On the Use of the Split Infinitive in the Asian Varieties of English

A split infinitive consists of a particular type of syntactic tmesis in which a word or phrase, especially an adverb, occurs between the in finitive marker to and the verb. The earliest instances of the split infinitive in Engli sh date back to the 13th century, in which a personal pronoun, an adverb or two or more words could appear in such environments (Visser 1963-1973, II: 1038-1045). Eve n though its use dropped throughout the 16th and the 17th centuries, it bega n to gain ground again from the 18th century, resisting the severe criticisms of gr ammarians from the first half of the 19th century (Calle-Martín and Miranda-García 2009: 347-364; Perales-Escudero 2011: 316-319). Given the historical concerns about the constructio n, this paper analyses the attitudes towards the split infinitive in the Asian varieties of English, taking the British English practice as a point of departure. The paper has then been conceived with the following objectives: a) to compare the distributio n f the construction in British English and some varieties of Asian Englishes; and b) to explore the phenomenon from a variationist perspective, considering any li kely variation across speech and writing and across the spoken and written registers . The corpus used as a source of evidence is the International Corpus of English , both the British English and the Asian English components (i.e. India, Hong Kong, Singapor e and The Philippines).


Introduction
A split infinitive is defined as a type of tmesis in which a word or phrase, especially an adverb, occurs between the infinitive marker to and the verb.Different terms have been used to refer to this particular ordering of English, such as spiked adverb or cleft infinitive, although the term split infinitive has eventually superseded all its predecessors (Smith 1959: 270).
On historical grounds, the origin of the split infinitive is generally associated with the new finite verb order of Middle English, according to which the adverb tended to appear before finite verbs, eliminating all possibility of ambiguity in the position of adverbials. 2 The early instances of the construction date back to the 13th century, where a pronominal, an adverb or even two or more words could appear in such environments (Calle-Martín 2015 forthcoming; Visser 1963-1973II: 1038-1045).After its rise in Middle English, the split infinitive is found to have a sporadic use until the second half of the 19th century.In Calle-Martín and Miranda-García's historical analysis of the construction, the split infinitive is documented with a rate of 6.85 occurrences (every 10,000 sentences) in the historical period 1640-1850, a fact which corroborates the constrained diffusion of the phenomenon until the year 1850 (2009: 350-351; also Burchfield and Fowler 1996: 737;Mitrasca 2009: 101).The definite rise of the construction took place from the second half of the 19th century, resisting the severe criticisms of grammarians on the grounds of a) the prescriptivist objection to its alleged lack of prestige (Crystal 1984: 27-28); and b) the impossibility of such splitting in other languages, either Classical or Germanic (Crystal 1985: 16).
The pros and cons of the split infinitive have been largely discussed over the last one hundred years, especially from the point of view of its ban in contemporary usage (Close 1987: 217-229;Fischer 2007: 262-267).In a recent publication, Perales-Escudero has traced the history of its proscription in English proposing to consider it a 19th-century reaction associated with the ideology of Teutonic purity in view of the impossibility of this splitting in languages such as German.The Latin-origin hypothesis is then rejected in the light of his close reading of the sources, insofar as there are not written records proving that the proscription stems from the enforcing Latinate standards (2011: 318-319).Even though a word of caution is still the rule in many contemporary usage guides (Howard 1997: 341;Sorenson 1997: 579;Fowler and Burchfield 1998: 738;Partridge 1999: 309;etc.), the split infinitive has safely managed to withstand the proscription and today its misguidedness is no longer open to debate. 3iven these historical concerns about the phenomenon, the present paper analyses the attitudes towards the split infinitive in some varieties of Asian English, taking the British English practice as a point of departure.The working hypothesis is that the ban against the construction could have also left its imprint in these postcolonial varieties of English, thus hindering the subsequent diffusion of the construction.In the light of this, this paper has been then conceived with the following objectives: a) to compare the distribution of the split infinitive in British English and some of the Asian varieties of English; and b) to explore the phenomenon from a variationist perspective, considering any likely variation across speech and writing and across the spoken and written registers.For the purpose, the use of the split infinitive is examined in some East and south-East Asian varieties of English, in particular the varieties spoken in Hong Kong, India, Singapore and the Philippines.
Despite their parallel developments, South Asian English has been often described as being characterized by unity and diversity (Schilk et al. 2012: 137;Zipp and Bernaisch 2012: 167), creating some tension between the unity of South Asian English and the specific developments of each of the individual varieties.Our main concern here is to evaluate the level of unity or diversity towards the split infinitive in Asian Englishes (AsEs), especially compared with the conservative attitude towards the phenomenon in British English (GB).
The present paper has been organized as follows.After the introduction, section 2 explains the methodology followed and the source of evidence upon which this study is based.Section 3, in turn, deals with the empirical analysis of the corpus data, evaluating the level of variation across the different varieties of English, across speech and writing and across registers.Finally, section 4 presents the conclusions together with some suggestions for further research into the topic.

