Syntactic complexity and language contact: A corpus-based study of relative clauses in British English and Indian English

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. Alicante Journal of English Studies 30 (2017): 149-182 Syntactic complexity and language contact: A corpusbased study of relative clauses in British English and Indian English Iván Tamaredo Universidade de Santiago de Compostela [email protected] ABSTRACT The aim of the present paper is to test the claim that contact simplifies language (cf. Kusters, 2008) by comparing the domain of relative clause formation in British English, a L1 variety, and Indian English, a L2 variety. According to Hawkins (1999), the processing cost of relativizing a noun phrase increases down the Accessibility Hierarchy (Subject > Direct Object> Indirect Object > Oblique > Genitive> Object of Comparison) proposed by Keenan and Comrie (1977). Subject relative clauses are thus easier to process than direct object relatives, and so on. The results of a corpus study of the British and Indian components of the International Corpus of English show that the Accessibility Hierarchy has an indirect effect on the production of relative clauses in British English and Indian English: whereas the distribution of relative clauses with respect to the hierarchy is very similar in both varieties, the number of complex relatives, i.e., with coordination or further embedding, decreases in the lower positions in Indian English. These results thus suggest that language contact plays a significant role in relative clause use and accounts for certain differences between L1 and L2 varieties of English in this grammatical domain. Keywords: syntactic complexity, language contact, relative clauses, British English, Indian English, variation 150 Alicante Journal of English Studies 1. Introduction Language and dialect contact is a pervasive situation in the world nowadays. Bilingualism and multilingualism are the norm, while monolingualism is actually a restricted phenomenon (Valdés, 2012). As argued, for instance, by Trudgill (2009: 109), the evolution of language may be very different under the influence of contact: I have argued that language contact involving widespread adult language learning leads to an increase in simplification including loss of morphological categories. […] And I have argued, more hypothetically, that small community size and isolation may promote the spontaneous growth of morphological categories; and that they may also promote the growth of irregularity, redundancy, and low transparency. Thus, studying the effects of contact on language variation and change is an issue of utmost importance. The present contribution has two main goals, which are reflected in the title: (1) to propose a metric of syntactic complexity of relative clauses on the basis of the preferences of speakers, since the “locus of contact is the language processing apparatus of the individual multilingual speaker” (Matras, 2009: 3); and (2) to assess the effects of contact on relative clause complexity. To do so, relative clauses will be analysed in two varieties of English, British English (BrE) and Indian English (IndE), since they represent two different types of varieties: BrE is a native variety, and IndE is a L2 variety that developed under contact conditions and with a strong exonormative pressure. Two sets of results will be presented. Firstly, the distribution of relativizers and the different positions of the preposition in relative clauses that relativize a prepositional complement NP will be commented on. Second, the frequency of simple and complex relative clauses will be examined in the two varieties at hand, in order to discover contact effects.1The inclusion of relativizer choice and preposition placement in the study is motivated by the intrinsic interest that variation with respect to these different structural options has for a study of relative clause formation. However, it serves an additional function: previous research on relative clauses in English has focused mostly on the factors underlying the selection of relativizers and the placement of the preposition in prepositional complement relatives. Therefore, this analysis can be used to compare the present results with what has been found in earlier studies, and to test their validity and generalizability. The article is structured as follows. Section 2 introduces the metric used here to quantify the syntactic complexity of relative clauses, and reviews the effects of language contact found in previous research. Section 3 deals with the formation of relative clauses in Standard English and in Indian English. Next, section 4 describes the data and the methodology of the study, followed by the results in section 5. Section 6 focuses on the discussion of the results, and, finally, section 7 presents some conclusions. Syntactic complexity and language contact 151 2. Syntactic complexity Measuring grammatical complexity is definitely a complicated task. There is yet no agreement as regards the best way to determine the complexity of grammatical features, and less so if we move to the level of entire linguistic systems, such as phonology, morphology, or syntax (and even less so at the level of whole languages or dialects). Quantifying syntactic complexity is particularly difficult due to the abstract nature of the objects to be measured, which in this case comprise schematic rules and constructions (Dahl, 2009: 62-63). Previous operationalizations of syntactic complexity