Principles in the emergence and evolution of linguistic features in World Englishes

Table of Contents

Abstract

List of abbreviations

1. Introduction
1.1 Aim of this study
1.2 Methods
1.3 Basic assumptions

2. Current research
2.1 Sociolinguistic approaches
2.1.1 Schreier – linguistic endemicity
2.1.2 Andersen – center and periphery
2.1.3 Trudgill – new-dialect formation
2.2 Language evolution
2.2.1 Mufwene – language as an organism
2.2.2 Croft – Theory of Utterance Selection
2.3 Linguistic formalism
2.3.1 Chomsky – Universal Grammar
2.3.2 Bickerton – Language Bioprogram Hypothesis
2.3.3 Chambers – vernacular universals
2.4 Synthetic approaches
2.4.1 Tomasello – usage-based approach
2.4.2 Ansaldo – Adaption Theory
2.4.3 Bybee – usage-based functionalism
2.5 Summary

3. Selection of varieties of English
3.1 Traditional L1 variety: Scottish English
3.2 High-contact L1 variety: New Zealand English
3.3 Indigenized L2: Chicano English
3.4 Creole: Bonin Island English/Ogasawara Mixed Language
3.5 Pidgin: Norfolk Island/Pitcairn English

4. Selection of linguistic features
4.1 Morphosyntactic features
4.1.1 F34: alternative forms for 2nd person plural pronouns
4.2 Phonological features
4.2.1 Monophthongization
4.2.2 Th -movement

5. Discussion
5.1 Sociolinguistic approaches
5.1.1 Theoretic outline
5.1.2 Matching the data
5.2 Language evolution
5.2.1 Theoretical outline
5.2.2 Matching the data
5.3 Linguistic formalism
5.3.1 Theoretical outline
5.3.2 Matching the data
5.4 Synthetic approaches
5.4.1 Theoretical outline
5.4.2 Matching the data
5.5 Conclusion – principles at work

6. Considerations
6.1 Does the input matter? Languages vs. dialects in contact
6.2 A different perspective on universals
6.3 Reflection upon material and methods
6.4 Outlook

Bibliography

Appendix I
Appendix II
Appendix III
Appendix IV
Appendix V
Appendix VI

Abstract

This study deals with different explanatory models for the emergence or existence of linguistic features in varieties of the English language. After a brief overview of the current research, five non-standard varieties from all over the world, ranging from a traditional dialect to pidgins and creoles, are analyzed in two morphosyntactic and two phonological features. The theoretical approaches are discussed with reference to the features, providing recommendations for or advise against certain explanatory models. Finally, Bybee's usage-based functionalist approach and the usage-based syn­the­sis of new-dialect formation according to Ansaldo are highlighted as plausible expla­nations for the features. Formalist, descriptive universals are rejec­ted in favour of functionalist, cognitive universals in human language processing, acquisi­tion and evolution, as they occur in language contact or speaker contact sce­narios – the dri­ving force of language change.

List of abbreviations

illustration not visible in this excerpt

1. Introduction

1.1 Aim of this study

In this study, five varieties of English will be compared in two phonological and two morphosyntactic features. In advance, different theories or explanations for processes in language development will be discussed. Later, the theories will be taken into consideration for the explanation of the prevalence, varying in degree, of the features discussed in the different varieties. Hereby, certain principles, general processes and tendencies in language development are supposed to be uncovered, confirmed or refuted. Different models of explanations will be matched to certain features, and recommendations on each theory will be made, hopefully resulting in the contribution of an integrated model.

1.2 Methods

In order to achieve the aims, a set of five different varieties of the English language, ranging from traditional Scottish English dialect to an exotic pidgin spoken on 9remote Pacific Pitcairn Island, will be analyzed in four features. Two of them will focus on phonology, another two focus on morphosyntax. The selection of features was made partly in order to grant fair chances to each explanatory model, and partly because of their high pervasiveness in most varieties. The five varieties of English follow the five categories of non-standard varieties in the World Atlas of Varieties of English, WAVE (Kortmann/Lunkenheimer 2012: 3f.). As a next step, explanatory models and theories from a number of mainly sociolinguistic paradigms, such as lan­guage universals represented by Chambers, dialectology by Trudgill, evolutionary linguistics represented by Mufwene, a usage-based approach by psycholinguist To­masello, or an attempt of an integrated model by Ansaldo, are discussed, interpreted and matched with data from the varieties and features.

While doing so, we will encounter a large number of terms and concepts that originate from biology, especially genetics, in descriptions of linguistic phenomena. This is due to the fact that – as we will see – they share characteristics in many res­pects. Nevertheless, one aim of this study is a balanced discussion of each model, which hopefully leads to a juxtaposition of the advantages and disadvantages of each model in the explanation of the chosen features in the five varieties.

1.3 Basic assumptions

Ca veats

WAVE and eWAVE samples and feature rankings are often dependent on only one informant, and often the body of source material is rather slim, or in some cases, outdated. As languages, contact languages in particular, are instable and may change in a short period of time, written evidence from half a century ago (e.g. Ross 1964 on Pitcairn/Norfolk English) cannot be left unquestioned. In addition, the selection both of the features and the varieties was not randomly done. This might skew the results, just as the selection of explanatory models or their adaption on contact languages.

Generalizations

The final recommendation for the best matching fields of explanatory models, fea­tures, and varieties will be a generalization, but it is supposed to give a hint where the strengths and weaknesses of each theory are, and what further models might take into account.

