CHAPTER 3 METHODOLOGY
3.4 Data Analysis
To efficiently extract verb noun collocations, the corpus creating and Sketch Diff functions will be completely utilized. In order to understand the source of a concordance, 12,100 text files in TOEFL 11 will be firstly divided into 11 files based on the L1 languages and each file is named by the country codes according to the source of its country. The country codes are shown in Table 3.1. The 11 files will be zipped into one zip file and uploaded onto the Sketch Engine, named as TOEFL 11. On the other hand, the native speaker corpora, the BNC written corpus and academic texts of COCA, will be combined together as one large native speaker corpus and uploaded to the platform as well, named as BNCCOCA.
Table 3.2 The language codes of 11 languages
Language Arabic Chinese French German Hindi Italian
Language codes ARA ZHO FRA DEU HIN ITA
Language Japanese Korean Spanish Telugu Turkish
Language codes JPA KOR SPA TEL TUR
Then, as for the target items searched in the interface, the researcher will utilize the function of compiling wordlist in the Sketch Engine to create word lists of both nouns and verbs, which are ranked by the frequency and retrieved from the learner corpus. Both nouns and verbs in the wordlist will be typed in the interface to search for miscollocations. The production of word lists from learner corpora to retrieve collocations was also adopted in the
studies of Laufer (2011) and Liu (2013) although both studies only generated a word list based on frequent nouns since nouns express the semantic process of the predicates and serve as crucial indicators of verb noun collocations (C. Gledhill, 2007; Juknevičienė, 2008; Liu, 2002).
However, in order to ensure the full coverage of verb noun collocations, the present study will produce the wordlists of both verbs and nouns and consider frequent verbs and nouns as the target items that are tested with the Sketch Diff.
The threshold of frequency count is 200 in the present research. Both verbs and nouns that occur above 200 times in the learner corpus will be the keywords that are typed in the interface of the Sketch Diff. After the researcher types each noun and verb from the word lists in the interface of Sketch Diff, she will observe the column of object_of for nouns as the target items and object for verbs as the target items, meaning that the noun is the object of the verb or the verb’s object is the noun, to search for full coverage of verb noun collocations, as shown in Figure 3.9. Take time for example, time is the object of allow, comprising allow_ time as presented in the first line of the following figure. It should be noted that some miscollocations will be searched repeatedly if the researcher searches for both nouns and verbs. For instance, if the researcher searches for both do and mistake, the wrong collocate do_ mistake, which occur 26 times, will be searched twice. Thus, the researcher will check the miscollocations and delete the repeated ones.
Figure 3.9 The columns of object of and object from the noun “time” and the verb “do.”
As indicated above, the frequency of collocations in two corpora will be viewed to find out the potential miscollocations. Verb noun collocations, which occur at least 3 times in TOEFL 11 while do not occur in the native speaker corpus, will be considered as possible collocation errors. Since the corpus of TOEFL 11 is not large enough, the present study sets the low criterion of occurrence, which is three times. In addition, previous research, such as the study of Liu (2013), also set the standard of three-time occurrence in learner production and none in native corpora as the miscollocations analyzed in the research. The present study thus adopts the same standard of miscollocation definition.
The repeated searched miscollocations will be deleted as mentioned. The researcher will later clicks on those collocates to see the concordances, and clicks on the filename before the
concordance in order to check the original source of the concordance, which is the country code, as illustrated in Figure 3.10. Then, the potential collocation errors will be double-checked by a native speaker to delete the acceptable collocations. In the end, both verb noun miscollocations and their sources of L1s will be recorded. The whole data extraction process is summarized in Figure 3.11.
Figure 3.10 The source tracing function (After a user clicked on file2263628 before the concordance, the country code, ARA, in the yellow note will pop up).
Figure 3.11 The procedure of data extraction.
To sum up, in order to answer the three research questions, the researcher will analyze the data in the following steps. First, verb noun collocations will be considered as possible errors when they occur in the learner corpus instead of the native speaker corpus. Additionally, only the collocations that occur more than three times in the learner corpus will remain in the finding. The collocations that occur less than three times will be ignored in the result. The final list of miscollocations will be checked by a native speaker to delete the acceptable collocations.
Second, the collocation errors will be presented based on their frequency to observe the frequent collocation errors made by EFL learners of eleven L1s in response to the first research question. Third, in terms of the types of miscollocations, the present study will categorize verb noun miscollocations into different types, such as deviant verbs, deviant nouns, deviant preposition usages and others , based on the error types summarized by Nesselhauf (2003) in the section 2.2.2.