include, among others, • the number of rules that operate in the syntax of a language; the more, the more complex (cf. Szmrecsanyi and Kortmann, 2012: 9), • the degree of clausal embedding in a structure or the upper limit allowed by the grammar of a language; the more embedded clauses, the more complex (Karlsson, 2009: 192), • and the number of phrasal nodes that a syntactic unit (e.g., a phrase or a clause) dominates; the more nodes, the more complex (Szmrecsanyi, 2004: 1033).2 Givón (2009: 4-5) argues that complexity can be measured as the level of hierarchical organization of a system. In the case of languages, this means that syntactic complexity increases as linguistic elements are hierarchically grouped into phrases, clauses, and sentences. A simple transitive clause can have a 3-level hierarchical organization, as in example (1); coordinated clauses add an extra level of structure, as in example (2); and embedded clauses contain a 5-level hierarchy, as in example (3) (examples adapted from Givón, 2009: 4-5). (1) Simple transitive clause: 152 Alicante Journal of English Studies (2) Coordinated clauses: (3) Embedded relative clause: As can be deduced from this brief review, metrics of syntactic complexity can differ in the nature of the objects that they measure and in the perspective from which they approach those objects: some try to estimate the complexity of the syntax of a language by counting the number of syntactic rules or by calculating the upper limit of clausal embedding allowed by the grammar, while others focus on individual structures and quantify the number of nodes or the degree of embedding that they contain. The former are systemic metrics of complexity and the latter are structural metrics (Dahl, 2004: 42- 45). Systemic metrics measure the complexity of the rules of a grammar that produce the structures used by speakers, while structural metrics focus on the complexity of the structures that are the outputs of those rules. This is an important distinction because these two different types of metrics may provide diverging results. For instance, despite the Syntactic complexity and language contact 153 fact that, of all grammatical components, morphology and syntax have been traditionally considered the least vulnerable to the effects of language contact (Thomason and Kauffmann, 1988: 51-52), recent studies (e.g. Mukherjee and Gries, 2009; Schröter and Kortmann, 2016; Suárez-Gómez, 2017) have shown that contact languages or dialects tend to contain grammatical structures that are the result of transfer from one of the languages/dialects involved in the contact situation to the other(s). These innovations introduce new variants in the grammar and, therefore, increase its complexity from a systemic point of view by expanding the set of rules/constructions available. However, these innovative uses may result in simpler structures, i.e., with fewer phrasal or clausal nodes and fewer levels of hierarchical organization. In these cases (and many others), systemic and structural metrics provide opposite assessments of the complexity of the syntax of a language. A further distinction should be made between absolute and relative complexity metrics (Kusters, 2008; Miestamo, 2008). On the one hand, in absolute metrics complexity is understood as an objective property of grammars, which are in turn conceptualized as autonomous entities independent from considerations related to language use. A language is considered to be more complex the more elements (e.g. phonemes, morphemes, or syntactic patterns) and the more connections between elements it has. Relative metrics, on the other hand, understand complexity as rooted in the preferences of language users. Therefore, a language is more complex if it is harder for its users to process or learn. As with the systemic/structural distinction, these two different types of metrics may provide opposite results. Using again as an example the transfer of grammatical features from the native language/dialect to another in contact situations, an innovation may result, from an absolute perspective, in a more complex grammar with more rules/constructions, and even in more complex structures, i.e., with more phrasal or clausal nodes and more levels of hierarchical organization. However, from a relative point of view, these innovations may be easier to process or learn for speakers due to the fact that they are patterns found in the speakers’ native languages. The approach adopted in the present paper in order to quantify the complexity of relative clauses is structural and relative. It is structural because it measures the complexity of individual instances of relative clauses; it is relative because the relative clauses that are considered as complex are the ones that can be characterized as difficult to process for speakers on the basis of independent evidence. The next section (§ 2.1) describes this approach to syntactic complexity. 