Data and methodology
The source of evidence is the International Corpus of English (henceforth ICE), consisting of one-million word samples of nativeand official-language national varieties of English worldwide.For the sake of comparison, each ICE component has been compiled with the same rationale in terms of dimension (1 million words with 60% and 40% of speech and writing, respectively), chronology (from 1990), informants (native speakers educated through the medium of English, aged 18 or above) and annotation (textual mark-up, word-class tagging or syntactic parsing). 4The present study relies on the following components of ICE, Great Britain (GB), India (IndE), Singapore (SingE), Hong-Kong (HKE) and The Philippines (PhilE).Table 1 below reproduces the word-count of the source data in all the varieties surveyed.In geographical terms, the ICE components provide us with data from the south and the south-eastern Asian varieties of English, the former comprising Indian English while the latter includes the Englishes of Singapore, Hong Kong and the Philippines.In linguistic terms, on the other hand, the varieties analysed here are all members of the Outer Circle following Kachru's Concentric Circle model of the spread of English (Kachru 1985: 11-36;2005: 13-14;also Crystal 1997: 60-61).
In this model, varieties of English are classified as belonging to the Inner Circle (where English functions as a native language), the Outer Circle (English is not a native language but is historically and governmentally relevant) and the Expanding Circle (English used as a foreign language without any historical and/or governmental domain).
The Inner Circle, on the one hand, is here represented by British English, serving as a control group for the standard British English practice.The Englishes of India, Singapore, Hong Kong and The Philipines, on the other hand, are members of the Outer Circle where English plays an important second language role in these countries.These Asian varieties of English are taken to be stable (Mesthrie 2004: 807).According to Schneider's Dynamic Model, the Englishes of India, Hong Kong and the Philippines are already in phase 3 (nativization), which is "the most vibrant one, the central phase of both cultural and linguistic transformation" (2007: 41).In linguistic terms, "this stage results in the heaviest effects on the restructuring of the English language itself; it is at the heart of the birth of a new, formally distinct Post-Colonial English" (Schneider 2007 44).These three varieties are, however, well advanced in the process of nativization and already moving towards phase 4, the phase of endonormative stabilization (Setter, Wong and Chang 2010: 116).The English of Singapore is, in turn, the most advanced variety with evidence of phase 4 where "the country's unique, territory-based, and multicultural identity construction has paved the way for a general acceptance of the local way of speaking English as a symbolic expression of the pride of Singaporeans in their nation" (Schneider 2007: 160;Seoane and Suárez-Gómez 2013: 5).
AntConc 3.2.4 has been used for the automatic retrieval of the instances (Anthony 2011).The process, however, was not straightforward.First, the complete concordances for the word to were generated.Next, manual disambiguation was needed to weed out the irrelevant prepositional uses and identify tokens of the split infinitive construction, as shown in examples 1-2 below.
(2) Uh I will like to in the next few slides discuss other than this visual aspect and the noise aspect some of the other measures that we take to control the uhm the problem (ICE-HK:S2B-046#140:1:B).
More often than not, however, the separation of the infinitive marker to and the infinitive results from the interpolation of nonlexical noises such as uh/uhm, and discourse-functional lexical expressions such as you know, I mean, like, sort of or kind of, which allow for the speaker to pause while collecting his/her thoughts in the flow of conversation (Fox 2010: 1).These instances have been ruled out on the grounds that they do not serve the same kinds of grammatical functions as an adverbial actually does, as shown in examples 3-5: (3) Uh it helps you to uh develop your application base on some rules (ICE-HK:S2A-059#11:1:A). ( 4) We first have to uhm contact the company and get the application (ICE-HK:S1A-012#X9:1:Z).
(5) Maybe I should get my friends to you know send it to me (ICE-SIN:S1A-039#222:1:A).

