clements & Rialland 2008:74). Switches from one prosodic type to another are necessary in order to arrive at an areal distribution. In spite of this, there has not been, to our knowledge, an attempt to identify the precise mechanisms by which contact prosodic systems have emerged in Afro-European contact ecologies. We make such an effort in the following by drawing on the preceding analyses of CAF and EGS and by adducing additional evidence from African varieties of English and French, and Afro-European creole languages. We argue that three mechanisms are at work, namely stress-to-tone mapping (§4.1), paradigmatization (§4.2), and idiosyncratization (§4.3). We also discuss the persistence of tone in language contact with respect to the composition of linguistic ecologies (§5). 4.1. Stress-to-tone mapping. Most CAF and EGS words bear H tone where the superstrate etymons bear stress. We refer to the mechanism through which stress in the superstrate is converted into H tone in contact varieties and languages as stress-to-tone mapping. Stress-to-tone mapping relies on the perception by native speakers of tone languages of the stress vs. nonstress contrast as an H vs. L contrast. Beyond CAF and EGS, stress-to-tone mapping has also shaped the prosodic systems of a wide range of Afro-European contact varieties and languages that emerged in ecologies with SL agentivity in tone languages. Stress-to-tone mapping in other afro-european contact varieties. Gussenhoven and Udofot (2010) and Gussenhoven (2017) present acoustic and phonological evidence that Nigerian English has inherent word-tone patterns like CAF and EGS. Nigerian English also has two tones, H and L. The syllable with primary stress in the British English cognate receives an H, as in member /HL/. While content words feature a culminative and obligatory H, function words that normally remain unstressed in British English are L-toned in Nigerian English, as in of /L/, a/L/, and he /L/. Ghanaian English also has a two-tone H vs. L contrast, fixed word-tone patterns, a culminative and obligatory H in content words, and L-toned function words (Criper 1971, CriperFriedman 1990). The prosodic systems of African Romance varieties other than CAF and EGS have not been studied in great detail. The available sources suggest, however, that stress-totone mapping also characterizes the varieties of French spoken in Côte d’Ivoire (Boutin & Turscan 2009) and Mali (Bordal & Skattum 2014). In both varieties, a high tone is realized on the right edge of every content word, as in CAF. Stress-to-tone mapping in afro-european creoles. Like the colonial contact varieties, all English-lexifier creoles of Africa are spoken in ecologies that exclusively feature tone systems. Among these languages, the prosody of Pichi, the major adstrate of EGS, has been studied in greatest depth so far (Yakpo 2009, 2019a). Other works cover aspects of the prosodic systems of Krio (Berry 1970, Hancock 1971, Nylander 1984, Finney 2004), Nigerian Pidgin (Faraclas 1996), Cameroon Pidgin English (Nkengasong 2016), and Ghanaian Pidgin English (Huber 2013). The sources suggest that the prosodic systems of these languages show the characteristic outcomes of stress-to-tone mapping. They have two-tone systems in which most content words of English origin feature an obligatory and culminative H tone, while African-sourced content words show diverse tone patterns, including ones with no H tone at all. Tone systems that evolved through stress-to-tone mapping are also found in the African Portuguese-lexifier creoles Forro (Ferraz 1979, Maurer 2008) and Angolar On the outcomes of prosodic contact 23 (Maurer 1995), both spoken on São Tomé, and Lung’le (Traill & Ferraz 1981, Maurer 2009), spoken on Príncipe; see Hagemeijer 2011 for the sociohistorical context. Angolar, for example, has the characteristic two-tone contact prosodic system based on stress-to-tone mapping, albeit with many idiosyncrasies (see §4.3 below). As expected, the H vs. L contrast is based on the stress contrast of Portuguese, for example, kalôlô /LHL/ ‘heat’ from Port. caˈlor, and kotho /LH/ ‘heart’ from Port. coraˈção (examples from Maurer 1995). Afro-European creole languages with Indo-European lexifiers and African substrates of the Atlantic-Congo family are also spoken on the other side of the Atlantic. Research on the origins of Africans deported to the Caribbean during the European slave trade has established that the vast majority came from exclusively tonal ecologies (for Suriname, for example, see Arends 1995, Migge 1998). During the height of European slavery regimes, Africans starkly outnumbered Europeans. Much of the colonial Caribbean therefore also featured a type (i) constellation characterized by intonation-only superstrates, tonal substrates, and SL agentivity in tone languages (see §2.1). Caribbean creoles that were isolated from the influence of European languages from the early colonial period until recently still have two-tone systems that show the unmistakable signs of stress-to-tone mapping. Saramaccan, a creole spoken by the descendants of Africans who established independent polities in the Amazonian rainforests of Suriname in the seventeenth century, has a prosodic system with two tones and numerous tonal processes (Rountree 1972, Good 2004). Portugueseand English-sourced words have an H tone on the syllable that bears stress in the lexifier; compare wómì /HL/ ‘man’ (< Port. ˈhomem ‘man’) and àkí /LH/ ‘here’ (< Port. aˈqui ‘here’) (examples from Good 2009). The same holds for Saramaccan’s sister language Ndyuka (see §4.3). Less-isolated Caribbean creoles like Jamaican Creole and Sranan (coastal Suriname) have intonation-only systems (Gooden 2003, Barth 2016). However, the sociohistorical context suggests that a shift from tone to intonation-only systems took place earlier in these Caribbean creoles. The end of the European slave trade in the nineteenth