2.1. Relative clause complexity Relative clauses are characterized by having a gap in their structure, i.e., one element of the clause is ‘missing’ and has to be retrieved from a NP in the main clause in which it is embedded, called the head NP (Biber et al., 1999: 608; Huddleston and Pullum et al., 2002: 1034). The other major component is the relativizer, which can be overt, such as relative pronouns or adverbs that explicitly mark the function of the relativized NP, or covert, in which case we have a particle that does not overtly signal the function of the gap in the relative clause (e.g. that or the zero relativizer in English).3 In relative 154 Alicante Journal of English Studies constructions, a dependency is established between the gap (or the word that subcategorizes for it in the clause) and the head NP, sometimes mediated by an overt relative pronoun. There is experimental evidence which shows that relative clauses, and other structures that involve a dependency between a gap and an antecedent NP, are difficult structures to process (cf. Hawkins, 1999, and references therein): as soon as a relative construction is encountered, the addressee has to store in memory all the information related to the antecedent until the gap is located, and at the same time the words and syntactic/semantic dependencies in the way from the antecedent to the gap must be correctly parsed. Therefore, as argued by Hawkins (1999), processing relative clauses becomes easier the simpler they are (i.e., the fewer phrasal nodes and levels of hierarchical organization they have) and the shorter the distance between the antecedent and the gap, because this means that the head NP has to be kept in memory for a shorter time and that there are fewer additional words and syntactic/semantic operations that have to be processed simultaneously with gap identification. Keenan and Comrie (1977, 1979), on the basis of typological data from a sample of about fifty languages, postulate that NPs are not all equally relativizable. They propose a so-called Accessibility Hierarchy (Keenan and Comrie, 1977: 66), which, as they suggest, represents the ease with which different NP positions can be relativized: (4) The Accessibility Hierarchy Subject (SU) > Direct Object (DO)> Indirect Object (IO) > Prepositional Complement (PCOMP) > Genitive (GEN) > Object of Comparison (OC) The ease of relativizing a NP decreases down the hierarchy, with SU NPs being the easiest and OCs the hardest. According to Keenan and Comrie (1977: 67), languages differ with respect to which positions in the hierarchy they relativize, but this variation is constrained: if a NP position can be relativized, then all positions higher in the hierarchy must also be relativizable, but not the other way around. This means that, for instance, if a language allows relativization on IOs, then it must also allow it on DOs and SUs, but not necessarily on all other positions further down the hierarchy. Keenan and Comrie (1977: 88) also propose an explanation for the hierarchy in terms of processing difficulty: comprehension becomes more difficult as we go down the hierarchy, with relatives formed on lower positions being more difficult to process than those formed on higher ones. Hawkins (1994, 1999, 2004) suggests that relative clause complexity increases as we go down the Accessibility Hierarchy, with more phrasal nodes in the way from the head NP to the gap. As Hawkins (1999: 255) puts it, as the nodes increase there are more structural relations and cooccurrence requirements to compute and more morphosyntactic and semantic operations that apply, such as case assignment, θ-role assignment and thematic dependency computations. There will also be more terminal nodes to process in larger domains, with more words to recognize and process phonologically and morphologically. Syntactic complexity and language contact 155 It has to be borne in mind that all these processing operations take place simultaneously with the resolution of the head NP-gap dependency, which, as mentioned before, is a procedure that already burdens the human processor substantially. The order of the NP positions in the hierarchy can, therefore, be explained in the following terms (Hawkins, 1999: 253-254): • A relativized SU is closer to the head NP than any other relativized positions, so it is the easiest to relativize. • A relativized DO is separated from the head NP by at least a verb and a subject. • A relativized IO presupposes the existence of a direct object and a subject in addition to the verb of the clause, and is thus separated from the head NP by at least those three elements. • A relativized PCOMP is embedded in a prepositional phrase, and set apart from the head NP by at least the preposition, the verb, and the subject.4 • A relativized GEN NP is part of a possessive phrase, which creates extra syntactic depth and puts this position at the bottom of the hierarchy.5 Diessel and Tomasello (2005) and Diessel (2009) propose some further refinements to the Accessibility Hierarchy. Drawing on data from the domain of L1 acquisition,6 they found that English-speaking children have more problems with transitive SU relative clauses (SU-TR) than with intransitive ones (SU-INT), a finding that they explain in terms of the additional referent that SU-TR relatives contain, i.e., a subject and an object “engaged in a transitive activity” (Diessel and Tomasello, 2005: 900). Furthermore, all SU relatives (both intransitive and transitive) were easier for children than DO, IO, PCOMP, and GEN relative clauses. Diessel and Tomasello (2005: 899) argue that this is because they are similar to simple sentences in that the initial NP, i.e., the head NP, is the one that expresses the subject, whereas in all other relative clauses the subject is expressed by another NP that is different from the head NP. DO, IO, and PCOMP relatives performed similarly: they all caused more problems than SU relatives, but were not different from one another, which, as suggested by Diessel and Tomasello (2005: 901), is due to the fact that they have a comparable structure with the same sequence of nouns and verbs, i.e., first the head NP, and then the subject NP, followed by the verb and the gap (NP [NP V …]REL). Finally, GEN relative clauses were the most difficult for children because they are different from the rest: they “establish the link between the head noun and the relative clause by a genitive attribute, which even many adult speakers find difficult to process” (Diessel and Tomasello, 2005: 901). 