The split infinitive across the AsEs varieties
The ICE corpora have provided us with a total of 785 instances of split infinitives, of which 104 belong to GB while the other 681 correspond to AsEs.Table 2 reproduces the number of split infinitives in the corpus (absolute figures and normalized frequencies), which have been classified in terms of a) the language variety and b) their speech and writing variation.For comparison, the figures have been normalized to tokens per million words.
Table 2.The split infinitive in the ICE components (absolute and n.f.) These data show that the split infinitive is more constrained in the British English practice.While the construction amounts to 97.9 instances in GB, it shows 109.9 occurrences in HKE, 112.9 in IndE,  5 This can be explained in the light of the imprint of English in some of these territories.India was under the rule of the British Empire since 1765 until independence in 1947, a nearly 200-year period which eventually derived in its configuration as a second official language in the country together with Hindi (Gargesh 2006: 94).For that reason, in contrast with other Asian varieties of English, "the syntax of Indian English, as opposed to phonology and lexis, is said to conform most to standard British English" (Saijala 2009: 39).On the contrary, the spread of English in the South-East Asian territories is a 20th-century phenomenon, when English managed to become the language of government and the legal system but also with a growing importance in education and the media (Crystal 1984: 57).While IndE seems to be more reluctant to these kinds of changes, the other Asian varieties are found to be freer from this strict ban towards the construction.SingE and PhilE, on the other hand, then show the other side of the coin with a wider diffusion of the split infinitive, amounting to 179.7 and 166.7 occurrences, respectively.The phenomenon is more frequent in SingE than in the other AsEs, plausibly as a result of the status of English in Singapore, considered to be more advanced according to Schneider's Dynamic model, with clear traces of endonormative stabilization (2007: 41).In PhilE, this high proportion of split infinitives can be explained as an influence of the superstratum 5 Indian English is generally reported to be the most conservative variety of Asian Englishes.In their analysis of the levelling between the present perfect and the simple past for the expression of the perfect in Asian Englishes, Seoane and Suárez-Gómez conclude that it is the variety with the highest percentage of present perfect forms, therefore more tightly in the line of the British English practice.The conservatism in this case is reflected in the mildness of the decline of the present perfect vis-á-vis the preterite (Seoane and Suárez-Gómez 2012: 12).language, American English,6 which has been recently reported as using the construction on a frequent basis.Perales-Escudero has investigated the use of the split infinitive in the Corpus of Contemporary American English (COCA), concluding a) that the construction is notably diffused in American English; and b) that the split infinitive is also register-dependent in American English in the sense that some combinations "are much more common in written registers than they are in spoken ones, and much more common in academic registers" (2011: 324-325).7

The split infinitive across speech and writing
Figure 1 below reproduces the distribution of the phenomenon across speech and writing.The split infinitive is observed to predominate in speech-based text types, however, the occurrence across the different varieties is far from uniform, with 113.8 occurrences in HKE, 125.3 in IndE, 138.2 in PhilE, 142.7 in GB and 217.04 in SingE.However, crucial differences arise when speech and writing are taken into consideration.While GB shows the most significant difference between speech and writing (142.7 and 30.6), the phenomenon is found to have a more balanced distribution in IndE (125.3 and 93.4) and HKE (113.8 and 102.2),where a sharp rise is confirmed if compared with the British English practice.SingE, in turn, is found to be one step further in the continuum inasmuch as the split infinitive amounts to 217.04 occurrences in oral-based texts, almost doubling the occurrence of the phenomenon in the written domain (with 121.4 occurrences).PhilE, on the other hand, shows the other side of the coin insofar as the split infinitive is found to be more frequent in writing than speech, amounting to 210.8 and 138.2 occurrences, respectively.
Even though the split infinitive is observed to predominate in speech-based text-types in all the varieties, it is worth noting that all AsEs show a substantial use of the construction in the written domain, especially if compared with the constrained attitude towards the phenomenon in GB.While the split infinitive just amounts to 30.6 in GB, this figure is more than tripled in the Asian varieties with 93.4 in IndE, 102.2 in HKE, 121.4 in SingE, and 210.08 in PhilE.These results again corroborate both the conservative attitude of both IndE and HKE towards the split infinitive and the wider diffusion of the phenomenon in SingE and, more importantly, in PhilE.The figures in SingE and PhilE corroborate that these varieties have already set free from the traditional objections to the split infinitive in GB, showing a more widespread use of the construction even in the written medium.