century gradually led to declining proportions of native speakers of African tone languages in the Caribbean. In the twentieth century, immigration from Asia, urbanization, and the expansion of education led to further recompositions of the linguistic ecologies of the Caribbean (for an overview of the Guyanas, see Yakpo & Muysken 2017). More porous social boundaries led to increasing opportunities for the non-European majority of the Caribbean colonies to acquire the European colonial languages that served as standards. In the process, numerous Caribbean creoles may have acquired intonation-only systems from the colonial superstrates. Even so, most Caribbean creoles retain features that hark back to a tonal past. Barth (2016) argues on the grounds of historical phonology that Sranan once had a tone system. Today, Sranan, like numerous other Caribbean intonation-only creoles, manifests ‘residual tone’ (Berry 1972) in a few grammatical functions and semantic fields. Not by chance, residual tone is found in parts of the linguistic system with a high density of Africanisms. For instance, Sranan ideophones bear fixed (series of) H or L tone(s) that cannot be altered. Many of these ideophones can be traced directly to African etymons. Compare pétépété /HHHH/ ‘thoroughly’ (< Akan/Ewe pétépété ‘thoroughly’) and tjùbùm /LL/ ‘with a plop sound’ (Smith & Adamson 2006:216). Other word classes in Sranan feature stress with characteristic effects like the shortening of unstressed syllables, often leading to consonant gemination, for example, paˈpa → pˈpa ‘father’ (van der Hilst 1988:51–54). Contact of tone and intonation-only systems has also engendered hybrid prosodic systems with the acoustic and phonological properties of tone and stress. The Iberian-lexi24 LANGUAGE, VOLUME 96, NUMBER 1 (2020) fier creole Papiamentu, spoken in the Netherlands Antilles, combines stress from its European heritage with tone (and various tonal processes) from its African lineage (RiveraCastillo & Pickering 2004).10The vowels of stressed syllables are longer and louder than unstressed ones, and unstressed syllables tend to be more centralized (i.e. more schwalike) than stressed ones (Remijsen & van Heuven 2005:223). At the same time, syllables marked for H and for L differ consistently in pitch in a way similar to that established for CAF and EGS (see §3.5). Although much still remains obscure about the origins of Papiamentu (see Martinus 1996), its hybrid system seems to have taken shape in a small ecology with a more balanced demographic ratio of Africans to Europeans, a prevalence of domestic instead of plantation slavery, significant exposure to the European lexifier by Africans, and to the creole by Europeans, and widespread multilingualism (see Jacobs 2012). The ecology therefore appears to have had a good measure of simultaneous RL and SL agentivity in tone and intonation-only languages when Papiamentu emerged (see also §2.1). The role of perception in stress-to-tone mapping. Stress-to-tone mapping is the crucial mechanism through which the prosodic systems of Afro-European contact varieties and languages took shape in type (i) contact constellations featuring SL agentivity in tonal adstrates and substrates (see §2.2). Stress-to-tone mapping originates in perceptual analogies between stress and H tone. These analogies lead speakers of tone languages to interpret the stress vs. nonstress contrast in the intonation-only superstrate or lexifier as a tonal H vs. L contrast in the contact variety or language. The production of both H tone and stress involves activation of the cricothyroid muscle and therefore demands more articulatory effort than that of L tones and absence of stress (Yip 2002:63). The articulatory effort involved in producing stress tends to entail raised rather than lowered pitch. Consequently, high pitch is a frequent correlate of stress in intonation-only systems. This is the case, for instance, with SEF, the superstrate of CAF (see §3.2) and with (British) English, the superstrate of African Englishes and the lexifier of the Afro-European English-lexifier Creoles (Fry 1958, Lieberman 1960). Speakers of tone languages are perceptually also more sensitive to pitch variations than to other acoustic cues of stress. Speakers of Mandarin Chinese selectively perceive the higher pitch of stressed syllables in English, rather than vowel length or loudness, thus making pitch the primary cue for distinguishing stressed from unstressed syllables (Wang 2008). Two influential models of the nonnative acquisition of phonology, the perceptual assimilation model for second language learners (PAM-L2; Best & Tyler 2007, So & Best 2010) and the speech learning model (SLM; Flege 1995), provide theoretical and empirical support for the role of perception in stress-to-tone mapping. Both models postulate that the perception of a nonnative phonetic input is linked to the phonology of a learner’s native language(s). The PAM-L2 proposes that additional language learners will assimilate nonnative phonemes to the native phonemes that are perceptually the most similar. For example, Best (1994:199) shows that English speakers associate the Zulu voiceless lateral fricative /ɬ/ with any of the voiceless coronal fricatives /s/, /ʃ/, or /θ/. The acoustic overlaps (i.e. coronality, friction, and voicelessness) engender sufficient perceptual similarities for English speakers to render them with a corresponding native phoneme. Stress vs. nonstress is also such a nonassimilable conOn the outcomes of prosodic contact 25 10 According to Childs (1995), mixed prosodic systems that make use of lexical tone and lexical stress are also found in the Atlantic subfamily (spoken mostly between Senegal and Liberia) of the Atlantic-Congo family (see also §2.2). (pitch) accent systems (e.g. Goldsmith 1987). Pulleyblank (1986:158) deconstructs the notion of pitch accent early on by showing that ‘pure’ tone languages also show numerous limitations in the number of tones that can be assigned to a prosodic unit. Positional restrictions on the placement of tones in African languages have been documented more recently by Downing (2010). Restrictions are apparently so common in tone languages that van der Hulst is prompted to ‘entertain the hypothesis that perhaps all languages are accentual’ (2010:9). We, in contrast, maintain along with others that the opposite holds, and that languages like CAF and EGS, in which pitch enters into the composition of morphemes, are tone languages (cf. Gussenhoven 2004, Hyman 2006). This study was explicitly limited to type (i) contact constellations (INSUP + TOSUB/ADS = TOCON). We nevertheless also assume the opposite to be the case in type (ii) constellations (TOSUP + INSUB/ADS = INCON). In ecologies featuring SL agentivity in intonation-only adstrates, a tonal superstrate is likely to undergo the inverse mechanism of tone-to-stress conversion. This leads us back to the controversial claims cited in the introductory paragraph of this article. Maintaining, as does McWhorter (1998), that creolization leads to the abandonment of tone is counterintuitive against the backdrop of the extensive imposition of African substrate structures on creoles in categories that are less stable under SL agentivity than prosody (e.g. in the tense-mood-aspect system (see Winford & Migge 2007) and in the grammar of space (see Yakpo 2017)). Such a claim also has no empirical basis, as shown in §4. In a similar vein, Trudgill’s (2010) suggestion that tone disappears during contact is not raised independently of the issue of simplification. Since tone is so stable in natively acquired languages, even morphologically isolating contact languages, like many of the creoles covered, will therefore retain tone if they emerge under SL agentivity in tone languages. Further, Salmons (1992:45) does not differentiate between the tonal and nontonal baselines of SL agentivity when he proposes the ‘vulnerability of tone to contact’ in arguing for his tone-to-stress cycle. With evidence involving SL agentivity in nontonal languages he suggests tone loss as a general tendency. Among the examples given is the shift from tone to stress in Norwegian and Swedish contact varieties occasioned by substrate and adstrate transfer from nontonal Sami and Finish (see §2.2). In ecol - ogies where native languages with intonation-only systems dominate, tone is, of course, not likely to be selected into an emerging contact variety. The persistence of tone in contact languages that arose under SL agentivity in tone languages is, in turn, framed as unusual by Salmons, although it merely represents the other side of the coin of tone loss (1992:34–35, with respect to the presence of tone in Saramaccan, Krio, and Lingala). On a more general note, it is remarkable that languages with differing genealogical affiliations and typological profiles, such as the colonial varieties CAF and EGS, Afro-European creoles, and African adstrates, all end up with similar prosodic systems when spoken in the same tone-dominant ecology. Contact varieties like CAF and EGS, as well as the Afro-European creoles, therefore partake just like other languages in the convergence processes that lead to the worldwide areal clustering of prosodic systems. The presence of some intonation-only features in these languages is charac - teristic of areal ‘buffer zones’ (cf. Stilo 2005), in which tone and intonation-only systems collide. Once demographic and social dominance relations, the stratal composition, and the direction of transfer and agentivity in an ecology have been taken into account, the outcomes of prosodic contact are no longer typologically as unusual as they may first seem. 32 LANGUAGE, VOLUME 96, NUMBER 1 (2020) Appendix: Speaker characteristics11 On the outcomes of prosodic contact 33 11 Speakers are ordered by the number of languages spoken. SP age/gend languages occupation education 1 28/FCAF, Sango Civil servant University 2 28/M CAF, Sango Student University 3 39/F CAF, Sango Secretary Secondary school 4 59/M CAF, Sango, Banda Electrician Middle school 5 58/M CAF, Sango, Goula Trader Secondary school 6 32/M CAF, Sango, Langba Civil servant Secondary school 7 44/F CAF, Sango, Yakoma Civil servant Secondary school 8 28/F CAF, Sango, Yakoma, Zande Student University 9 41/F CAF, Sango, Kaba, Banziri Secretary Secondary school 10 33/M CAF, Sango, Kaba, Ngbaka Civil servant Secondary school 11 9/M EGS, Pichi, Bube Student Primary school 12 18/F EGS, Pichi, Bube Student Secondary school 13 19/M EGS, Pichi, Bube Student Primary school 14 83/F EGS, Pichi, Bube Grandmother Primary school 15 34/M EGS, Pichi, Fang Mechanic Middle school 16 42/F EGS, Pichi, Fang Employee Secondary school 17 54/M EGS, Pichi, Fang, French (Cameroon) Civil servant University 18 58/F EGS, Pichi, Fang, Kombe, French (Gabon) Seamstress Middle school central african republicequatorial guinea REFERENCES Ansaldo, Umberto. 2009. Contact languages: Ecology and evolution in Asia. Cambridge: Cambridge University Press. Arends, Jacques. 1995. Demographic factors in the formation of Sranan. The early stages of creolization (Creole language library 13), ed. by Jacques Arends, 233–85. Amsterdam: John Benjamins. Avanzi, Mathieu; Guri Bordal; and Gélase Nimbona. 2014. The obligatory contour principle in African and European varieties of French. INTERSPEECH 2014, 1312–16. Online: https://www.isca-speech.org/archive/interspeech_2014/i14_1312.html. Barth, Emily. 