2.2. Complexity and language contact The other important goal of the present article, as stated in its title, is to analyse the effects of contact on language variation. Language contact (and concomitant L2 acquisition) has been shown to have a simplifying effect on grammar (cf., for instance, the collection of papers in Miestamo et al., 2008). As argued by Trudgill (2009, 2011), among others, this 156 Alicante Journal of English Studies is “due to the relative inability of adult humans to learn new languages perfectly” (Trudgill, 2009: 99): in the process of learning a language, adults simplify its grammar by decreasing its redundancy and opacity, regularizing paradigms, and eliminating semantic distinctions coded by different morphological categories. Many different studies have provided evidence to support the claim that language contact results in grammatical simplification due to the influence of L2 acquisition and use (McWhorter, 2001, 2007; Kusters, 2003, 2008; Parkvall, 2008; Sinnemäki, 2009). However, not all types of contact have this effect. Trudgill (2011) proposes a typology of contact situations and their influence: • High-contact situations with short-term adult L2 acquisition result in grammatical simplification. • High-contact situations involving long-term childhood bilingualism tend to lead to complexification due to “additive borrowing” (Trudgill, 2011: 27; italics in original), i.e., the incorporation of new features into the grammar derived from another language that coexist with the existing features, without substituting any of them. • Low-contact situations result in complexification due to a spontaneous (i.e. nonborrowed) increase in morphological categories, redundancy, opacity, and irregularity. Therefore, it is only in the first type of contact situations, i.e., those in which there is a high number of L2 users, that grammatical simplification takes place. In the domain of varieties of English, previous research has demonstrated that those varieties with a history of language contact are less complex than low-contact ones. A series of studies conducted at the Freiburg Institute for Advanced Studies (Kortmann and Szmrecsanyi, 2009, 2011; Szmrecsanyi and Kortmann 2009a, 2009b; cf. Szmrecsanyi and Kortmann (2012) for a summary of these studies) revealed complexity differences between high- and low-contact varieties of English. High-contact varieties, i.e., highcontact L1s, indigenized L2s, pidgins, and creoles, have, in general, fewer grammatical features that “add contrasts, distinctions, or asymmetries without providing a communicative or functional bonus” (Szmrecsanyi and Kortmann 2012: 16) than lowcontact varieties. They also present more features that result in fewer contrasts, distinctions, and asymmetries, which are easier to acquire and use for L2 speakers. In addition, L2 varieties show a lower degree of grammaticity, i.e., a lower frequency of grammatical markers, and less irregularity than L1 varieties. IndE is a L2 variety of English and an example of Trudgill’s (2011) first type of contact situations, namely those characterized by short-term adult L2 acquisition and use (§ 3.1). Therefore, it is expected to display grammatical simplification vis-à-vis BrE. Finally, and more relevantly for the purposes of the present paper, simplification effects due to language contact can also be found in the domain of relative clauses. As shown in a recent article by Suárez-Gómez (2017), IndE, Singapore English (SgE) and Hong Kong English (HKE), all of them L2 Asian varieties of English, favour simpler relative structures more strongly than BrE. In Suárez-Gómez’s study, relative clauses are Syntactic complexity and language contact 157 more complex if they decrease transparency by containing elements that are not overtly expressed (i.e. zero relativizers) or that increase redundancy, for instance, via agreement (i.e. wh-pronouns), and by introducing discontinuities in the structure (i.e. relatives that are not adjacent to the head NP that they modify). Moreover, IndE, SgE and HKE speakers relativize the higher positions in Keenan and Comrie’s Accessibility Hierarchy more frequently than BrE users, thus producing more SU relatives and fewer DO and, especially, PCOMP and GEN relatives than in the L1 variety. As mentioned in section 2.1, the difficulty of relativizing a NP increases as we move further down the hierarchy, so SU relatives are easier to process than DO, PCOMP and GEN ones. The literature reviewed in this section points to the conclusion that simplification dominates in those languages or dialects affected by language contact and short-term adult L2 acquisition, and that it also affects the domain of relative clauses. The following section focuses on this grammatical domain in the two varieties of English that are at the center of the present paper, i.e., IndE, an L2, and BrE, an L1, as well as on a hypothesis related to the distribution of relativization strategies in the two varieties. 