The split infinitive across registers: dialogues and monologues
The ICE corpora have also been designed to account for any likely variation in the written and the spoken samples.Figure 2 presents the distribution of the split infinitive across the spoken component of ICE, distinguishing whether they occur in dialogues or monologues.For comparison, the figures have been normalized to tokens per million words.The results confirm the same tendency in the different varieties under scrutiny in the sense that the split infinitive predominates in monologues over dialogues, SingE in particular.These figures tentatively confirm the on-going diffusion of the split infinitive in all these varieties, giving room for the construction in monologues, considered to be less spontaneous than face-to-face communication.

Fig. 2. The split infinitive across dialogues and monologues (n.f.)
Dialogues are subdivided in ICE into private and public dialogues.Private dialogues include direct conversations and telephone calls while public dialogues display class lessons, broadcast discussions and interviews, parliamentary debates and business transactions, among others.As reproduced in Figure 3, the split infinitive presents a different use in GB and AsEs.In GB the split infinitive is favoured in private dialogues (with 188.2 and 135.2 occurrences, respectively).The other side of the coin, however, is witnessed in the other AsEs to such an extent that the split infinitive finds more room in public dialogues, as its occurrence in all cases exceeds that of private dialogues.Interestingly enough, the bulk of public dialogues in ICE includes broadcast discussions and parliamentary debates, giving then an idea of the level of diffusion of the split infinitive in AsEs, particularly if compared with the constrained use of the construction in GB.
Monologues are classified into scripted and unscripted.The former display broadcast news together with broadcast and nonbroadcast talks while the unscripted material contains spontaneous commentaries, unscripted speeches, demonstrations and legal presentations, among others.As in the case of dialogues, there are again two different attitudes towards the split infinitive in GB and AsEs.In GB, on the one hand, the use of the phenomenon in scripted monologues is negligible, amounting to 31.3 and 174.7 occurrences in scripted and unscripted monologues, respectively.AsEs, on the other hand, present a substantial diffusion of the construction in scripted monologues, to such an extent that in some cases it outnumbers that of unscripted monologues, HKE and PhilE in particular.This fact confirms the increased diffusion of the phenomenon in AsEs, presenting a parallel use of the construction both in scripted and nonscripted material.In addition to these general tendencies, SingE stands out for the number of split infinitives in unscripted monologues (with 350.2 occurrences), therefore doubling in some cases the figures obtained from the other varieties surveyed.This is plausibly connected with the status of English in Singapore, considered to be more advanced than the others according to Schneider's Dynamic model (2007: 48-52).Already immersed in the phase of endonormative stabilization, our data show how the split infinitive is in an on-going process of diffusion in Singapore and, more importantly, how that process is finding more ground in spontaneous material, unscripted monologues in particular.Fig. 3.The split infinitive in terms of spoken variation (n.f.)

The split infinitive across registers: printed and non-printed material
In this same vein, Figure 4 below reproduces the distribution of the phenomenon in the written component of ICE distinguishing whether it occurs in printed and non-printed material.These data confirm an overwhelming preference for the split infinitive in non-printed material in all the varieties as a result of the spontaneous nature of this textual category.However, the figures also allow us to reach the following conclusions.GB, on the one hand, is again more reluctant to use the construction in printed texts (with 40.7 occurrences) especially if compared with HKE (100.5),SingE (98.09) and PhilE (204.1).SingE and PhilE, on the other hand, again present the highest number of split infinitives, the latter in particular regardless of the printed or non-printed nature of the texts.In ICE, non-printed material includes correspondence (i.e.social and business letters) and non-professional writing (i.e.student essays and examination scripts), the split infinitive predominating in letter writing across the different varieties.Printed material, in turn, consists of the following types of writing, i.e. academic writing, popular writing, instructional writing, persuasive writing, creative writing and reportages.The split infinitive is subjected to a higher level of variation here, mostly preferred in popular writing, academic writing and reportages.Persuasive and creative writing would then be at the bottom of the continuum with a more constrained use of the construction.