2016. Tone loss in Sranan Creole: Re-thinking contact driven change. Papers in Historical Phonology 1.76–89. DOI: 10.2218/pihph.1.2016.1694. Berry, Jack. 1970. A note on the prosodic structure of Krio. International Journal of American Linguistics 36.266–67. DOI: 10.1086/465120. Berry, Jack. 1972. Some observations on ‘residual tone’ in West Indian English. Paper presented at the Conference on Creole Languages and Educational Development, University of the West Indies at St. Augustine, Trinidad, 24–28 July. Best, Catherine T. 1994. The emergence of native-language phonological influences in infants: A perceptual assimilation model. The development of speech perception: The transition from speech sounds to spoken words, ed. by Judith C. Goodman and Howard C. Nusbaum, 167–224. Cambridge, MA: MIT Press. Best, Catherine T., and Michael D. Tyler. 2007. Nonnative and second-language speech perception: Commonalities and complementarities. Language experience in second language speech learning: In honor of James Emil Flege, ed. by Ocke-Schwen Bohn and Murray J. Munro, 13–34. Amsterdam: John Benjamins. Boersma, Paul, and David Weenink. 2018. Praat: Doing phonetics by computer. Version 6.0.41. Online: http://www.praat.org. Bolekia Boleká, Justo. 2007. La enculturación bubi desde los préstamos léxicos del pidgin-english: Procesos de lexicalización progresiva. Salamanca: University of Salamanca dissertation. Bordal, Guri. 2012a. A phonological study of French spoken by multilingual speakers from Bangui, the capital of the Central African Republic. Phonological variation in French: Illustrations from three continents (Studies in language variation 11), ed. by Randall Gess, Chantal Lyche, and Trudel Meisenburg, 23–43. Amsterdam: John Benjamins. Bordal, Guri. 2012b. Prosodie et contact de langues : Le cas du système tonal du français centrafricain. Oslo: University of Oslo, and Paris: Université Paris Ouest Nanterre la Défense dissertation. Bordal, Guri. 2013. Le français centrafricain : Un français à tons lexicaux. Revue française de linguistique appliquée 18.91–102. DOI: 10.3917/rfla.182.0091. Bordal, Guri. 2015. Traces of the lexical tone system of Sango in Central African French. Prosody and language in contact, ed. by Elisabeth Delais-Roussarie, Mathieu Avanzi, and Sophie Herment, 29–49. Heidelberg: Springer. DOI: 10.1007/978-3-662-45168 -7_3. Bordal, Guri, and Ingse Skattum. 2014. La prosodie du français en Afrique—traits panafricains ou traits de la langue première ? Le cas de locuteurs natifs de quatre langues : Sango, bambara, wolof et tamasheq. La phonologie du français : Normes, périphéries, modélisation : Mélanges pour Chantal Lyche, ed. by Jacques Durand, Gjert Kristoffersen, Bernard Laks, and Julie Peuvergne, 119–52. Paris: Presses universitaires de Paris Ouest. Bordal Steien, Guri; Akissi Béatrice Boutin; and Robert Beyom. 2016. French in the Central African Republic: A speaker from Bangui. Varieties of spoken French, ed. by Sylvain Detey, Jacques Durand, Bernard Laks, and Chantal Lyche, 236–46. Oxford: Oxford University Press. Boutin, Béatrice Akissi, and Gabor Turscan. 2009. La prononciation du français en Afrique : La Côte d’Ivoire. Phonologie, variation et accents du français, ed. by Jacques Durand, Bernard Laks, and Chantal Lyche, 131–51. Paris: Hermès. Boyeldieu, Pascal. 1975. Études yakoma, langue du groupe oubanguien (RCA) : Morpho - logie—synthématique. Paris: Société d’Études Linguistiques et Anthropologiques de France (SELAF). Bruce, Gösta. 2004. An intonational typology of Swedish. Speech Prosody 2004, Nara, 175–78. Online: https://www.isca-speech.org/archive/sp2004/sp04_175.html. Bull, Tove. 1995. Language maintenance and loss in an originally trilingual area in North Norway. International Journal of the Sociology of Language 115.125–34. DOI: 10 .1515/ijsl.1995.115.125. Castillo Rodríguez, Susana. 2013. Language and the hispanization of Equatorial Guin - ea. A political history of Spanish: The making of a language, ed. by José Del Valle, 350–63. Cambridge: Cambridge University Press. DOI: 10.1017/CBO9780511794339 .029. Childs, George Tucker. 1995. Tone and accent in Atlantic: An evolutionary perspective. The complete linguist: Papers in memory of Patrick J. Dickens (Afrikanische Sprachen und Kulturen), ed. by Anthony Traill, Rainer Vossen, and Megan Biesele, 195–215. Cologne: Rüdiger Köppe. Clements, George N., and Yetunde Laniran. 2007. Downstep and high tone raising: Interacting factors in Yoruba tone production. Paris: Laboratoire de Phonétique et Pho - nologie (LPP) UMR 7018, ms. Online: https://halshs.archives-ouvertes.fr/halshs-00180 368. Clements, George N., and Annie Rialland. 2008. Africa as a phonological area. A linguistic geography of Africa, ed. by Bernd Heine and Derek Nurse, 36–85. Cambridge: Cambridge University Press. DOI: 10.1017/CBO9780511486272.004. Courtenay, Karen R. 1974. On the nature of the Bambara tone system. Studies in African Linguistics 5.303–23. Criper, Lindsay. 1971. A classification of types of English in Ghana. Journal of African Languages 10.6–17. Criper-Friedman, Lindsay. 1990. The tone system of West African Coastal English. World Englishes 9.63–77. DOI: 10.1111/j.1467-971X.1990.tb00687.x. de Blij, Harm. 2009. The power of place: Geography, destiny, and globalization’s rough landscape. Oxford: Oxford University Press. de Lacy, Paul. 2002. The interaction of tone and stress in optimality theory. Phonology 19.1–32. DOI: 10.1017/S0952675702004220. Delais-Roussarie, Elisabeth; Brechtje Post; Mathieu Avanzi; Caroline Buthke; Albert Di Cristo; Ingo Feldhausen; Sun-Ah Jun; Philippe Martin; Trudel Meisenburg; Annie Rialland; Rafèu Sichel-Bazin; and Hi-Yon Yoo. 2015. Into34 LANGUAGE, VOLUME 96, NUMBER 1 (2020) national phonology of French: Developing a ToBI system for French. Intonation in Romance, ed. by Sónia Frota and Pilar Prieto, 63–100. Oxford: Oxford University Press. DOI: 10.1093/acprof:oso/9780199685332.003.0003. Devonish, Hubert. 2001. On the Sierra Leone–Caribbean connection: Hot on the trail of ‘tone-shifted items’ in Anglo-West African. Due respect: Essays on English and English-related Creoles in the Caribbean in honour of Professor Robert Le Page, ed. by Pauline Christie, 184–205. Kingston: University of the West Indies Press. Diki-Kidiri, Marcel. 