3. Relative clauses in English As mentioned in section 2.1, relativizers can be overt or covert. In English, overt relativizers consist of wh-pronouns,7 i.e., who, whom, whose, and which, and covert relativizers include that and zero (Huddleston and Pullum et al., 2002: 1034). Whpronouns are more commonly found in formal written contexts (although who is also frequently used in spoken language, mostly in SU relatives with animate antecedents; cf. Cheshire, Adger and Fox, 2013), while that and zero are more frequent in speech and informal texts (Biber et al., 1999: 612). Relative clauses can also be classified with respect to the relation between the relative clause and the head NP into restrictive and nonrestrictive: restrictive relatives delimit the set of entities that the head NP can refer to, while the nonrestrictive type merely adds additional information about the antecedent without limiting the set of possible referents (Huddleston and Pullum et al., 2002: 1034- 1035; Denison and Hundt, 2013: 140). The choice of relativizer in English is determined, among other factors, by the animacy of the antecedent, the relation between the relative clause and the head NP, and the function of the gap in the clause (Quirk et al, 1985: 1247-1248; Biber et al., 1999: 609). In nonrestrictive relatives, zero is not possible and that is very infrequent; there is variation between who/whom/whose, used mostly with human antecedents, and which, which is mainly restricted to nonhuman referents (Quirk et al., 1985: 1258). In restrictive relatives, on the other hand, there is more competition between the different relativizers (Quirk et al., 1985:1249-1252; Biber et al., 1999: 612-620; Huddleston and Pullum et al., 2002: 1054-1056): • With human antecedents, who is favoured in SU relatives (although that is also frequent) and that/zero in the rest. Whom and whose are restricted to DO/PCOMP and GEN relatives respectively. 164 Alicante Journal of English Studies • the numbers in boldface represent configurations that are statistically significant: the ones followed by ‘T’ are types, and those with an ‘A’ are antitypes (see section 4). Table 2 does not provide percentages because of the complexity of the data: there are four variables that are hierarchically organized, and three of them have more than two levels. With this kind of data, it is very difficult to decide out of which total the percentages should be calculated. For instance, taking as an example the number of whpronouns in SU position in spoken informal texts in restrictive relatives, i.e., 32, we could calculate the following percentages, among others, depending on the focus of the study: • 41.56% out of the total instances of SU relatives in spoken informal texts (77), • 15.24% out of the total instances of wh-pronouns in restrictive relatives (210), • 8.62% out of the total instances of wh-pronouns in both restrictive and nonrestrictive clauses (371), • 6.72% out of the total instances of restrictive relatives (476), • 19.51% out of the total instances of relative clauses in spoken informal texts (164). Since what is of interest here is the general distribution of relativizers in BrE as a function of each and all of the other variables in Table 2, the data can be better summarized in graphical form. Figure 1 is a visual representation of Table 2. Syntactic complexity and language contact 165 Figure 1: Relativizer choice in BrE 166 Alicante Journal of English Studies Figure 1 plots the proportion of cases of each configuration in relation to the rest, on the basis of the four variables selected: restrictiveness, text type, relativizer, and relativized position. The first division in Figure 1 is that between restrictive and nonrestrictive relative clauses (vertical axis), so that each of the two largest boxes represents the proportion of restrictive and nonrestrictive relatives. Within each of them, we first have a division into text types, and then into relativized positions (horizontal axis): there are four rows that represent the four text types distinguished here and, within each of them, there are four smaller rows, one per relativized position. On the vertical axis there is another division, that between wh-pronouns, that, and zero relativizers. The size of each of the partitions in the plot reflects the proportion of cases of a specific configuration in comparison with the rest. For instance, the black box on the top left corner of Figure 1 represents the proportion of restrictive relative clauses occurring in spoken informal texts with a wh-pronoun as a relativizer and with a SU relativized position. The distribution presented in Table 2 and Figure 1 is statistically significant (χ2 = 2010.34, d.f. = 181, p < 0.001). The picture that emerges from this distribution is a complex one. In restrictive clauses, that is favoured in BrE in spoken (both informal and formal) texts, since it is the most frequent option: it is in fact a type in SU, DO, and PCOMP positions in spoken informal texts. That occurs more often in SU and DO positions, although it is not an infrequent choice in PCOMP relatives (especially in spoken informal texts). Zero is the most frequent relativizer in restrictive relatives in written informal texts. It is favoured in DO and PCOMP positions and strongly disfavoured, or even forbidden, in SU position:13 It is a type in PCOMP relatives in spoken informal texts and in DO/PCOMP in written informal ones, and an antitype in SU in spoken informal/formal and written formal texts. Wh-pronouns are favoured in restrictive relatives in written formal texts (type in PCOMP position in written formal