The Variations in Verb-Preposition Combinations in the GloWbE Corpus and its Usage in Informal Englishes

This paper is based on the Corpus of Global Web-based English (GloWbE) compiled by Mark Davies in 2013. The GloWbE corpus consists of web data from 20 different English speaking countries. This research is constructed on the GloWbE corpus to investigate the variations in certain verb-preposition combinations in informal Englishes. As the corpus is divided into two sections, such as – general and blog, this study is based on the blog section to compare web data from two inner-circle countries such as The USA and Great Britain and two outer-circle countries such as – India and Bangladesh. The reason for selecting the blog is that, in the GloWbE corpus, the blog section consists of informal data, whereas the general section consists of formal data. It is to be noted that the inner-circle countries use English as their native tongue, whereas the outer-circle countries use English as their second or foreign language. This paper argues that the verb-preposition combinations or prepositional verbs vary in their frequency and meaning in the countries mentioned above. This paper investigates the following five prepositional verbs from The Cambridge Grammar of the English Language proposed by Huddleston and Pullum in 2002– come up with, look out for, put up with, stand up to, and get along with in the GloWbE corpus to prove that the frequency and meaning of these phrases vary from country to country based on social, cultural and political contexts as seen in the results. The investigation shows the trends of the above five prepositional verbs in the four countries explicitly.


Introduction
"The word corpus is Latin for the body (plural corpora). In linguistics, a corpus is a collection of texts (a 'body' of language) stored in an electronic database. Corpora are usually large bodies of machine-readable text containing thousands or millions of words" (McEnery, Hardie and Baker 2006, p. 48). Kennedy (1998, p. 143) observes, " […] corpus-based studies of prepositions reveal that many of them frequently occur in recurring collocations and that systematic analysis of the phrases which prepositions form part of reveals a variety of different semantic roles". The foundation of this paper is based on Davies (2013), Davies (2015), Davies and Fuchs (2015) and Davies (2020). About the GloWbE corpus, Davies (2020) reports, "at the most basic level, researchers can see the frequency of a word or phrase in all 20 countries […]". He also discusses the different functions of 'List', 'Chart', 'Collocates', 'Context', and 'KWIC' (meaning -Key Word In Context) in the GloWbE corpus in English-Corpora.org: a guided tour (cf. 2020).
"Language is inherently variable, both across time (diachronically) and at any specific point in time (synchronically)" (Krug & Schlueter 2013, p. 01). "Computer corpus-based studies of variation in English became possible with the availability of the Brown Corpus from 1964" (Kennedy, 1998, p. 181). Later, due to the development of electronic technology, more and more corpora were built to research different aspects of various languages. Subsequently, the Corpus of Global Web-based English (GloWbE) was compiled and released by Mark Davies in 2013(cf. Davies, 2015. "GloWbE is based on 1.9 billion words in 1.8 million web pages from 20 different English-speaking countries" (Davies and Fuchs 2015, p. 01). According to Davies and Fuchs (2015, p. 01), "Because of its large size, its architecture and interface, the corpus can be used to examine many types of variation among dialects, which might not be possible with other corpora […]". This is the prime reason that the GloWbE corpus has been selected to investigate the variations in verb-preposition combinations in informal Englishes. As "the texts in the corpus consist of informal blogs (about 60% of the corpus) and other web-based materials, such as newspapers, magazines, company websites, and so on" (Davies, 2015), this paper focused on 'Blog' category to analyze the variations. Moreover, as the GloWbE corpus contains texts from "Six Inner Circle and 14 Outer Circle countries" (Davies and Fuchs 2015, p. 02), for this study, two inner-circle countries -The USA and Great Britain and two outer-circle countries -India and Bangladesh have been selected to compare the data.
The reason for selecting these countries are manifoldsfirstly, the inner-circle countries use English as their primary language or native language, whereas the outer-circle countries use English as their second or foreign language (cf. Kachru 1985, p.12); secondly, I assume, that The USA and Great Britain have a large number of data available to compare with the data from India and Bangladesh; thirdly, I also assume that there are socio-cultural differences among the inner circle and the outer-circle countries which may also indicate regional language variation. I argue that the verb-preposition combinations or prepositional verbs vary in their frequency and meaning in the countries above. This paper examines the following prepositional verbs from The Cambridge Grammar of the English Language by Huddleston and Pullum (2002, p. 287)come up with, look out for, put up with, stand up to, and get along with in the GloWbE corpus to prove that the frequency and meaning of these phrases vary from country to country.
In this paper, I analyzed the verb-preposition combinations concerning the GloWbE corpus. To support my arguments, I used related quotations, charts, tables and figures with proper citations. I defined the related topics and described the background of corpora to set the ground for the research. Later, I examined the five verbs with examples of the first ten hits in the four countries. Finally, I discussed the results of the corpus data to demonstrate which countries have more or less usage of the specified verb-preposition combinations. Baldwin (cf. 2005, p. 09) worked with corpus data to find out prepositional verbs and gave a list of '100 most-frequent verbs' and '10 most-frequent transitive prepositions' separately, from the written section of the British National Corpus (BNC). However, he did not list any prepositional phrases or idioms in his study. Gardner and Davies (cf. 2007, p. 358-359) also pointed out the 'Top 100 Phrasal Verb Lemmas in BNC'. Liu (cf. 2011) studied the Corpus of Contemporary American English (COCA) and the British National Corpus (BNC) to find out the most frequent English phrasal verbs and found that 20 out of 30 phrasal verbs are "more common in American English" (p. 671), however, he did not specify any prepositional verbs in his study. Unlike them, I will look into the GloWbE corpus, which is relatively new compared to the BNC, for my selected verb-preposition combinations. In this regard, I have discussed the definitions of a phrasal verb, idiom and prepositional verb below.