1977. Le sango s’écrit aussi … : Esquisse linguistique du sango, langue nationale de l’Empire Centrafricain. (Langues et civilisations à tradition orale 24.) Paris: Société d’Études Linguistiques et Anthropologiques de France (SELAF). Diki-Kidiri, Marcel, and Jean-Pierre Caprile. 1982. Contacts de langues et contacts de cultures 4 : L’expansion des langues africaines : Peul, sango, kikongo, ciluba, swahili. (LACITO documents Afrique 8.) Paris: Société d’Études Linguistiques et Anthropologiques de France (SELAF). Downing, Laura J. 2010. Accent in African languages. A survey of word accentual patterns in the languages of the world, ed. by Harry van der Hulst, Rob Goedemans, and Ellen van Zanten, 381–428. Berlin: De Gruyter Mouton. DOI: 10.1515/97831101989 66.1.381. Downing, Laura J. 2015. Convergence of prosody under contact: Two African case studies. Paper presented at HALS Workshop 4, Helsinki University. Downing, Laura J., and Annie Rialland (eds.) 2016a. Intonation in African tone languages. (Phonology and phonetics 24.) Berlin: De Gruyter Mouton. DOI: 10.1515/978 3110503524. Downing, Laura J., and Annie Rialland. 2016b. Introduction. In Downing & Rialland 2016a, 1–16. DOI: 10.1515/9783110503524-001. Faraclas, Nicholas G. 1984. A grammar of Obolo. Bloomington: Indiana University Linguistics Club. Faraclas, Nicholas G. 1996. Nigerian Pidgin. London: Routledge. Ferraz, Luiz Ivens. 1979. The creole of São Tomé. Johannesburg: Witwatersrand University Press. Finney, Malcolm Awadajin. 2004. Tone assignment on lexical items of English and African origin in Krio. Creoles, contact, and language change: Linguistic and social implications, ed. by Geneviève Escure and Armin Schwegler, 221–36. Amsterdam: John Benjamins. Flege, James Emil. 1987. The production of ‘new’ and ‘similar’ phones in a foreign language: Evidence for the effect of equivalence classification. Journal of Phonetics 15. 47–65. DOI: 10.1016/S0095-4470(19)30537-6. Flege, James Emil. 1995. Second language speech learning: Theory, findings, and problems. Speech perception and linguistic experience: Issues in cross-language research, ed. by Winifred Strange, 233–77. Timonium, MD: York. François, Alexandre. 2012. The dynamics of linguistic diversity: Egalitarian multilingualism and power imbalance among northern Vanuatu languages. International Journal of the Sociology of Language 2012.85–110. DOI: 10.1515/ijsl-2012-0022. Fry, Dennis B. 1958. Experiments in the perception of stress. Language and Speech 1.126– 52. DOI: 10.1177/002383095800100207. Fyle, Clifford N., and Eldred Durosimi Jones. 1980. A Krio-English dictionary. Oxford: Oxford University Press. Goldman, Jean-Philippe. 2011. EasyAlign: An automatic phonetic alignment tool under Praat. INTERSPEECH 2011, 3233–36. Online: https://www.isca-speech.org/archive /interspeech_2011/i11_3233.html. Goldsmith, John. 1987. Tone and accent, and getting the two together. Berkeley Linguistics Society 13.88–104. DOI: 10.3765/bls.v13i0.1827. Good, Jeff. 2004. Tone and accent in Saramaccan: Charting a deep split in the phonology of a language. Lingua 114.575–619. DOI: 10.1016/S0024-3841(03)00062-7. Good, Jeff. 2009. A twice-mixed creole? Tracing the history of a prosodic split in the Saramaccan lexicon. Studies in Language (Special issue: Language change in contact languages: Grammatical and prosodic considerations, ed. by J. Clancy Clements and Shelome Gooden) 33.459–98. DOI: 10.1075/sl.33.2.09goo. On the outcomes of prosodic contact 35 Gooden, Shelome. 2003. The phonology and phonetics of Jamaican Creole reduplication. Columbus: The Ohio State University dissertation. Online: http://rave.ohiolink.edu /etdc/view?acc_num=osu1070485686. Güldemann, Tom. 2008. The Macro-Sudan belt: Towards identifying a linguistic area in northern sub-Saharan Africa. A linguistic geography of Africa, ed. by Bernd Heine and Derek Nurse, 151–85. Cambridge: Cambridge University Press. DOI: 10.1017/CBO97 80511486272.006. Gussenhoven, Carlos. 2004. The phonology of tone and intonation. Cambridge: Cambridge University Press. Gussenhoven, Carlos. 2006. Between stress and tone in Nubi word prosody. Phonology 23.192–223. DOI: 10.1017/S0952675706000881. Gussenhoven, Carlos. 2017. On the intonation of tonal varieties of English. The Oxford handbook of World Englishes, ed. by Markku Filppula, Juhani Klemola, and Devyani Sharma, 569–98. Oxford: Oxford University Press. DOI: 10.1093/oxfordhb/97801997 77716.013.29. Gussenhoven, Carlos, and Inyang Udofot. 2010. Word melodies vs. pitch accents: A perceptual evaluation of terracing contours in British and Nigerian English. Speech Prosody 2010, Chicago, paper 015. Online: https://www.isca-speech.org/archive /sp2010/sp10_015.html. Hagemeijer, Tjerk. 2011. The Gulf of Guinea creoles: Genetic and typological relations. Journal of Pidgin and Creole Languages 26.111–54. DOI: 10.1075/jpcl.26.1.05hag. Hammarström, Harald; Robert Forkel; and Martin Haspelmath (eds.) 2017. Glottolog 3.0. Jena: Max Planck Institute for the Science of Human History. Online: http:// glottolog.org. Hancock, Ian F. 1971. A study of the sources and development of the lexicon of Sierra Leone Krio. London: SOAS University of London dissertation. Online: https://eprints .soas.ac.uk/id/eprint/28465. Hancock, Ian F. 1986. The domestic hypothesis, diffusion and componentiality: An account of Atlantic Anglophone creole origins. Substrata versus universals in Creole genesis, ed. by Pieter Muysken and Norval Smith, 71–102. Amsterdam: John Benjamins. Hao, Yen-Chen. 2012. Second language acquisition of Mandarin Chinese tones by tonal and non-tonal language speakers. Journal of Phonetics 40.269–79. DOI: 10.1016/j .wocn.2011.11.001. Haugen, Einar. 1972. The analysis of linguistic borrowing. The ecology of language: Essays by Einar Haugen, ed. by Anwar S. Dil, 79–109. Stanford, CA: Stanford University Press. Heine, Bernd. 1978. Some generalizations on African-based pidgin languages. Afrika und Übersee 61.219–29. Hock, Hans Henrich. 