texts), and disfavoured in the other three text types (antitype in DO in spoken informal texts, in SU/DO/PCOMP positions in spoken formal ones, and in DO/PCOMP relatives in written informal texts). They occur more commonly overall in SU position, although they are also very frequent in PCOMP position in written formal texts, where they are a type. Wh-pronouns are the most common option in GEN relatives, which are very infrequent overall. There is, however, one case with that and a stranded preposition of: (11) […] he is carrying this famous letter [that the whole world is waiting to see the contents of] (ICE-GB:S2A-008 #134:3:A) In nonrestrictive relative clauses, the only available option in BrE are wh-pronouns, since there are no instances of that or zero in the data. They occur more commonly in SU position and in spoken formal texts, where they are a type. As regards IndE, Table 3 and Figure 2 show the distribution of relativizers in this variety (χ2 = 2010.34, d.f. = 181, p < 0.001):14 Syntactic complexity and language contact 167 RESTRICTIVE NONRESTRICTIVE wh-pro that zero wh-pro that zero SP. INF. SU 22 0 A 0 9 0 0 DO 5 4 7 2 0 0 PCOMP 3 0 7 0 0 0 GEN 0 0 0 0 0 0 SP. FOR. SU 21 12 1 55 T 1 0 DO 3 12 T 8 7 0 0 PCOMP 3 3 6 1 0 0 GEN 1 0 0 0 0 0 WR. INF. SU 17 4 0 31 T 2 0 DO 3 6 26 T 6 0 0 PCOMP 4 1 4 6 0 0 GEN 0 0 0 0 0 0 WR. FOR. SU 63 12 0 A 31 1 0 DO 7 5 6 1 1 0 PCOMP 21 0 5 5 0 0 GEN 2 0 0 1 0 0 TOTAL 304 160 Table 3: Relativizer choice in IndE 168 Alicante Journal of English Studies Figure 2: Relativizer choice in IndE Syntactic complexity and language contact 169 In IndE restrictive relative clauses, wh-pronouns are preferred in spoken informal and written formal texts. They are more frequent in SU position, although no statistical types or antitypes were found by the HCFA test in this case. There is competition between whpronouns and that in spoken formal texts, with a similar number of cases of both relativizers. That is disfavoured in spoken informal texts, with only 4 cases attested in the corpus (it is an antitype in SU position in this text type), and it occurs more frequently in DO position overall, becoming a type in spoken formal texts. There is also competition in written informal texts, in this case between wh-pronouns and zero, though the latter is more frequent. Zero is preferred in DO and PCOMP positions (it is a type in DO position in written informal texts), and it is very infrequent in SU relatives15 (it is an antitype in SU position in written formal texts). In IndE nonrestrictive relative clauses, wh-pronouns are the default choice. They occur more commonly in SU position and in spoken formal and written formal texts, where they are types.16 With respect to preposition placement in PCOMP relatives, Table 4 and Figure 3 show the distribution of the three different strategies, pied-piping, stranding, and deletion, as a function of restrictiveness and text type in both BrE and IndE (χ2 = 197.94, d.f. = 40, p < 0.001): BRITISH ENGLISH INDIAN ENGLISH Restrictive Nonrestrictive Restrictive Nonrestrictive SP. INF. Pied-piping 3 A 2 2 0 Stranding 14 5 0 0 Deletion 15 T 0 5 0 SP. FOR. Pied-piping 1 0 2 1 Stranding 2 1 3 0 Deletion 0 0 7 T 0 WR. INF. Pied-piping 0 A 2 4 6 T Stranding 14 1 4 0 Deletion 12 0 1 0 WR. FOR. Pied-piping 33 T 4 21 T 5 Stranding 1 A 0 0 0 Deletion 0 A 0 3 0 TOTAL 110 64 Table 4: Preposition placement in BrE and IndE 170 Alicante Journal of English Studies Figure 3: Preposition placement in BrE and IndE Syntactic complexity and language contact 171 In BrE, pied-piping is favoured in restrictive relatives in written formal texts, where it is a type, and disfavoured (antitype) in spoken informal and written informal texts. Preposition stranding is more common in restrictive clauses in spoken informal and written informal texts, although in this case it is not a statistically significant type, and it is infrequent in written formal ones, in which it becomes an antitype. Similarly, deletion is preferred in restrictive relative clauses in spoken informal (where it is a type) and written informal texts, and disfavoured in written formal ones (where it is an antitype). There are not many cases of nonrestrictive PCOMP relatives, so statistical significant results were not found, but a preference for pied-piping can be observed in written (both formal and informal) texts, while stranding is more common in spoken (again, both formal and informal) ones. In IndE, pied-piping is more common in restrictive clauses in written formal texts, in which it is a type, and it is in competition with stranding in written informal ones. In spoken (both informal and formal) texts, on the other hand it is disfavoured, although it is not a statistically significant antitype. Stranding is only frequent in IndE restrictive relatives in written informal texts, where, as just mentioned, it is in competition with piedpiping. Deletion is more common in spoken texts, being a type in formal ones, and disfavoured in written (both informal and formal) texts. With respect to nonrestrictive relative clauses, pied-piping is the only variant attested in IndE in the data, and it is especially frequent in written informal texts, where it is a type. 5.2. Complexity effects Tables 5 and 6 show the individual effects of relativized position (χ2 = 5.95, d.f. = 4, p > 0.05) and relative clause complexity (χ2 = 6.34, d.f. = 1, p < 0.05) respectively. BRITISH ENGLISH INDIAN ENGLISH SU 233 (48.95%) 152 (50%) SU-INT 97 (20.38%) 82 (26.97%) SU-TR 136 (28.57%) 70 (23.03%) DO 