Literature Review
Further, "Phrasal verbs also called (idiomatic) multi-word verbs, consist of a verb, an adverb (adverbial particle) and a preposition. Some verbs are called prepositional verbs since they consist of a verb and a preposition" (Ostyn-Rudzka 2003, p. 01). According to Huddleston and Pullum (2002), "The term 'phrasal verb' implies that the combinations concerned form syntactic constituents belonging to the category verb" (2002, p. 274). Besides, phrasal verbs can have multiple, context-sensitive meanings (cf. Gardner and Davies 2007, p. 345). In Dagut and Laufer's (1985) study of English phrasal verbs on Israeli students, they noticed that, out of literal, figurative and completive phrasal verbs, students used figurative phrasal verbs the least (cf. Dagut and Laufer 1985: 74-77), which may indicate that, "the figurative, or idiomatic, phrasal verbs were considered semantically more difficult than other types of phrasal verbs" (Liao and Fukuya 2004: 197). Similarly, it will be interesting to see the frequency of idiomatic verbpreposition combinations in GloWbE corpus.
To approach the further discussion on multi-word verbs, idioms needed to be defined. "An idiom is an expression larger than a word whose meaning cannot be systematically derived from meanings that the parts have when used independently of each other" (Huddleston and Pullum 2002, p. 273). In a verbpreposition combination, Huddleston and Pullum classified preposition as (i) specified and (ii) unspecified and grouped specified preposition in two typesmobile preposition and fixed preposition. Verb and a specified proposition create a verbal idiom, and a fixed verb and preposition combination are called a fossilized combination (cf. Huddleston and Pullum 2002, p. 275, 277).
Moreover, they also pointed out, "[…] the term 'prepositional verb' applies to the sequence of verb + preposition, […]", and "prepositional verbs […] are those which select a PP [Prepositional Phrase] complement containing a specified preposition together with its complement" (2002, p.274). Furthermore, they provided a structural list of six verbpreposition combinations, given below:  (Adapted from Huddleston and Pullum: 2002, p.277) Prep in chart one means preposition, O is for Object, and PC means Predicative Complement. The chart explains the structure of prepositional verbs. As it can be noticed that the verb is always in the initial position, and it can be followed by either prepositions or objects, while the predicative complement only takes prepositions before them. From the chart, it is clear that it be a prepositional verb. The proposition need not be always in the immediate position after the verb.