1986. Principles of historical linguistics. Berlin: Mouton de Gruyter. Hockett, Charles F. 1967. Where the tongue slips, there slip I. To honor Roman Jakobson: Essays on the occasion of his seventieth birthday, 910–36. The Hague: Mouton. Hualde, José Ignacio. 2012. Stress and rhythm. The handbook of Hispanic linguistics, ed. by José Ignacio Hualde, Antxon Olarrea, and Erin O’Rourke, 153–71. Malden, MA: Wiley-Blackwell. DOI: 10.1002/9781118228098.ch8. Hualde, José Ignacio, and Pilar Prieto. 2015. Intonational variation in Spanish: European and American varieties. Intonation in Romance, ed. by Sónia Frota and Pilar Prieto, 350–91. Oxford: Oxford University Press. DOI: 10.1093/acprof:oso/9780199685 332.003.0010. Huber, Magnus. 2013. Ghanaian Pidgin English. The survey of pidgin and creole languages, vol. 1: English-based and Dutch-based languages, ed. by Susanne Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber, 167–75. Oxford: Oxford University Press. Online: https://apics-online.info/surveys/16. Huttar, George L., and Mary L. Huttar. 1994. Ndyuka. London: Routledge. Hyman, Larry M. 2006. Word-prosodic typology. Phonology 23.225–57. DOI: 10.1017 /S0952675706000893. Hyman, Larry M. 2009. How (not) to do phonological typology: The case of pitch-accent. Language Sciences 31.213–38. DOI: 10.1016/j.langsci.2008.12.007. 36 LANGUAGE, VOLUME 96, NUMBER 1 (2020) Hyman, Larry M. 2018. What tone teaches us about language. Language 94.698–709. DOI: 10.1353/lan.2018.0040. Hyman, Larry M.; Peter Jenks; Geoffrey Bacon; Nicolas Baier; Emily Clem; Matthew Faytak; Spencer Lamoureux; Florian Lionnet; John Merril; and Hannah Sande. 2015. Areal features and linguistic reconstruction in Africa. Paper presented at the 8th World Congress of African Linguistics, Kyoto University, 20–24 August. Ingham, Jacqueline. 2015. Perception, production and perceptual learning in the second language: A study of perceptual learning by L1 Bengali speakers of L2 English. Papers from the Lancaster Postgraduate Conference in Linguistics and Language Teaching (LAEL PG) 9.49–69. Jacobs, Bart. 2012. Embedding Papiamentu in the mixed language debate. Journal of Historical Linguistics 2.52–82. DOI: 10.1075/jhl.2.1.05jac. Jahr, Ernst Håkon. 1984. Language contact in Northern Norway: Adstratum and substratum in the Norwegian, Sami and Finnish of Northern Norway. Acta Borealia 1.103–12. DOI: 10.1080/08003838408580308. Jun, Sun-Ah, and Cécile Fougeron. 2000. A phonological model of French intonation. Intonation: Analysis, modeling and technology, ed. by Antonis Botinis, 209–42. Dordrecht: Springer. DOI: 10.1007/978-94-011-4317-2_10. Jun, Sun-Ah, and Cécile Fougeron. 2002. Realizations of accentual phrase in French intonation. Probus 14.147–72. DOI: 10.1515/prbs.2002.002. Koffi, Ettien N. 2018a. Differential analysis of lexical pitch in accent and tone languages. Linguistic Portfolios 7:8. Online: https://repository.stcloudstate.edu/stcloud_ling/vol7 /iss1/8. Koffi, Ettien N. 2018b. Guide d’orthographe et de lecture de l’agni morofou. Bongouanou: Centre d’Alphabétisation et de Traduction en Agni (CATA). Kwok, Veronica P. Y.; Guo Dan; Kofi Yakpo; Stephen Matthews; Peter T. Fox; Ping Li; and Li-Hai Tan. 2017. A meta-analytic study of the neural systems for auditory processing of lexical tones. Frontiers in Human Neuroscience 11:375. DOI: 10.3389 /fnhum.2017.00375. Ladd, D. Robert. 2008. Intonational phonology. Cambridge: Cambridge University Press. Lang, Jürgen. 2009. Les langues des autres dans la créolisation : Théorie et exemplification par le créole d’empreinte wolof à l’île Santiago du Cap Vert. Tübingen: Gunter Narr. Leben, William Ronald. 1973. Suprasegmental phonology. Cambridge, MA: MIT dissertation. Online: http://www.ai.mit.edu/projects/dm/theses/leben73.pdf. Lieberman, Philip. 1960. Some acoustic correlates of word stress in American English. The Journal of the Acoustical Society of America 32.451–54. DOI: 10.1121/1.1908095. Lim, Lisa, and Umberto Ansaldo. 2015. Languages in contact. Cambridge: Cambridge University Press. Lipski, John M. 2015. ‘Toned-up’ Spanish: Stress → pitch → tone(?) in Equatorial Guinea. Romance linguistics 2013: Selected papers from the 43rd Linguistic Symposium on Romance Languages (LSRL), New York, 17–19 April, 2013, ed. by Christina Tortora, Marcel den Dikken, Ignacio L. Montoya, and Teresa O’Neill, 233–56. Amsterdam: John Benjamins. Lüpke, Friederike. 2016. Uncovering small-scale multilingualism. Critical Multilingualism Studies 4.35–74. Online: https://cms.arizona.edu/ojs3/multilingual/article/view /100. Martín del Molino, Amador. 1993. La ciudad de Clarence: Primeros años de la actual ciudad de Malabo, capital de Guinea Ecuatorial, 1827–1859. Malabo: Ediciones Centro Cultural Hispano-Guineano. Martinus, Efraim Frank. 1996. The kiss of a slave: Papiamentu’s West-African connections. Amsterdam: University of Amsterdam dissertation. Matisoff, James A. 2001. Genetic versus contact relationship: Prosodic diffusibility in South-East Asian languages. Areal diffusion and genetic inheritance: Problems in comparative linguistics, ed. by Alexandra Y. Aikhenvald and R. M. W. Dixon, 291–327. Oxford: Oxford University Press. Matras, Yaron. 2009. Language contact. Cambridge: Cambridge University Press. On the outcomes of prosodic contact 37 Maurer, Philippe. 1995. L’angolar : Un créole afro-portugais parlé à São Tomé : Notes de grammaire, textes, vocabulaires. Hamburg: Buske. Maurer, Philippe. 2008. A first step towards the analysis of tone in Santomense. Roots of Creole structures: Weighing the contribution of substrates and superstrates (Creole language library 33), ed. by Susanne Maria Michaelis, 253–61. Amsterdam: John Benjamins. Maurer, Philippe. 