Fourth, in order to answer the third research question, which is the main purpose of the present study, the collocation errors, which occur in six out of eleven L1 groups, will be considered as universal errors that shared by most EFL learners. These errors are universal and intralingual errors that are difficult for most EFL learners, no matter what L1 language they use.
As for errors only made by learners of specific L1s, the researcher will attempt to categorize those collocation errors based on language families and draw a general conclusion. Since the possible causes of these errors are difficult to explain, the present study will try to offer some insight in the possible causes of these errors but will not explicitly conclude the causes of these errors.
Figure 3.12 The data analysis in the present study
CHAPTER 4 RESULTS AND DISCUSSION
4.1 Frequency
After the researcher deleted the collocation errors searched repeatedly and a native speaker deleted the acceptable collocations, 109 types of verb noun collocation errors, which occurred at least three times in TOEFL 11 and none in BNCCOCA, were found in TOEFL 11, with 807 occurrences in total. The types of collocation errors made by 11 groups of EFL learners were categorized into four main groups, including deviant prepositions of verbs, deviant verbs, deviant nouns and others by consulting the error type categorizations from Nesselhauf (2003), as shown in Table 4.1. The complete concordances of these miscollocations are presented in Appendix 1.
Table 4.1 The collocation errors in TOEFL 11 (ranked by frequency)
Types
*travel places (43)
*specialize themselves (3) specialize in 49
*listen music (20)
*listen a lecture (5)
*listen their story (3)
listen to 28
*participate activities (17)
*participate events (3) participate in 20
*go vacation (4)
*graduate university graduate from 9
*concentrate subjects concentrate on 8
*prepare the travel
*prepare an advertisement prepare for 8
*arrive the place arrive at 7
*care their community care about 7
*contribute their
communities contribute to 6
*adapt the society adapt to 4
*contract a guide contract with 4
*major subjects major in 4
*succeed the business succeed in 3
*learn hardware learn about 3
(2)Deviant
*experiment a culture (3) experience 28
delexical
simple verbs
*use celebrities hire celebrities 6
*face accidents have 5
*see hardships experience/ encounter 3
*sharp your skills sharpen 3
*watch advertising (4) advertisement 7
*create advertising (3)
*help neighborhood neighbors 5
*manipulate customers customers’ desire 5
*reach the destiny destination 5
*outnumber the decrease outnumber the
(number of cars…) 4
*learn the intuition fact 2
(4)Others
*have a travel take a trip 30
44
*conclude ideas make conclusions 7
*conclude the fact make conclusions 4
*ride a ski go skiing 3
Total frequency
807
Among the four types, the most and second frequent error types are the deviant preposition of verbs and deviant verbs. In terms of deviant prepositions of verbs, learners tend to neglect prepositions when they use certain verbs, such as agree, travel, specialize, etc. In this type, the most frequent pattern is the incorrect usage of “agree,” comprising 8 miscollocation types out of 109 types, such as *agree ideas, *agree the opinion and *disagree
the topic, with 120 occurrences in TOEFL 11. The suggested form should be agree with, like agree with ideas and agree with the opinion, which reveals that learners seemed to ignore the
preposition “with” when they used agree in their writing.The second and third frequent patterns in the deviant prepositions of verbs are travel and
specialize. Four examples of collocations containing the incorrect usage of travel were found,
with 53 occurrences in TOEFL 11. Learners seemed to connect the verb travel directly with any places, such as *travel places, *travel Japan and *travel a park. The preposition to was neglected by those learners, as in travel to places, Japan and a park. The similar prepositional errors were found in the usage of specialize, such as *specialize subjects and*specialize themselves, which occur 49 times in the learner corpus. The correct usage should
be specialize in subjects and specialize in respectively. The following are the examples of
travel and specialize:
[1] We have to travel more places to get more information (ARA).
[2] Last winter, I traveled Japan alone, and I studied the Japan to plan my trip (KOR).
[3] I strongly believe that specializing one subject is more appropriate than knowing
entire knowledge base (HIN).
[4] Other people prefer to specialize themselves in one specific subject (FRA).