Method
The methodology in this paper follows a quantitative approach to find out the real-world usage of the five verb-preposition combinations. The data have been collected from the GloWbE corpus, and by explicitly using the 'list', 'chart' and 'context' functions of the website, the examples have been narrowed down. As Davies compiled the corpus in 2013, the examples have not been changed to date. As a result, the usage of the five verb-preposition combinations could be extracted from the online database. First, I looked into the frequency of the verb-preposition combinations of each of the five verbs. Then I compared the frequencies among the four countries to see the variation. After that, I examined whether some prepositional verbs occur more or less in a specific context. As Davies (2020, p. 05) observes, "words do not occur in isolation, and learners need to understand the patterns that a given the word takes".

Corpus Data and Results
For my research, as I have chosen four countries, two from the inner circle (The USA and Great Britain) and two from the outer circle (India and Bangladesh), I would like to begin by showing the number of web sites and web pages in the 'Only Blogs' section of those countries in the GloWbE corpus. From the table above, it is clear that the inner-circle countries have far more web sites and web pages than the outer-circle countries. Therefore, more data will be available to investigate in the US and GB than in the other two countries. Now, if I look into the five prepositional verbs in all the countries (Figure 4.1) in the GloWbE, it is evident that all these verbs occur more in the general section of the corpus than in the blog section. It is logical to assume that, as the general section has more data, the verbs occur more frequently in that section.   Table 4.2. Out of the total raw frequency of 19280 in 20 countries, it occurs more in the US English than in GB, IN or BD English. From table 4.2 (second row), it can also be noticed that the size of the corpus varies from country to country. Out of 1900 million words, the US has 386.8 million, GB has 387.6 million, IN has 96.4 million, and BD has 39.5 million words in total in the blog section. The third row indicates the normalized frequency (NF) per million words. The formula 1 to get the normalized frequency (NF) is - The data from table 4.2 shows, across the four varieties of English, after the normalized frequency, come up with is more frequent in the US and GB English, i.e. in the inner-circle countries, than in IN and BD. Furthermore, if I use the 'Collocates' function of the GloWbE corpus as in figure 4.2 (next page), and select the 'blog' section for the phrase come up with, to search for two words before and two words after, I find a list of words (figure 4.3) that 'hangs out with' (Davies 2020, p   The drawback of this search is that figure 4.3 shows data from all the twenty countries listed in the GloWbE corpus. As it will be a time-consuming task to separate four specific countries for this function, I left it for another research. However, as Davies (2020, p. 06) points out, "Collocates (nearby words) can provide beneficial insight into the meaning and usage of a word or phrase" and "also move beyond strict 'word meaning' to show 'what we are saying' about different topics" (2020, p.08). I looked into the chart function as in figure 4.4 to investigate the first ten sentences or hits in each of the four countries that I am discussing. The search result (first ten hits) for the US (figure 4.5) shows that the combination was primarily used on blogs and in online newspapers. Also, the context was primarily political.  Similarly, the search result (first ten hits) for GB shows that the verb come up with was used in informal contexts, not to describe any political news. The sense in which it is used is similar to the US English, which indicates 'bring forth' (WordNet Search -3.1). In Indian English, the first ten hits on come up with primarily represent dialogues in either festival contexts or job contexts. However, while searching for detailed contexts, some links did not work. In the case of Bangladeshi English, the first ten hits point to banking websites, personal blogs and travel blogs. There was no political context, and from the banking sites, it seemed that the language in which come up with is used is more formal than the US and GB English.
The second verb-preposition combination I searched in GloWbE is look out for. Table 4.3 shows that the phrase look out for was more frequently used in GB and IN English than the US and BD English. As GB has the most significant corpus size (387.6 million), among the four countries, it has the highest number of raw frequency (1165). Normalized frequency is also the highest in GB English with 3.01 per million. Besides, it is evident from table 4.3 that one of the outer-circle countries, i.e. India, has higher usage (per million) of look out for compared to the inner circle countrythe US. 32 Apart from this, I used the chart function in the GloWbE corpus to see the first ten example sentences with a look out for in it. The first ten hits for US English ( figure 4.4) shows various examples from personal, business and sports blogs. The phrase look out has two senses in WordNet Search -3.1 such as -(1) to be vigilant, and (2) to protect someone's interests (cf. http://wordnetweb.princeton.edu), and in the examples, shows both the meaning senses in the US English. For example, in the second hit, look out for indicates to protect the interest of a person's brother, while in the eighth hit, the meaning is to be vigilant in case of autocrats' misuse of technology.