142 (29.83%) 92 (30.26%) PCOMP 95 (19.96%) 57 (18.75%) GEN 6 (1.26%) 3 (0.99%) TOTAL 476 (100%) 304 (100%) Table 5: Frequency of relative clauses per relativized position in BrE and IndE17 172 Alicante Journal of English Studies BRITISH ENGLISH INDIAN ENGLISH SIMPLE 332 (69.75%) 237 (77.96%) COMPLEX 144 (30.25%) 67 (22.04%) TOTAL 476 (100%) 304 (100%) Table 6: Frequency of simple and complex relative clauses in BrE and IndE Even though the global distribution of simple and complex relative clauses in BrE and IndE is statistically significant, we do not find any significant types or antitypes with respect to the variables relativized position and relative clause complexity, which means that they do not have individual effects on the distribution in the varieties at hand. We find that, in both varieties, there are more instances of relative clauses in SU position (SU-INT + SU-TR), then in DO, PCOMP, and, finally, in GEN position. In IndE, there are also more cases of SU-INT than SU-TR relatives, whereas BrE shows the opposite distribution: SU-TR > SU-INT. With respect to relative clause complexity, simple relatives are much more frequent than complex ones. Statistically significant results are found when we investigate the conjoined effect of both variables. Table 7 and Figure 4 show the interaction of relativized position and relative clause complexity in BrE and IndE (χ2 = 36.05, d.f. = 13, p < 0.001). BRITISH ENGLISH INDIAN ENGLISH Simple Complex Simple Complex SU 149 (31.31%) 84 (17.64%) 112 (36.84%) 40 (13.16%) SU-INT 65 (13.66%) 32 (6.72%) 61 (20.06%) 21 (6.91%) SU-TR 84 (17.65%) 52 T (10.92%) 51 (16.78%) 19 (6.25%) DO 107 (22.48%) 35 (7.35%) 69 (22.70%) 23 (7.56%) PCOMP 74 (15.55%) 21 (4.41%) 53 (17.43%) 4 A (1.32%) GEN 2 (0.42%) 4 (0.84%) 3 (0.99%) 0 (0%) TOTAL 476 (100%) 304 (100%) Table 7: Conjoined effect of relativized position and relative clause complexity in BrE and IndE Syntactic complexity and language contact 173 Figure 4: Conjoined effect of relativized position and relative clause complexity in BrE and IndE 180 Alicante Journal of English Studies Dahl, Östen (2004): The Growth and Maintenance of Linguistic Complexity. Amsterdam: John Benjamins. Dahl, Östen (2009): “Testing the assumptions of complexity invariance: The case of Elfdalian and Swedish”. In G. Sampson, D. Gil and P. Trudgill, ed., Language Complexity as an Evolving Variable. Oxford: Oxford University Press, 50-63. Denison, David and Marianne Hundt (2013): “Defining relatives”. Journal of English Linguistics, 41(2): 135-167. Diessel, Holger (2009): “On the role of frequency and similarity in the acquisition of subject and non-subject relative clauses”. In T. Givón and M. Shibatani, ed., Syntactic Complexity: Diachrony, Acquisition, Neuro-cognition, Evolution. Amsterdam: John Benjamins, 251-276. Diessel, Holger and Michael Tomasello (2005): “A new look at the acquisition of relative clauses”. Language, 81(1): 1-25. Givón, Talmy (2009): The Genesis of Syntactic Complexity: Diachrony, Ontogeny, Neuro- Cognition, Evolution. Amsterdam: John Benjamins. Gries, Stefan Th. (2004): HCFA 3.2. A program for R. Available at: <http://www.linguistics.ucsb.edu/faculty/stgries/>. Gut, Ulrike and Lilian Coronel (2012): “Relatives worldwide”. In M. Hundt and U. Gut, ed., Mapping Unity and Diversity World-Wide: Corpus-based Studies of New Englishes. Amsterdam: John Benjamins, 215-241. Hawkins, John A. (1994): A Performance Theory of Order and Constituency. Cambridge: Cambridge University Press. Hawkins, John A. (1999): “Processing complexity and filler-gap dependencies across grammars”. Language, 75(2): 244-285. Hawkins, John A. (2004): Efficiency and Complexity in Grammars. Oxford: Oxford University Press. Hilpert, Martin (2013): Constructional Change in English: Developments in Allomorphy, Word Formation, and Syntax. Cambridge: Cambridge University Press. Hofmann, Thomas (2005): “Variable vs. categorical effects: Preposition pied piping and stranding in British English relative clauses”. Journal of English Linguistics, 33(3): 257-297. Huddleston, Rodney, Geoffrey K. Pullum, Laurie Bauer, Betty Birner, Ted Briscoe, Peter Collins, David Denison, David Lee, Anita Mittwoch, Geoffrey Nunberg, Frank Palmer, John Payne, Peter Peterson, Lesley Stirling and Gregory Ward (2002): The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Izumi, Shinichi (2003): “Processing difficulty in comprehension and production of relative Clauses by learners of English as a second language”. Language Learning, 53(2): 285-323. Karlsson, Fred (2009): “Origin and maintenance of clausal embedding complexity”. In G. Sampson, D. Gil and P. Trudgill, ed., Language Complexity as an Evolving Variable. Oxford: Oxford University Press, 192-202. Keenan, Edward and Bernard Comrie (1977): “Noun phrase accessibility and universal grammar”. Linguistic Inquiry, 8(1): 63-99. Keenan, Edward and Bernard Comrie (1979): “Data on the noun phrase accessibility hierarchy”. Language, 55(2): 333-351. Kortmann, Bernd and Benedikt Szmrecsanyi (2009): “World Englishes between simplification and complexification”. In L. Siebers and T. Hoffmann, ed., World Englishes – Problems, Properties and Prospects: Selected Papers from the 13th IAWE Conference. Amsterdam: John Benjamins, 265-285. Syntactic complexity and language contact 181 Kortmann, Bernd and Benedikt Szmrecsanyi (2011): “Parameters of morphosyntactic variation in World Englishes: Prospects and limitations of searching for universals”. In P. Siemund, ed., Linguistic Universals and Language