Definitions

Before starting with the analysis we need to define certain (socio-)linguistic con­cepts. When speaking about a linguistic situation of a geographical or political area, we speak about:

die ethnische und/oder regionale Verteilung sowie die soziale Distribution und Hie­rarchie der Sprachen und/oder Sprachvarianten, die zu einem gegebenen Zeit­punkt auf einem bestimmten (meist politisch-administrativ abgegrenzten) Territorium ent­sprechend den dort herrschenden ethnischen, politischen, sozial­ökonomischen und kulturellen Bedingungen zur Kommunikation verwendet wer­den. (Hansen et al. 1996: 13)

Furthermore, this very situation is characterized by historical processes leading to its coming into existence, and the intensity, aims and fashion of language policy (ibid.). For this reason we will deal with the history, the characteristics, functions, prestige, areal and social distribution and political circumstances of each variety in order to get to know them in great detail. The linguistic situation has various names and facets in the different approaches, for instance linguistic ecology. A language com­munity can be defined as a "group of people who regard themselves as using the same language" (Hansen et al. 1996: 14), which implies a certain amount of linguis­tic awareness. This, next to a shared ethnic and cultural identity, can create a commu­nal spirit and form a certain togetherness. This definition will be most productive for our purpose here. Nevertheless, we have to take into account two more perspectives on linguistic communities. A code community includes speakers who use one lan­gu­age as means of communication in intra-national exchange, regardless whether L1 or L2, but not as a foreign language (ibid.). A primarily social perspective is pro­vided by the term communication community that is characterized by fairly stable so­cial relations and hierarchies within a community which leads to a specific selection and distribution of linguistic means and features, regardless whether only one or more languages are used. One individual can be member of a number of communication communities, depending on the dialogue partner and situation (ibid.). This might be a helpful perspective in colonial contact situations with clear-cut social hierarchies. As a next step, we need to look at the linguistic potential of a language community, i.e. "die Zahl und Art der zur Kommunikation innerhalb der Gemeinschaft verwendeten Sprachen und/oder Sprachvarianten, deren Status und sozialkommunikative Funktion sowie auch deren regionale Verbreitung" (Hansen et al. 1996: 16). As we are mainly interested in contact situations here, always more than one language will be en­coun­tered. It is important to see whether there is a hierarchy, a substrate, adstrate or su­per­strate situation of the English language, which variety is more prestigious, to what extent and why the output is more or less influenced by what input.

Hansen et al. (1996: 20) find two main influences for the attitude of speakers towards the languages or varieties used in their communication community. Firstly, it is the ethnic or regional origin and the social or socio-economic status; secondly, it is the language's or variety's prestige within this community. Those two factors do not necessarily have to be congruent. Overt prestige is important for the variety's ap­pro­val within the community and is mainly influenced by the speakers' associated social role and status. Covert prestige, on the other hand, seems to contradict the public opinion. As a symbol for a minority's ethnic and social identity, it can become an expression of group solidarity and distance from the establishment, which in turn can also lead to depreciation or rejection by the majority. A variety's or language's prestige can influence the willingness to acquire or master it, or to adopt certain features.

For a clear understanding of variety categories we need to define terms such as L1, L2, creole, and pidgin. An L1 is the speaker's mother tongue. It is the language that a person learns as a child and that he or she uses primarily. An L2 is a second language, "a language that sb learns to speak well and that they use for work or at school, but that is not the language they learned first" (Hornby 2010: 1380).

A pidgin is "a simple form of a language, especially English, Portuguese or Dutch, with a limited number of words, that are used together with words from a local language" in contact situations of people who do not speak the same language or one speaker does not speak a language well (Hornby 2010: 1144). In this definition, the focus lies upon simplification, imperfection, language contact and the typical colo­nial languages. A more complex definition by Daniel Long (2007: 4) gives an extra perspective on the socioeconomic situation of pidgin communication:

A pidgin is a language system that evolves when speakers of two, three, or more languages come into contact with each other and cannot easily understand one another's language. Typically, the language of the people with 'power' (through economics, technology, warfare, sheer numbers, etc.) is learned imperfectly by the other groups. These groups acquire lexical morphemes from the powerful lexifier (or superstrate) language, but their understanding of grammatical mor­phemes and syntax (the way words are joined together to make meaningful sen­tences) is influenced by their various native languages (the substrate langua­ges). Their misinterpretations (reinterpretations) of the grammar of the target language result in the grammatical simplification and restructuring of the language.

In addition, there might be many "complex relationships that the speaker can conceive of (in her mind, in her native language) but cannot verbalize in the pidgin due to its grammatically [and lexically] limited nature" (Long 2007: 6). Only a restricted pool of expressions and structures is available making formulations outside this corridor impossible.

Important here is that a pidgin, by definition, is not spoken as a first language in contrast to a creole language. "[P]idgins have no native speakers. A user of a pidgin is by definition a native speaker of some other language. A nativized pidgin is a creole" (ibid.). A creole is "a language formed when a mixture of a European lan­guage with a local language (especially an African language spoken by slaves in the West Indies) is spoken as a first language" (Hornby 2010: 359). This definition, though, is a bit narrow because it focuses on European languages, which is not ne­cessary. Creolization is a process of "expansion through the nativization of a pidgin, and the creole language that the children create is a full-fledged language in which there are grammatical structures to express the cognitive relationships their minds come up with" (Long 2007: 6). In Long's words (ibid.), we will later refer to creo­loids, a contact language with the typical admixture of two or more languages but with less dramatic processes of restructuring, simplification and expansion. A highly recommended introduction on pidgins and creoles can be found in Sebba 2009.