2009. Principense (Lung’le): Grammar, texts, and vocabulary of the Afro-Portuguese creole of the Island of Príncipe, Gulf of Guinea. London: Battlebridge. McWhorter, John H. 1998. Identifying the creole prototype: Vindicating a typological class. Language 74.788–818. DOI: 10.2307/417003. Mertens, Piet. 2004. Le prosogramme : Une transcription semi-automatique de la proso - die. Cahiers de l’Institut de Linguistique de Louvain 30.7–25. DOI: 10.2143/CILL.30.1 .519212. Migge, Bettina. 1998. Substrate influence in the formation of the Surinamese Plantation Creole: A consideration of the sociohistorical data and linguistic data from Ndyuka and Gbe. Columbus: The Ohio State University dissertation. Online: http://rave .ohiolink.edu/etdc/view?acc_num=osu1487953567769312. Morgades Besari, Trinidad. 2011. Los criollos (Fernandinos-Kriös) de Guinea Ecuatorial. La Gaceta de Guinea Ecuatorial 162, Miscelánea. Mufwene, Salikoko S. 1996. The founder principle in creole genesis. Diachronica 13.83–134. DOI: 10.1075/dia.13.1.05muf. Mufwene, Salikoko S. 2001. The ecology of language evolution. Cambridge: Cambridge University Press. Newman, Paul. 1973. Grades, vowel-tone classes and extensions in the Hausa verbal system. Studies in African Linguistics 4.297–346. Nimbona, Gélase. 2014. Étude contrastive de la prosodie du kirundi et du français : Analyse des transferts prosodiques du kirundi au français parlé au Burundi. Louvain: Université catholique de Louvain. Online: https://dial.uclouvain.be/pr/boreal/object /boreal:152240. Nimbona, Gélase, and Guri Bordal Steien. 2019. Modes monolingues dans des écologies multilingues : Les études phonologiques des français africains. Langue française (Special issue: Les français d’Afrique. En Afrique. Hors d’Afrique, ed. by Guri Bordal Steien and Cécile Van den Avenne) 202.1–26. Nkengasong, Nkemngong. 2016. A grammar of Cameroonian Pidgin. Newcastle upon Tyne: Cambridge Scholars. Nylander, Dudley K. 1984. The relation between the middle tone and ‘empty category principle’ violations in Krio. Studies in African Linguistics 15.163–75. Parkvall, Mikael. 2000. Out of Africa: African influences in Atlantic creoles. London: Battlebridge. Pasch, Helma. 1993. Phonological similarities between Sango and its base language: Is Sango a pidgin/creole or a koiné? Topics in African linguistics: Papers from the XXI Annual Conference on African Linguistics, University of Georgia, April 1990, ed. by Salikoko S. Mufwene and Lioba Moshi, 279–92. Amsterdam: John Benjamins. Pasch, Helma. 1997. Sango. Contact languages: A wider perspective (Creole language library 17), ed. by Sarah G. Thomason, 209–70. Amsterdam: John Benjamins. Pierrehumbert, Janet Breckenridge. 1980. The phonology and phonetics of English intonation. Cambridge, MA: MIT dissertation. Online: http://dspace.mit.edu/handle/1721 .1/16065. Post, Marike. 2013. Fa d’Ambô. The survey of pidgin and creole languages, vol. 2: Portuguese-based, Spanish-based, and French-based languages, ed. by Susanne Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber, 81–89. Oxford: Oxford University Press. Online: https://apics-online.info/surveys/38. Pulleyblank, Douglas. 1986. Tone in lexical phonology. Dordrecht: Reidel. Quilis, Antonio. 1978. Métrica española. Madrid: Ediciones Alcalá. R Core Team. 2018. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. Online: https://www.R-project.org/. Remijsen, Bert, and Vincent J. van Heuven. 2005. Stress, tone and discourse prominence in the Curaçao dialect of Papiamentu. Phonology 22.205–35. DOI: 10.1017/S095 2675705000540. 38 LANGUAGE, VOLUME 96, NUMBER 1 (2020) Rivera-Castillo, Yolanda, and Lucy Pickering. 2004. Phonetic correlates of stress and tone in a mixed system. Journal of Pidgin and Creole Languages 19.261–84. DOI: 10.1075/jpcl.19.2.02riv. Roca, Iggy. 2005. Saturation of parameter settings in Spanish stress. Phonology 22.345–94. DOI: 10.1017/S0952675705000655. Rountree, S. Catherine. 1972. Saramaccan tone in relation to intonation and grammar. Lingua 29.308–25. DOI: 10.1016/0024-3841(72)90028-9. Salmons, Joe. 1992. Accentual change and language contact: Comparative survey and a case study of early Northern Europe. Stanford, CA: Stanford University Press. Samarin, William J. 2000. The status of Sango in fact and fiction: On the one-hundredth anniversary of its conception. Language change and language contact in pidgins and creoles, ed. by John McWhorter, 301–34. Amsterdam: John Benjamins. Samarin, William J. 2013. Sango. The survey of pidgin and creole languages, vol. 3: Contact languages based on languages from Africa, Australia, and the Americas, ed. by Susanne Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber, 13–24. Oxford: Oxford University Press. Online: https://apics-online.info/contributions/59. Schneider, Edgar W. 2003. The dynamics of New Englishes: From identity construction to dialect birth. Language 79.233–81. DOI: 10.1353/lan.2003.0136. Smith, Norval. 2017. Krio as the Western Maroon Creole language of Jamaica, and the /na/ isogloss. Language contact in Africa and the African diaspora in the Americas: In honor of John V. Singler (Creole language library 53), ed. by Cecelia Cutler, Zvjezdana Vrzić, and Philipp Angermeyer, 251–74. Amsterdam: John Benjamins. Smith, Norval, and Lilian Adamson. 2006. Tonal phenomena in Sranan. STUF—Language Typology and Universals 59.211–18. DOI: 10.1524/stuf.2006.59.2.211. So, Connie K., and Catherine T. Best. 2010. Cross-language perception of non-native tonal contrasts: Effects of native phonological and phonetic influences. Language and Speech 53.273–93. DOI: 10.1177/0023830909357156. Stilo, Donald L. 2005. Iranian as a buffer zone between the universal typologies of Turkic and Semitic. Linguistic convergence and areal diffusion: Case studies from Iranian, Semitic and Turkic, ed. by Éva Ágnes Csató, Bo Isaksson, and Carina Jahani, 35–63. London: Routledge Curzon. Thomason, Sarah Grey, and Terrence Kaufman. 1988. Language contact, creolization, and genetic linguistics. Berkeley: University of California Press. Thornell, Christina. 