Conclusions
The present paper examines the split infinitive in GB and AsEs, paying particular attention to the quantitative dimension of the phenomenon.For the purpose, the study has been based on the International Corpus of English, which has provided us with material for comparative analysis of the East and South-East Asian varieties of English surveyed.The ICE corpus design in terms of dimension, chronology and profile of the informants has ensured the compatibility across the individual corpora, thus becoming the ideal input for investigating a linguistic construction such as the split infinitive.In our case, GB has been taken as the touchstone for comparison to note any likely deviation from the standard British English practice.The data obtained have allowed us to reach the following conclusions.
First, the split infinitive is generally more constrained in British English than in the Asian varieties of English, where a looser attitude towards the construction is confirmed.This trend, however, cannot be equally corroborated in all the varieties surveyed.IndE and HKE, on the one hand, present the lowest number of split infinitives, and they stand out for their moderate use of the construction, remaining still closer to the British English practice.SingE and PhilE, on the other hand, are located at the top of the continuum showing evidence of a more widespread use of the construction, a fact plausibly associated with the spread of English in these territories.The spread took place throughout the 20th-century and, as a result, these varieties plausibly developed a more positive attitude towards the construction, not under the shelter of the 19th-century objections.Within this group, the frequency of the split infinitive is particularly conspicuous in SingE, a fact which is surely justified in the light of the status of English in Singapore, considered to be more advanced according to Schneider's Dynamic model (already in phase 4 -endonormative stabilization).
Second, the split infinitive has also been investigated across speech and writing.Even though the construction is overwhelmingly favoured in speech-based text types in all the varieties of English, this paper reports a sharp increase of the phenomenon in written texts in all AsEs in general, especially if compared with the constrained GB practice, therefore confirming that these post-colonial varieties have already set free from the traditional objections to the split infinitive, showing a substantial diffusion of the phenomenon also in the written medium.Following the previous trend, IndE and HKE are again the most conservative varieties in contrast with SingE and PhilE, the latter in particular with 210.08 instances.This is plausibly associated with the American ascendancy of PhilE, where the split infinitive is confirmed to have gained substantial ground in both speech and writing.
Third, the split infinitive has also been analysed from the perspective of register variation.As for the spoken component of ICE, our study reports an overwhelming preference for the construction in monologues over dialogues in the different language varieties surveyed, despite their less spontaneous nature than face-to-face communication.However, a close examination of the data leads us to postulate a different use of the construction in AsEs, especially in terms of the typology of dialogues and monologues.While the split infinitive is favoured in private dialogues in GB, in AsEs the construction finds more room in public dialogues.In this same vein, while in GB the split infinitive is found negligible in scripted monologues, the other AsEs present a substantial diffusion of the phenomenon in scripted monologues, PhilE in particular.
The written component of ICE also allows the classification of the phenomenon in terms of the printed or non-printed nature of the texts.Our analysis confirms an outstanding preference for the split infinitive in non-printed material in all the varieties as a result of the spontaneous character of this category.However, GB again is observed to be significantly reluctant to use the construction in printed texts (just 40.7 occurrences) in sharp contrast with AsEs where the split infinitive is disseminated irrespective of the printed or the nonprinted nature of the texts, especially in HKE (100.5 occurrences), SingE (98.09) and PhilE (204.1).
Split infinitives are more often than not disregarded in many present-day English grammars as a result of the longstanding influence of the 20th-century prerogatives, the only references being just limited to the inclusion of brief notes about their frequency and their stylistic implications (Thompson and Martinet 1960: 248;Alexander 1988: 305;Huddleston and Pullum 2002: 582).This study has shed light on the on-going diffusion of the phenomenon, not only in British English but also more conspicuously in these post-colonial varieties of English, SingE and PhilE in particular.In our opinion, the traditional tenets published in the literature should be re-examined in view of this quantitative piece of evidence as the construction has gained substantial ground in the last decades.A call is made here for more insight into the topic to gain a wider scope not only synchronically, to explore both regional and sociolinguistic variation, but also diachronically to analyse the origin and development of the construction in Middle English.