2. Current research

We can find a broad range of perspectives on linguistic phenomena such as isolated speech communities, language contact, internally and externally motivated change, and the resulting features in varieties of English all over the world. In this chapter, a number of these perspectives will be presented in brief. Here, different attitudes towards new dialect formation or contact language formation become apparent, in­cluding arguments for and against each approach. The perspectives are grouped ac­cording to the main statements they make, even though it is difficult to form groups. There are overlapping aspects in the different theories, but also mutually exclusive aspects.

We start with Schreier in order to explain the phenomenon of linguistic contact and isolation, and continue by outlining the relevant principles that might be at work in contact situations. Here, universalistic, evolutionistic, psycholinguistic, and mainly sociolinguistic approaches will be presented, discussed, and used as a starting point for the discussion in chapter 5 in which data from chapters 3 and 4 will be included.

2.1 Sociolinguistic approaches

Almost all approaches in chapter 2 have a sociolinguistic dimension as contact situ­a­tions are often central aspects of their statements. The approaches in 2.1 are mainly concerned with the speakers and their geographic and social movements, though, and put emphasis on the social facet.

2.1.1 Schreier – linguistic endemicity

Daniel Schreier (2003) stepped into the tradition of usage of terminology from biology for linguistic purposes, see 2.3. Endemicity in biology or medicine means the existence or distribution of species, both animals and plants, only in a certain locally restricted area (Duden 2009: 274). Insularity is geographical isolation including its special case of local restriction on a single island. Linguistic endemicity then is the occurrence of certain linguistic phenomena in geographically isolated speech communities. In his research, Schreier mainly focused on the South Atlantic island of Tristan da Cunha, probably the most remote inhabited island in the world. The closest settlements are St. Helena 1,400 miles north and Cape Town 1,800 miles east (Schreier 2003: 252). His findings can be adapted to most of the varieties we will discuss here as they share characteristics such as a small population, and geo­gra­phi­cal and social isolation.

2.1.2 Andersen – center and periphery

Henning Andersen (1988) describes new-dialect formation with an expanded model of the classic Adoption Theory. In this theory, the simplistic standard scenario for contact-induced language change is outlined. Speakers with different traditional norms get into contact, adopt each other's norms and adjust their usage accordingly. Features that were formerly marked differences in the two varieties are obliterated or merged; traditional features will not continue to be passed on. As a result, a linguistic boundary – an isogloss – disappears (Andersen 1988: 40). Here, the new and vital point is the model of center and periphery. Andersen describes a phenomenon that was already discovered by Ferdinand de Saussure, namely that "regardless of which language areas they work with, there are palpable differences between the kinds of developments that characteristically occur in central and in peripheral speech areas" (ibid.: 39). Depending on the characteristics of the dialect, speakers are conservative or open toward change. This creates reduction and regularization, or differentiation and complication, i.e. language change.

2.1.3 Trudgill – new-dialect formation

Trudgill (2004) proposes three stages of new-dialect formation, each subdivided in several steps. As he states, the relative influence of "language contact versus dialect contact [...] may be rather hard to disentangle" (2004: 5). With small adaptations, his theory of new-dialect formation seems to work for both dialect and language contact.

Trudgill describes a process of new-dialect formation in six steps. After coming together in a special location, speakers of different dialects or mutually intelligible languages mix in a multi-dialectal society. The variants will reduce over time in accordance with certain deterministic, social factors such as socio-economic status of the speakers and the associated prestige. Too heavily marked features will be reduced or abandoned because they are not locally shared. Unmarking as a subtype of levelling often happens in favour of forms "which offer greatest structural simplicity" (Trudgill 2004: 86). An interdialect develops. This is a dialect that has a set of fea­tures which was not actually present in any of the of the contributing dialects, and which arise out of interaction between them. The interdialect's structure and system may be simpler, as complex as, or – owing to hypercorrection – even more complex than a speaker's initial dialect. After reallocating leftovers of initial dialects to phono­logical, morphological, or register-related niches, the five necessary steps of koi­néisation are completed. After that, in a final step, the new dialectal features are focussed; this is their norms and stability are consolidated within the communication community (Trudgill 2004: 84-89). According to him, the six steps occur in three different chronological stages, roughly corresponding with three successive genera­tions of speakers (Trudgill 2004: 89).

2.2 Language evolution

2.2.1 Mufwene – language as an organism

Salikoko S. Mufwene is mostly known for his concept of language ecology and evolution including the ideas of the founder principle and the feature pool which he has elaborated since the mid-1990s (Mufwene 2012a). He defines language as "a complex adaptive system and as a piece of technology that was built incrementally and has been modified several times over by its users and makers (speakers and signers alike) to meet their current communicative needs, under the influence of habits developed previously" (Mufwene 2012b: 3), or in another text, as a "Lamarckian species, whose genetic makeup can change several times in its lifetime. It is also a parasitic species, whose life and vitality depend on (the acts and dispositions of) its hosts, i.e. its speakers, on the society they form, and on the culture in which they live" (Mühlhäusler 2005: 266).

In the course of time, speakers of a language introduce "variation and therefore competition and selection, as different innovators often introduce variants (forms or structures) for the same functions" (Mufwene 2012b: 3). First, change happens in idiolects on an individual level. Later, this change, if accepted by the speech community, will be applied by a larger number of speakers. Over time, the language self-organizes communal norms, reduces variation, and certain standards rise to population level (ibid.). He "focused on how indirect external ecological factors (e.g. population movements, the particular dialect mix of the allopatric^[1] population, the kinds of languages spoken by the people they came in contact with in the colony, and population structure, which determines patterns of social interaction) influenced language change" (ibid.).