Variation. Berlin/New York: Mouton de Gruyter, 264- 290. Kusters, Wouter (2003): Linguistic Complexity: The Influence of Social Change on Verbal Inflection. Utrecht: LOT. Kusters, Wouter (2008): “Complexity in linguistic theory, language learning and language change”. In M. Miestamo, K. Sinnemäki and F. Karlsson, ed., Language Complexity: Typology, Contact, Change. Amstredam: John Benjamins, 3-22. Maas, Utz (2009): “Orality versus literacy as a dimension of complexity”. In G. Sampson, D. Gil and P. Trudgill, ed., Language Complexity as an Evolving Variable. Oxford: Oxford University Press, 164-177. Matras, Yaron (2009): Language Contact. Cambridge: Cambridge University Press. McWhorter, John H. (2001): “The world’s simplest grammars are creole grammars”. Linguistic Typology, 5(2-3): 125-166. McWhorter, John H. (2007): Language Interrupted: Signs of Non-native Acquisition in Standard Language Grammar. Oxford: Oxford University Press. Miestamo, Matti (2008): “Grammatical complexity in a cross-linguistic perspective”. In M. Miestamo, K. Sinnemäki and F. Karlsson, ed., Language Complexity: Typology, Contact, Change. Amstredam: John Benjamins, 23-41. Miestamo, Matti; Kaius Sinnemäki, and Fred Karlsson (2008): Language Complexity: Typology, Contact, Change. Amsterdam: John Benjamins. Mukherjee, Joybrato (2007): “Steady states in the evolution of New Englishes: Present-day Indian English as an equilibrium”. Journal of English Linguistics, 35(2): 157-187. Mukherjee, Joybrato and Stefan Gries (2009): “Collostructional nativization in World Englishes: Verb-construction associations in the International Corpus of English”. English World-Wide, 30(1): 27-51. Parkvall, Mikael (2008): “The simplicity of creoles in a cross-linguistic perspective”. In M. Miestamo, K. Sinnemäki and F. Karlsson, ed., Language Complexity: Typology, Contact, Change. Amstredam: John Benjamins, 265-285. R Core Development Team (2015): R: A Language and Environment for Statistical Computing. Viena: The R Foundation for Statistical Computing. URL: <http://www.R-project.org>. Quirk, Randolph (1957): “Relative clauses in educated spoken English”. English Studies, 38(1- 6): 97-109. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik (1985): A Comprehensive Grammar of the English Language. Essex: Longman. Sharma, Devyani (2010): “Indian English”. In B. Kortmann and K. Lunkenheimer, eds., The Mouton World Atlas of Variation in English. Berlin/New York: Mouton de Gruyter, 523-530. Schneider, Edgar (2007): Postcolonial English: Varieties around the World. Cambridge: Cambridge University Press. Schröter, Verena and Bernd Kortmann (2016): “Pronoun deletion in Hong Kong English and colloquial Singaporean English”. World Englishes, 35(2): 221-241. Sinnemäki, Kaius (2009): “Complexity in core argument marking and population size”. In G. Sampson, D. Gil and P. Trudgill, eds., Language Complexity as an Evolving Variable. Oxford: Oxford University Press, 126-140. Suárez-Gómez, Cristina (2014): “Relative clauses in Asian Englishes”. Journal of English Linguistics, 42(3): 245-268. 182 Alicante Journal of English Studies Suárez-Gómez, Cristina (2017): “Transparency and language contact in the nativization of relative clauses in World English”. English World-Wide, 38(2): 211-237. Szmrecsanyi, Benedikt (2004): “On operationalizing syntactic complexity”. In G. Purnelle, C. Fairon and A. Dister, ed., Le poids des mots. Proceedings of the 7th International Conference on Textual Data Statistical Analysis. Louvain-la-Neuve, March 10-12, 2004. Louvain-la- Neuve: Presses Universitaires de Louvain, 1032-1039. Szmrecsanyi, Benedikt and Bernd Kortmann (2009a): “The morphosyntax of varieties of English worldwide: A quantitative perspective”. Lingua, 119(11): 1643-1663. Szmrecsanyi, Benedikt and Bernd Kortmann (2009b): “Between simplification and complexification: Non-standard varieties of English around the world”. In G. Sampson, D. Gil and P. Trudgill, ed., Language Complexity as an Evolving Variable. Oxford: Oxford University Press, 64-79. Szmrecsanyi, Benedikt and Bernd Kortmann (2012): “Introduction: linguistic complexity – Second language acquisition, indigenization, contact”. In B. Kortmann and B. Szmrecsanyi, ed., Linguistic Complexity: Second Language Acquisition, Indigenization, Contact. Berlin/Boston: Walter de Gruyter, 6-34. Thomason, Sarah and Terrence Kaufmann (1988): Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Tottie, Gunnel (1997): “Relatively speaking: Relative marker usage in the British National Corpus”. In T. Nevalainen and L. Kahlas-Tarkka, ed., To Explain the Present: Studies in the Changing English Language in Honour of Matti Rissanen. Helsinki: Société Néophilologique, 465-481. Trudgill, Peter (2009): “Sociolinguistic yypology and complexification”. In G. Sampson, D. Gil and P. Trudgill, ed., Language Complexity as an Evolving Variable. Oxford: Oxford University Press, 98-109. Trudgill, Peter (2011): Sociolinguistic Typology: Social Determinants of Linguistic Complexity, Oxford: Oxford University Press. Valdés, Guadalupe (2012): “Multilingualism”. Linguistic Society of America. Available at: <http://www.linguisticsociety.org/resource/multilingualism> [accessed 24 November 2017]. Wasow, Thomas (1997): “Remarks on grammatical weight”. Language Variation and Change, 9(1): 81-105. Wasow, Thomas and Jennifer Arnold (2003): “Post-verbal constituent ordering in English”. In G. Rohdenburg and B. Mondorf, ed., Determinants of Grammatical Variation in English. Berlin/New York: Mouton de Gruyter, 119-154.