Figure 4.6 First ten examples of look out for in the US English in blog
However, the first ten hits in the search result for GB shows that mostly look out for were used to pay attention to something or search for something. Example sentences are from personal blogs, medical blogs and online magazines. In Indian English, the first ten hits using look out for are from personal blogs and websites and online photography magazine. The ordinary senses of the phrase are to protect a family's interest and search for something (figure 4.7).  In Bangladeshi English, however, there were only 29 example sentences, out of which I looked into the first ten hits. Figure 4.8 shows the first ten hits in BD English, where most websites are personal blogs, except a university website (hit no. 10).

Figure 4.8 First ten hits of look out for in BD English in the blog section
The senses of look out for are mostly to be vigilant and to search for something. However, as BD has far less data on the phrase look out for than the other three countries, it cannot be concluded whether the senses are used only in personal blogs and educational websites.
The third verb-preposition combination that I examined in GloWbE corpus is put up with. Table 4.4 on the following page shows that put up with is most frequent in GB English than the other three countries. The highest normalized frequencies are in GB (2.41), and the US (2.17) with the most hits, and the lowest is in BD (0.61) and IN (0.72). The chart function in GloWbE shows that put up with was mostly used in personal blogs in the US data. The first ten hits in US English (figure 4.9) shows that the prepositional verb put up with was used to mean 'to tolerate' or 'to endure' (WordNet Search -3.1) in all the cases.

Figure 4.9 First ten hits of put up with in US English in the blog section
As GB has the highest raw frequency (934) of put up with among the four countries, the first ten instances (figure 4.10) are from personal blogs, and the collocating words nearby the phrase put up with were primarily used in negative senses, for example, in hit 2, "… put up with all the other crap …", and in hit 5, "… put up with all the insults…" to show the meaning 'to put up with something or somebody unpleasant' (WordNet Search -3.1). The examples show that the phrase put up with was used in negative contexts in the personal blogs. For example, hit 1, "… put up with this shit." and in hit 5, "There are some people whom you hate, but you have to put up with them", all indicate the sense 'to tolerate'.
In BD English, the raw frequency was only 24 in the blog section. The first ten examples using the chart function in GloWbE corpus (figure 4.12) shows the usage of put up with in informal blogs. BD English also shows the negative usage of put up with in the first ten instances. For example, in hit 2, "… put up with their rape and sodomy…", and in hit 3, "… put up with the abuse …".
The fourth verb-preposition combination in my investigation list is stand up. Table 4.5 shows that; the raw frequency is the highest in the US among the four countries in the blog section of the GloWbE corpus. It also shows that the inner-circle countries, e.g. US and GB have more data than the outer-circle countries, e.g. IN and BD. If I investigate the data using the chart function of the GloWbE corpus, I see that the first ten hits in US English (figure 4.13) shows that the phrase stands up to have been used in political contexts the personal blog writers. There are seven senses of stand up in the WordNet Search -3.1. However, in US data, the meaning 'defend against' or 'resist' have been used, for example, in hit 1, "… stand up to EVIL …", and in hit 2, "… stand up to the Taliban…", the meaning 'defend against or 'resist' is appropriate.

Figure 4.13 First ten hits of stand up to in US English in the blog section
If I look into the GB data for the phrase stand up to, the first ten hits using the chart function give in Figure  4.14. The examples show that British English is much more variant than the American one. Although the examples are from informal blogs, the contexts vary from politics to sports to personal experiences.
However, the meaning sense is mostly 'to defend' or 'to resist'. For example, in hit 1, "… stand up to these greedy …", and in hit 2, "… stand up to these money-hungry…" shows the negative context of the phrase stand up to. Figure 4.14 First ten hits of stand up to in GB English in the blog section ISSN (Online): 2685-3906, ISSN (Print): 2685-0818 DOI: https://doi.org/10.31849/reila.6108 Vol. 3, No. 1, April, 2021 37 The first ten hits out of 58 raw data from Indian English in the blog section show that the context varies in the phrase stand up to (figure 4.15).