Exactly his inspiration from biology has led him to the re-introduction of the founder principle^[2] that is also called the doctrine of first effective settlement (Mühlhäusler 2005: 266). It explains how "structural features of creoles have been predetermined to a large extent (but not exclusively!) by characteristics of the vernaculars spoken by the populations that founded the colonies in which they developed" (Mufwene 1996: 84). The "ethnographic setting in which the lexifier [...] has come into contact with diverse languages (or populations) whose structural features [...] enter into competition with its own features" is called the ecology of a language (ibid.: 85). At this point, Mufwene puts great emphasis on the idiolectal level. He says that every individual's realization of his or her L1 is an incomplete abstraction of the L1 because of the speaker's limited input - depending on the ecology on the individual's level. This means two speakers of the same language must differ in their acquisition of their L1 due to input that is not identical.

When dealing with Mufwene, analogies to biology, especially genetics, become obvious. He first started to compare language with population, as speakers are the agents of a language – just as living creatures are the carriers of genes. A lan­gu­age is only a useful but abstract extrapolation of the (mutually intelligible) idio­lects of their speakers – just as a species is an abstraction of similarities of its (genetically compa­tible and similar) members. There is interindividual variation in genes and L1 reali­zation, the latter identified by linguistic features. As genes are inherited, lingu­istic features can be passed on, though rather via transmission and restructuring than by inheritance, but both being dependent on the ecology or surrounding. Further­more, there is change in both genetics and linguistic usage, which can be described as evolution (Mufwene 2002: 46). In the same way, the feature pool is an analogy to the gene pool (ibid.). Input from feature donors are more or less equally collected in a pool which causes competition among them. Finally, a selection is made and certain features make it to the output language with different probabilities, some of them only appearing sporadically in individuals, others being salient in the majority of the population; comparable to recessive and dominant genes, one could say (ibid: 46f.). The founder principle, in biology a term for a reduced gene pool owing to geographic or genetic isolation of (small) popu­lations, is used to describe the increased likeli­hood of an establishment of the founders' linguistic features existing in a population. Here, it is more likely that the lexifier's features will dominate and only few items (or genes) of the substrate will prevail.

2.2.2 Croft – Theory of Utterance Selection

Croft, himself influenced by Hull, developed a similar concept, the Theory of Ut­terance Selection. Language can be seen as a "population of linguistic features and grammar as a combination of idiolects" (Ansaldo 2009: 14). Social forces such as prestige or status are understood as a "mechanism that selects an innovative variant for subsequent propagation across the speech community" (Croft et al. 2006: 2). An utterance is taken as the analogue to DNA as it passes (grammatical) information from one speaker to another, forming the lingueme-genome- analogy . Speakers are interactors, genes are replicators; speakers, by exchanging linguistic features, repli­cate the lingueme. Variation exists not only among speakers but also within indivi­duals. And because neither two speakers nor two situations are alike, communication creates variation through imperfect replication of words, sounds and constructions speakers have heard before, an aspect also considered in usage-based models (chap­ter 2.4). Social circumstances such as the "social success" cause the "differential survival of the linguemes they produce" (Croft et al. 2006: 3). This accounts for change over time (ibid.: 2f.). Similarities to Mufwene's Feature Pool notion with its competition and selection become apparent.

2.3 Linguistic formalism

2.3.1 Chomsky – Universal Grammar

The notion of linguistic universals is highly debated and ambiguous. In general, uni­versals can refer to "a superficial descriptive property true of the expression of all languages", a descriptive universal, or to a cognitive universal, "a property true of all human minds" (Winford 2013: 224). Descriptive universals are associated with Jo­seph Greenberg's functional, empiricist approach in which a representative sample of languages is typologically compared in certain features, and cross-linguistic generali­zations are made. Cognitive universals have their origin in Noam Chomsky's forma­list approach of generative grammar, which proposes a set of universal principles that limit the possible forms of grammar, both in language acquisition and change. Here, central is the formulation of generalizations about the "essential nature of language, from which particular language-specific grammatical features can be derived", lea­ding to a theoretical construct of Universal Grammar, an innate human language fa­culty (Winford 2013: 225). While functionalists usually make empirical studies and derive their deductions about similarities in languages and in mechanisms of langu­age change world-wide from corpus-based data, their evidence and results seem to be more relevant and logical than those of formalists who have a starting point of innate grammar governed by universal principles, and try to support this view with data, as we will see below (ibid.).

The Universal Grammar Hypothesis can be characterized by four interrelated claims (see Goldberg 2009: 202):

1) domain-specificity: Language acquisition is constrained by representations or principles that are specific to language;2) universality: These representations or principles are universal;3) innateness: These representations or principles are not learned;4) autonomous syntax: These representations or principles depend on syntactic representations and not their functional correlates.

This means that every newborn child has the capacity to acquire a language (L1) relatively easily, the faculty of speech and basic grammatical processes are innate, genetic information is the source for learning every possible language because the input during first language acquisition triggers certain parameters to attain a distinct value, all of them following deep universal principles. The language faculty as the basis for linguistic universals is a black box, but functionalist-inductive approaches are interested in exactly those processes happening in this black box, "providing external explanations for observable universal properties of language and mainly address physical and cognitive constraints" (Siemund 2009: 334). The combination of all parameters is characteristic for each language. However, there is a lack of data on this theory and thus no empirical or developmental-psychological support (Nün­ning 2008: 95f.).