1997. The Sango language and its lexicon (Sêndâ-yângâ tî sängö). Lund: Lund University Press. Traill, Anthony, and Luiz Ivens Ferraz. 1981. The interpretation of tone in Principense Creole: Abstract phonological analysis. Studies in African Linguistics 12.205–15. Trudgill, Peter. 2010. Contact and sociolinguistic typology. The handbook of language contact, ed. by Raymond Hickey, 299–319. Malden, MA: Wiley-Blackwell. DOI: 10 .1002/9781444318159.ch15. Van Coetsem, Frans. 2000. A general and unified theory of the transmission process in language contact. Heidelberg: Universitätsverlag Winter. Van de Velde, Mark L. O. 2008. A grammar of Eton. Berlin: Mouton de Gruyter. van der Hilst, Eddy. 1988. Skrifi Sranantongo bun, leysi en bun tu [Write Sranantongo, read it well too]. Paramaribo. van der Hulst, Harry. 2010. Word accent: Terms, typologies and theories. A survey of word accentual patterns in the languages of the world, ed. by Harry van der Hulst, Rob Goedemans, and Ellen van Zanten, 3–53. Berlin: De Gruyter Mouton. DOI: 10.1515 /9783110198966.1.3. Vydrine, Valentine. 2004. Areal and genetic features in West Mande and South Mande phonology: In what sense did Mande languages evolve? Journal of West African Languages 30.113–25. Online: https://main.journalofwestafricanlanguages.org/index.php /downloads/category/101-volume-30-number-2-new. Walker, James A., and William J. Samarin. 1997. Sango phonology. Phonologies of Asia and Africa (including the Caucasus), vol. 2, ed. by Alan S. Kaye, 861–80. Winona Lake, IN: Eisenbrauns. Wang, Qian. 2008. Perception of English stress by Mandarin Chinese learners of English: An acoustic study. Victoria: University of Victoria dissertation. Online: https://dspace .library.uvic.ca//handle/1828/1282. On the outcomes of prosodic contact 39 Weinreich, Uriel. 1953. Languages in contact. New York: Linguistic Circle of New York. Welmers, William Everett. 1959. Tonemics, morphotonemics, and tonal morphemes. General Linguistics 4.1–9. Williamson, Kay. 1978. From tone to pitch-accent: The case of Ijo. Kiabàrà: Journal of the Humanities 1.116–25. Winford, Donald, and Bettina Migge. 2007. Substrate influence on the emergence of the TMA systems of the Surinamese creoles. Journal of Pidgin and Creole Languages 22.73–99. DOI: 10.1075/jpcl.22.1.06win. Witzel, Michael. 1999. Early sources for South Asian substrate languages. Mother Tongue (Special issue on South Asia), 1–70. DOI: 10.11588/xarep.00000113. Yakpo, Kofi. 2009. A grammar of Pichi. Nijmegen: Radboud University Nijmegen dissertation. Online: http://hdl.handle.net/2066/79407. Yakpo, Kofi. 2010. Gramática del Pichi. (Laboratorio de recursos orales 13.) Barcelona: Ceiba Ediciones. Yakpo, Kofi. 2012. Reiteration in Pichi: Forms, functions and areal-typological perspectives. The morphosyntax of reiteration in creole and non-creole languages (Creole language library 43), ed. by Enoch O. Aboh, Norval Smith, and Anne Zribi-Hertz, 251–84. Amsterdam: John Benjamins. Yakpo, Kofi. 2013. Wayward daughter: Language contact in the emergence of Pichi (Equatorial Guinea). Journal of African Languages and Linguistics 34.275–99. DOI: 10.1515 /jall-2013-0009. Yakpo, Kofi. 2016. ‘The only language we speak really well’: The English creoles of Equatorial Guinea and West Africa at the intersection of language ideologies and language policies. International Journal of the Sociology of Language (Special issue: Exploring glottopolitical dynamics in Africa: The Spanish colonial past and beyond, ed. by Susana Castillo-Rodríguez and Laura Morgenthaler García) 2016(239).211–33. DOI: 10 .1515/ijsl-2016-0010. Yakpo, Kofi. 2017. Unity in diversity: The homogeneity of the substrate and the grammar of space in the African and Caribbean English-lexifier Creoles. Language contact in Africa and the African diaspora in the Americas: In honor of John V. Singler (Creole language library 53), ed. by Cecelia Cutler, Zvjezdana Vrzić, and Philipp Angermeyer, 225–50. Amsterdam: John Benjamins. Yakpo, Kofi. 2018. ¿El nacimiento de una lengua afrohispana?: La influencia del español en el idioma criollo inglés de Guinea Ecuatorial. Trans-afrohispanismos: Puentes cultu - rales críticos entre África, Latinoamérica y España, ed. by Dorothy Odartey-Wellington, 243–59. Leiden: Brill. DOI: 10.1163/9789004364080_015. Yakpo, Kofi. 2019a. A grammar of Pichi. (Studies in diversity linguistics 23.) Berlin: Language Science. DOI: 10.5281/zenodo.2546450. Yakpo, Kofi. 2019b. Inheritance, contact, convergence: Pronominal allomorphy in the African English-lexifier Creoles. English World-Wide 40.202–27. DOI: 10.1075/eww .00028.yak. Yakpo, Kofi. 2020. Social factors. The Routledge handbook of language contact, ed. by Evangelia Adamou and Yaron Matras. London: Routledge, to appear. Yakpo, Kofi, and Pieter Muysken. 2017. Language contact and change in the multilingual ecologies of the Guianas. Boundaries and bridges: Language contact in multilingual ecologies (Language contact and bilingualism 14), ed. by Kofi Yakpo and Pieter C. Muysken, 3–19. Berlin: De Gruyter Mouton. Yip, Moira. 2002. Tone. Cambridge: Cambridge University Press. Zamora, Armando. 2010. Gramática descriptiva del fá d’ambô. Barcelona: Ceiba Ediciones. Zamponi, Raoul. 2009. Focus in Fang and other Bantu A70 languages. Siena: Università degli Studi di Siena, ms. Online: https://www2.hu-berlin.de/predicate_focus_africa /data/Zamponi_-_Focus_in_Fang.pdf. Bordal Steien [Received 9 September 2017; Inland Norway University of Applied Sciences revision invited 13 January 2018; Holsetgata 31 revision received 17 September 2018; 2318 Hamar, Norway revision invited 18 February 2019; [[email protected]] revision received 17 June 2019; accepted pending revisions 17 July 2019; 40 LANGUAGE, VOLUME 96, NUMBER 1 (2020) Yakpo revision received 12 August 2019; The University of Hong Kong accepted 12 August 2019] School of Humanities Pokfulam, Hong Kong [kofi.hku@hk] On the outcomes of prosodic contact 41