Figure 4.15 First ten hits of stand up to in IN English in the blog section
The examples are personal blogs, blogs related to movies and songs, newspaper blogs and business blogs. Here, the meaning sense of stand up to is mostly 'to defend against'. For example, in hit 3, "… stand up to great strain …", and in hit 5, "… stand up to punishment …" shows the variety of usage in Indian English.
BD English has only 20 raw data, and I checked the first ten out of those (figure 4.16). Examples indicate stand up was used in websites and personal blogs related to technology, mobile phones and vehicles, and also in one of the newspaper articles. For example, in the first hit, the sample was from a technology blog and in the tenth hit, the sample was from a newspaper website.  Like the previous data collection, I used the chart function to get the first ten hits of get along with in the US English.

Discussion
In the previous section, I collected data for the five prepositional verbscome up with, look out for, put up with, stand up to, and get along with from the 'only blogs' segment of the GloWbE corpus and analyzed those data with related examples to show the variation in results. From the data, it is evident that, in all the cases, the US and GB have more data, i.e. raw frequency of the five verb-preposition combinations, than IN and BD. There may be many reasons for this. For example -Great Britain has a long history of colonization in the US, and later, in the Indian subcontinent. Kachru (cf. 1985, p.12) points out that the spread of English happened because of the "extended periods of colonization, essentially by the users of the inner circle varieties". Thus, the US and GB have a larger corpus of web data than the outercircle countries -India and Bangladesh. Another reason, perhaps, is that India received independence later in 1947 and Bangladesh in 1971, both countries with different native tongues; thus, English did not have the privilege of being the primary language. Moreover, in the case of Bangladesh, Bengali linguistic identity was the primary catalyst to receive independence (cf. Kabir 1987, p.482-484).
In the case of collocations, I found that some verbpreposition combinations are used in some specific contexts. For example, come up with in US data (first ten hits) mainly was used for political contexts, whereas, come up with in Indian English (first ten hits) was mostly used in the festival and cultural contexts. Furthermore, some phrases are exclusively used in negative contexts; for example, put up with in all the four countries, was used in most negative situations when someone has to tolerate something reluctantly. In the case of stand up to, this phrasal verb is also used in political contexts in the US data, while it has multiple usage contexts in GB data, unlike the US.
The research explores the GloWbE database on the prepositional verbs extensively and illustrates the usage in individual countries. As prepositional verbs are part and parcel of everyday conversations, the informal usage with examples shows the trending topics of discussion in the websites in those specific countries. However, future research can improve a similar topic by adding more English speaking countries and, subsequently, adding more data or examples from the corpus. Comparisons can be made with other corpora such as British National Corpus (BNC) etc. Future researches can also show the usage in the general section from the GloWbE corpus to find exciting outcomes.
To the end of the discussion, it can be said that, in the 'only blogs' section of the GloWbE corpus, the five verb-prepositions are used in numerous contexts, and their meanings vary depending on the country, type of socio-political situation, type of culture and also on the type of blog, newspaper, magazine or business company web sites.

Conclusion
English has a broad spectrum of domains in which it is used with varying degrees of competence by members of society, both as an intranational and an international language (Kachru 1985, p. 12). To sum up, it can be said that my assumptions in the introduction were appropriate, as there were more data in the US and GB English compared to the IN and BD, and, also, the five verb-preposition combinations were used in varied contexts in those countries. Although some verbs were in low frequency yet, the fact illustrates less use of those verbs, particularly in the outer-circle countries, which perhaps justifies Liao and Fukuya's (cf. 2004, p. 194) concept that the second language learners of English try to avoid using phrasal verbs. From the corpus data, it was also clear that the verb-preposition combinations vary in their frequency and, to some extent, in meaning in the four countries. However, as the GloWbE has web data from only the years 2012 and 2013, it can be depicted as the limitation of the study. Perhaps, if the corpus had more data ranging from other periods, I could have found more frequency and more variations among the countries.