2.3.2 Bickerton – Language Bioprogram Hypothesis

Moreover, we can find Derek Bickerton's Language Bioprogram Hypothesis, LBH, a formalist approach that treats creole genesis as the outcome of L1 acquisition "in a context of restricted linguistic input from the surrounding community" (Thomason 2001: 178). Children growing up in a highly unstable linguistic environment, as in pre-pidgin-speaking plantation creole communities, construct a grammar that derives from their innate bioprogram, so to speak a genetically programmed grammar "hard-wired in every newborn human infant's brain" (ibid.). In a stable, 'normal' sur­roun­ding, the grammar of the community overlays the bioprogram features, wiping out genetic traces. This approach seems "shaky on the empirical evidence" (ibid.: 179) and was refuted in various recent studies on Caribbean English-lexicon creoles, showing no evidence for grammatical predetermination in Bickerton's sense (Win­ford 2013: 228). The LBH can only account for a small number of incidences of plantation creole L1 acquisition, and is thus not an option for the five varieties dis­cussed in this study.

2.3.3 Chambers – vernacular universals

The term of vernacular universals or vernacular roots was introduced by socio­linguist Jack Chambers "to refer to linguistic features which are absent from Stan­dard English, but which recur in many different non-standard varieties of English around the world" (Trudgill 2009: 307). Chambers (2004: 128) emphasises that because universals "arise naturally in pidgins, child language, vernaculars, and else­where, they are primitive features, not learned. As such, they belong to the language faculty, the innate set of rules and representations that are the natural inheritance of every human being." We can clearly recognize Chomsky in this approach. Chambers says that "[s]ociolinguists have amassed copious evidence in the past 35 years for a surprising conclusion: a small number of phonological and gram­matical processes recur in vernaculars wherever they are spoken. This conclusion follows from the observation that, no matter where in the world the vernaculars are spoken [...] these features inevitably occur" (Chambers 2004: 128). Their prevalence can only come into existence by diffusion by the dialect's founders, or by developing "independently as natural structural linguistic developments" (Chambers 2004: 128). Because, according to Chambers, any approach basing on diffusion is inadequate here – distances are just too great and universals occur in all types of varieties (ibid.: 128) – the discussion will focus on vernacular universals. Examples are con­sonant cluster simplification, copula deletion, or multiple negation (ibid.: 129). Pos­sible reasons for their occurrence may be underlying principles of cognitive over­load, motor eco-no­my, or avoiding redundancy (ibid.: 140), tendencies that are per­vasive in vernaculars but suppressed in the standard. These reasons, in fact, are of a rather cognitive-functional nature.

2.4 Synthetic approaches

2.4.1 Tomasello – usage-based approach

Michael Tomasello, co-director of the Max Planck Institute for Evolutionary Anthro­pology, Leipzig, is expert on developmental psychology and psycholingu­is­tics. His theory on L1 acquisition in young children can hardly be adapted to con­tact scena­rios, and young children's processes in language acquisition are different from those of adults, as "children operate with different psycholinguistic units than adults" (Tomasello 2000: 62). Nevertheless, it underlines the importance of linguistic role models, i.e. linguistic input in the early stages of language acquisition. Usage can be understood in two ways. On the one hand, a speaker can become used to a structure through permanent exposure; on the other hand, the speaker applies a structure in form of imitative and inductive learning. Imitative refers to the circumstance that children hear how their linguistic environment speaks. "In the early stages, children mostly use language the way they have heard adults using it. This leads to an inventory of item-based utterance schemas" (Tomasello 2000: 70). In other words, language use shapes the grammar and lexicon (Bybee 1999: 236). Inductive refers to the process of making generalizations from an exemplary usage on a grammatical structure, because "[c]hildren are focused on the adult's communicative intentions as they attempt to comprehend her immediate utterance, and communicative function is the main basis for their linguistic generalizations over time (otherwise they would be totally baffled by a language's many homonyms and proforms, among other things)" (Tomasello 2003: 324).

Furthermore, high-frequency use of words and phrases leads to a certain automa­ti­zation and phonological reduction, which in turn leads to a higher variability in the item's realization and perception or recognition, but high frequency can also lead to entrenchment or lexical strength, making the item resistant to change or conformity. Moreover, "linguistic capabilities are not presumed to be different in structure from other cognitive capabilities. Linguistic units are stored like other percepts that come from our experience [...] Thus there is no separation of lexicon and grammar, lexicon and phonology" (Bybee 1999: 236).

2.4.2 Ansaldo – Adaption Theory

In his 2009 paper, Umberto Ansaldo tried to integrate several approaches to typology in contact linguistics. With his draft of a functional-typological theory of language, he shows how to account for as many linguistic phenomena in language change as possible without having to employ universals. He assumes that there is no significant difference in contact-induced language change taking place in traditional types of contact languages such as pidgins, creoles and mixed languages (Ansaldo 2009: 2). Differences are only there in labels given for socio-historic reasons. His framework is inspired by Croft's Theory of Utterance Selection, added by an interpretation of the feature pool notion in language contact (ibid.: 5). He suggests "an evolutionary framework based on principles of selection, innovation and propagation, with the help of functional-typological analysis of the matrix. In this way, sociohistorical dy­namics and functional-typological features are integrated within the same frame­work" (ibid.: 27).

Ansaldo names three principles in Contact Language Formation, CLF. Firstly, we can find differential replication in contact scenarios, secondly, we understand contact scenarios as a complex typological matrix, and thirdly, in contact language formation "selection, innovation and propagation occur iteratively and feed into one another. The most likely candidates for selection and propagation are determined based pri­marily on sociohistorical analysis and typological make-up, within which frequen­cy patterns play a dominant role" (Ansaldo 2009: 27f.).

Andersen's criticism on terminology and concept borrowings from biological evo­lution (see section 2.3) is denied by Ansaldo because of two aspects. Firstly, it demands a complete overlap between two explanatory models in order to be useful, and secondly, it is Croft's field of conceptual – not biological – evolution that pro­vi­des the necessary foundations for a linguistic framework (An­saldo 2009: 5). As long as such analogies are fruitful, they are welcome.

As Goldberg (2009: 219) mentioned, the "(mostly minor) differences among various cognitive, functionalist or usage-based approaches pale in comparison to the stark contrasts between these approaches and traditional generative grammar." Accor­ding­ly, Ansaldo does not use Universal Grammar as an explanation here, because "[w]hat cannot be reconstructed does not necessarily indicate UG, universal cognitive patterns or other abstractions, but may simply indicate a gap in our knowledge" (2009: 3f.). Imperfections in the data of linguistic ecology and history or understan­ding of mechanisms may provide better explanation than "invisible hand changes [that] should be treated very carefully, as a last resort in trying to account for CLF" (Ansaldo 2009: 28). His criticism on Andersen's denial of evolutionary concepts was explained with their fruitfulness. We will see below whether or not a bias towards the explanation of formalist language change with universals might be a similar mistake.

2.4.3 Bybee – usage-based functionalism

Bybee supports a synthesis of functionalism and usage-based approaches, both being of cognitive and functional nature, combining language structure with language use. The central statement is that "the general cognitive capabilities of the human brain, which allow it to categorize and sort for identity, similarity, and difference, go to work on the language events a person encounters, categorizing and entering in me­mory these experiences" (Bybee 2006: 711). This results in a cognitive repre­sen­ta­tion of linguistic experience, both morphological and phonological, that can be called grammar. Important are three effects of usage. A high token frequency, the repeated occurrence of a word or phrase, leads to a faster rate of phonetic reduction through automatization and neuromotor routines. Furthermore, high-frequency tokens be­come lexically more entrenched than low-frequency ones, making them more resis­tant to morphological change.^[3] Analogical reformation is not necessary because high-frequent words are easily accessible in the memory. Thirdly, morphologically com­plex and high-frequent forms can lose their internal structure and become autono­mous from their etymological source – they grammaticize^[4] and become pro­ductive (Bybee 2006: 715). The storage must be very complex, as Bybee states:

Each token of use of an item affects its memory representation. Since tokens of use vary, the stored representation must include a range of variation. As words slowly and gradually reduce in production, the center of the range of variation gradually shifts. [...] Not only do lexical representations have to be fully speci­fied and represented in concrete phonetic units, these units cannot be an idea­lized systemic phonetic set of units, but rather must represent in some realistic way the range of variation occurring in the individual pronunciations that are constantly being mapped onto the existing representations. (Bybee 1999: 221)

2.5 Summary

In this chapter, the basic ideas of different theoretical approaches to linguistic change (in contact situa­tions) have been outlined. On the one hand, we can see profound dif­ferences between some approaches. On the other hand, other approaches seem to take similar or related paths. A combination of the approaches seems to be pro­duc­tive in accounting for linguistic processes and mechanisms of language change. As a next step, varieties of the English language will be presented with a focus on their linguistic ecology. After providing feature samples in chapter 4, our findings hitherto will be discussed and matched to the approaches in chapter 5.

3. Selection of varieties of English

In order to create a subset as interesting and representative as possible, five varieties of English were chosen for our purpose. We find one member of every variety type as classified in Kortmann/Lunkenheimer (2012: 3f.). The five types are:

1) L1t, a low-contact traditional L1 dialect or native-speaker variety, defined as "[t]ra­ditional, regional non-standard mother-tongue varieties, e.g. East Anglian Eng­lish and the dialects spoken in the Southwest, the Southeast and the North of Eng­land" (Kortmann/Lunkenheimer 2011),
2) L1c, a high-contact L1 variety, including "transplanted L1 Englishes and colonial standards (e.g. Bahamian English, New Zealand English), as well as language shift varieties (e.g. Irish English) and standard varieties (e.g. collo­quial American Eng­lish)" (ibid.),
3) L2, an indigenized non-native variety that compete with local native languages, and "that have a certain degree of prestige and normative status in their political commu­nities, like Pakistani English, [...] but also non-native varieties that compete with local L1 varieties for prestige and normative status, e.g. Chicano English and Black South Af­rican English" (ibid.),
4) Creoles, English-based contact languages and native language to many people, and "that developed in settings where a non-English-speaking group was under strong pressure to acquire and use some form of English, while access to its L1 spea­kers was severely limited (e.g. in plantation settings). Many creoles have become the native language of the majority of the population", e.g. Jamaican Creole (ibid.), and
5) Pidgins, "English-based contact languages that developed for communication be­tween two groups who did not share the same language, typically in restricted do­mains of use (especially trade)." Almost all pidgins in eWAVE can be considered expanded pidgins in contrast to prototypical pidgins, i.e. they are less restricted in the domains of use, and many people speak them as native or pri­mary languages (ibid.)^[5].

Furthermore, the chosen varieties have historically quite well-recorded influences with respect to the origin of their settlers. In other words, we know the linguistic eco­logy of these varieties quite well, which provides fair chances of explanation to all theoretical approaches. Another aspect is the broad but distinct variety of sub­strate in­fluences. We can find European, Asian, and Pacific languages in contact situations with non-standard varieties of English, creating quite a diverse impression, see ap­pendix I

Scottish English was chosen as L1t because of its distinct features distinguishing it from Standard British English, and its influence on other language types as a result of contact situations due to colonial seafaring in the past centuries, especially as of the seventeenth century colonial expansion which finally lead to an increase of English-speakers all over the world (Hansen et al. 1996: 25). All varieties discussed here have foun­ders who were British, partly Scottish, navy sailors – "men of little education" and probably speakers of a non-standard variety of English (Zet­tersten 1969: 133). An at­tractive L1c is New Zealand English, spoken almost at the opposite end of the world and influenced by native Maori. The indigenized L2 in this work will be Chicano English which is mainly spoken by Mexican immigrants to the United States but which took an interesting development. The Bonin Island English, also called Ogasa­wara Mixed Language, is an English-Japanese hybrid spoken on an archipelago south of Japan and will serve as creole. Last but not least, we will deal with the pid­gin spoken on Norfolk Island and Pitcairn with its Tahitian roots.

In the following, the five varieties will be introduced in order to gain insight in their sociolinguistic, historical and geographic situation, outlining their main charac­teris­tics and providing aspects for later discussion.

3.1 Traditional L1 variety: Scottish English

It seems to be rather difficult to define the term Scottish English. Aitken and others think of "Scottish English as a bipolar linguistic continuum, with broad Scots at one end and Scottish Standard English at the other" (Stuart-Smith 2004: 47). Maguire (2012: 55) expands this bipolar continuum to "a multi-dimensional sociolinguistic variation space" in which the speakers operate. This space is dependent on the spea­kers' socioeconomic class, level of education, identification as a Scot or a British, religion, urban or rural origin, age, and fashion of speaking, which still is an abstrac­tion from reality. By far the greatest differences between Standard English and ScE exist in pronunciation and intonation (Hansen et al. 1996: 71).

Scots is generally, with exceptions, spoken by working class people, and in informal situations with friends and family, mainly in the rural area. Scottish Standard English, in contrast, is typically spoken by educated middle class people in the urban areas of Edinburgh and Glasgow, and is used in more formal occasions (Stuart-Smith 2004: 47). Many speakers are able to switch between degrees of broader or standard-near Scots, what Aitken describes as style/dialect shifting or style/dialect drifting (ibid.). Most examples cited in WAVE are from the broad Scots end of the con­tinuum. For this reason, Scots features are mainly ranked B in the WAVE des­cription, as they do not account for all speakers in all contexts. This means, features are not pervasive in all occasions of language production; rather they depend on the situation's and speakers' sociolinguistic and socioeconomic characteristics (Smith 2012: 21).

How did this variety continuum evolve? Scots is often perceived of as Standard English spoken with a Scottish accent. The continuum itself results from dialect contact and language change over many centuries (Stuart-Smith 2004: 47), Smith (2012: 21) traces it back to the seventh century Old English dialect spoken in Northumbria influenced by further spread of English from the thirteenth century on­ward. Before the Anglian invasion, the area was predominantly Celtic-speaking, but a northern variety of Anglo-Saxon was introduced. About 150 years later, Vikings invaded Scotland from the south. At the time of the Norman Conquest, most people in the area of Scotland spoke a form of Celtic, while Norse was used in the far north and west, and Anglian was spoken in the south-east, with an increase of Anglian speakers from the twelfth century onwards (Stuart-Smith 2004: 47). Until 1500 A.D., a Lowlands variety of English known as Inglis – Gaelic was called Erse or Irish – developed under the main historical influence of Norse. Later, Norman French left its traces in Scots place names and literature (ibid.: 48). In 1398, Scots was declared the language of record, and flourished as a literary language, until influence of English increased after the 1603 Union of Crowns and 1707 Union of Parliaments (ibid.). In a process of language shift, Scots replaced Gaelic in the Lowlands and English replaced Gaelic in the Highlands (Maguire 2012: 53). From that time on, Standard Southern English became the written standard in Scotland while the spoken standard approximated as well, especially because of its prestige among the middle class. Today's spoken Scottish English in urban areas has a low overt prestige, and is considered as bad or degenerate. In contrast, rural varieties are considered good (Stuart-Smith 2004: 48). Despite the still prestigious role of Received Pronunciation in Scotland, most spea­kers do not assimilate, especially because a too obvious assi­milation in speech habits is perceived of as affected or hypercorrect, such as the two RP-oriented varieties Morningside accent and Kelvinside accent spoken in Edin­burgh or Glasgow respec­tively. These marked forms of RP are socially stigmatised for most speakers, and are mainly spoken by elderly middle-class women (Hansen et al. 1996: 71f.).

Today, there are roughly 5 million potential speakers of Scottish Standard English, of which two thirds speak Urban Scots. Still, it is difficult for both speakers and linguists to distinguish Scots from Scottish Standard English, and to determine whether or not it is an independent, autonomous language, facing the ongoing process of dialect levelling towards Standard English (ibid.: 49). Next to varieties of English, Scottish Gaelic is spoken, albeit only by 1.2% of the population, and passively understood by less than 2%. Only little phonological influence by Gaelic speakers on English is attested (ibid.: 50). Other ethnic minorities of Asian or African descent are statistically insignificant, except for agglomerations of immi­grants in urban areas, creating a bilingual, ethnically diverse culture in a number of public schools. Throughout its history, Scots has been under constant influence and was neither geographically nor linguistically isolated, and it shares many features with its neighbours northern English and (northern) Irish English (Maguire 2012: 54).

3.2 High-contact L1 variety: New Zealand English

More than 1,000 years ago, Polynesian people first set foot upon the islands of New Zealand as part of a long series of explorations in the Pacific. The islands were named Aotearoa, the land of the long white cloud. Even though the Maori people lost contact with other Polynesian people in the Pacific, the linguistic heritage is still strong (Hay et al. 2008: 3f.). "The ancestors of the present Maori people were Poly­nesian explorers who first arrived in New Zealand around AD 925. They came into increasing contact with English from the time of early European settlement, and were quick to adopt English as a language of trade and negotiation" (Bauer/Warren 2004a: 614).

The first Europeans to discover NZ were the Dutch in 1642, naming the islands after the Dutch province of Zeeland. In 1769, Captain James Cook made landfall and claimed the island for the British crown, though British sovereignty was gained not until 1840. From that time, the European settlements grew rapidly, and eighteen years later, European immigrants outnumbered the indigenous population (ibid.: 4). When gold was found and ranching became popular in the second half of the nine­teenth century, the number of immigrants grew even further, bringing settlers from Canada and Australia (Hansen et al. 1996: 168). Many Maori people died in tribal warfare, were killed or enslaved by the settlers, reducing their population by 50% to 46,000 (Hay et al 2008: 5). In 1991, NZ had 3.4 million inhabitants of which 326,000 were Maori (Hansen et al. 1996: 169). In 2003, the population reached 4 million of which 79% are of European descent, and 14.1% are Maori, a strongly growing group, and 6.6% Asian. 98% of the New Zealanders speak English, only 4.6% speak Maori, which means that only 26% of ethnic Maoris speak the Maori language, mainly the older ones.^[6] The Ministry of Maori Development has set up programmes in order to rescue and promote the Maori language by establishing Maori pre-, primary and se­condary schools, radio and TV stations (Hay et al. 2008: 10f.). "With the continuing perception of New Zealand comprising two groups, Maori and Anglos, multilingua­lism is not considered a significant feature of the distinctiveness of New Zealand" (Clyne 1997: 295). Even though there is no single official language in New Zealand, Maori was proclaimed an official language next to English in 1987, which probably was more for symbolic rather than linguistic rea­sons (ibid.).

Despite its young age of a mere 150 years, there are several attempts to explain the NZ variety of English. "Some twenty years ago [almost 40 years by now; author's note], New Zealand English was generally thought of as like Australian English, only more English" (Clyne 1997: 294). WAVE, too, shows that NZE is one of the least non-standard varieties worldwide, with only 21 percent of the possible non-standard features attested; but those features attested were quite pervasive, leaving a distinct image of differences (Kortmann/Wolk 2012: 917). But there is a variety of concep­tions. Apart from rather ridiculous and party ideologically tainted explanations,^[7] we can find a few informants shedding light on this variety of English. Not an "idle tongue, a rigid jaw, atrophied labial muscles [...] will account for most of the habits and mannerisms that colour New Zealand's speech", as stated in a printed speech in the 1948's New Zealand Voice (Gordon et al. 2004: 69), but it rather has to be a mix of language contact, immigrant's speech, Australian influence, and new-dialect for­mation (ibid.: 69-79). The mixture of dialects, which was the source of later NZE, was mainly influenced by varieties spoken on the British Isles. "For New Zealand, immigrants arrived from England, Scotland and Ireland in proportions of roughly 50:27:23" with only few immigrants from Wales and Australia, and less than one percent of immigrants came from North America (Trudgill 2004: 13). The crucial pe­riod of development of a unitary New Zealand variety of English is dated between 1840 and 1890 (ibid.: 24f.).

[...]

^[1] Allopatric speciation in biology is the separate existence and development of closely related species in different places due to sudden geographic isolation.

^[2] "Whenever an empty territory undergoes settlement, or an earlier population is dislodged by inva­ders, the specific characteristics of the first group able to effect a viable, self-perpetuating society are of crucial significance to the later social and cultural geography of the area, no matter how tiny the initial band of settlers may have been [...] in terms of lasting impact the activities of a few hundred, or even a few score, initial colonizers can mean much more for the cultural geography of a place than the contributions of tens of thousands of new immigrants generations later" (Zelinsky quoted in Mühl­häusler 2005: 267).

^[3] We can see this phenomenon in English irregular verbs. Highly frequent verbs, such as keep-kept-kept, are less likely to regularize or to change morphologically than less frequent verbs such as weep-weeped-weeped (instead of weep-wept-wept), as it can be found in various sources (Bybee 2006: 715).

^[4] This process can be found in the going-to future that employs a grammaticized form of the former local meaning of going to a place in order to do something (Bybee 2006: 719).

^[5] Additional definitions for creoles and pidgins can be found in various places. Most often they have in common that they are not definitely sure about what pidgins and creoles are, or blur their def­initions. Some even say they form one group, the pidgin-creole-continuum, and a distinction between them is not fruitful. Of course, both are contact languages, but nevertheless, there are a number of differences, e.g. in function, historical origin, formal characteristics. Most often, definitions of pidgins include spontaneous generation, restricted vocabulary, absence of complex grammatical features, that they are not L1, and focus on essentials; definitions on creoles include nativeness (L1), and reduction of redundant features (see Romaine 1988, 23f.).

^[6] Clyne (1997: 295) comes to similar results: "90% of the population are monolingual English spea­kers. [...] There is some revitalization of Maori through kotanga reo (language nests), but only 25% of the Maori population speak the language, including very few of the younger generation."

^[7] For instance, the nasal quality of Australian and New Zealand English was explained with hay fever caused by the large amount of pollen in the air. The monotony and dull quality in speech, perceived by an Australian politician, was traced back to a "loss of enjoyment of life occasioned by the 1930s De­pression. Finally, a radio commentator found reasons for the dialect in bad or false teeth plus laziness in mouth movement, which was a popular view and caused the demand for speech training for New Zealanders, especially children and youths (see Gordon et al. 2004: 68f.).