The second frequent type of collocation errors is the deviant verbs, in which the most frequent pattern is the the incorrect usages of “loose1,” with 15 examples of collocation errors and 73 occurrences. Loose means “to untie someone or something” while lose refers to “to
stop having a particular attitude, quality, ability etc., or to gradually have less of it” based on
the definitions of Longman Dictionary of Contemporary English. Loose is to let go of something and lose is to stop having something, which indicates that the meaning of loose shares a certain degree of meaning with lose, giving rise to learners’ misuse and confusion of these two words. Additionally, the forms of these words are similar, contributing to learners’misuse of collocations. When the forms of the words are similar, learners may feel confused and make an error, leading to learners’ approximation of the two words’ meanings (Liu, 1999). The more detailed concordance examples are revealed as follows.
[5] First of all, all advertising companies create new product and see how we react to it
even though they might loose some money for producing it (JPN).
[6] I do not have to loose time to make decisions, and also I never missed famous places
to visit (JPN).
1 Some researchers may doubt that the use of loose might be due to wrong spelling instead of learners’
confusion of word meanings and forms. However, EFL learners’ intention of using these collocations remains unknown; thus, the present study will not discuss this debatable aspect.
The second frequent pattern in deviant verbs is related to the incorrect use of delexical verbs, including do, have, make and take, which comprise the errors, like *do mistakes, *take a
decision and *do enjoyment. This corresponds to previous studies that since delexical verbs
do not have specific meanings, learners tend to consider delexical verbs as unrestricted combinations with collocations (Boers et al., 2013; Liu, 1999; Zinkgräf, 2008), which would give rise to the collocation errors.The above result is discussed based on the frequency of miscollocations. After discussing the more frequent miscollocation categories, the researcher will further indicate the number of error types among collocation errors in TOEFL 11 based on Table 4.1 in the following section.
4.2 Error Types
Among 109 miscollocation types, both deviant verb usages and deviant prepositions of verbs contained 46 types of miscollocations. 13 miscollocation types belonged to the deviant noun usages and 4 types were others. Although the number of error types between deviant verb usages and deviant prepositions of verbs were the same, the frequency of deviant prepositions of verbs was more than that of verb usages, which indicates that learners in TOEFL 11 seemed to have more problems with the usage of prepositions in verb noun collocations. The error types and ratio of error types are illustrated in Table 4.2 and Figure 4.1.
Table 4.2 Error types of 109 collocation errors in TOEFL 11.
Error Types The numbers of error types Frequency
Deviant verb usages 46 295
Deviant prepositions of verbs 46 395
Deviant noun usages 13 73
Others 4 44
Total 109 807
Figure 4.1 The ratio of error types in TOEFL 11
The frequent miscollocations and the error types of all collocations were presented above. And it would be our interests to observe among these collocation errors, which collocations belonged to universal errors shared by learners of most L1s (at least 6 L1s) and which miscollocations only appeared in certain learner groups. The two dimensions will be revealed in the following part.
4.3 Universal Errors
Among 109 types of collocation errors, 61 types of miscollocations were universal errors, with 28 deviant verb usages, 27 deviant prepositions of verbs, 2 deviant noun usages and 3 other deviant usages, as shown in Table 4.3. The frequency of universal errors, the same as the frequent errors, comprised the most deviant prepositions of verbs, followed by deviant verbs and deviant nouns.
Table 4.3 The error types of universal errors.
Deviant usages V P N O Total
Error types 28 27 3 3 61
Frequency 216 298 25 41 580
*V= deviant verb usages, P= deviant prepositions of verbs, N= deviant noun usages, O=others.
4.3.1 Deviant verb usages
28 types of deviant verb usages were found among universal errors. The error types were categorized into five groups based on the same incorrect usage of verbs, which would be the similar types of mistakes that can be further elaborated. The five groups of universal errors are the incorrect usage of do, loose, experiment, try and reach, ranked by their total frequency in the learner corpus.
1. The deviant use of do in collocations
Group 1 contains the incorrect usage of a delexical verb “do”, such as *do mistakes/
specialization/ many adventures/graduation and such enjoyment, with the total frequency of
75 times in TOEFL 11. Learners of eight L1s made the errors regarding do, containing Arabic, German, French, Spanish, Hindi, Japanese, Korean, Turkish and Telugu EFL learners, as revealed in Table 4.4.Table 4.4 The deviant use of do
NO. Miscollocations
Suggested verbs
ARA DEU FRA ITA SPA HIN JPN KOR TUR TEL ZHO Frequency
Shared
In terms of the suggested verbs above, the suggested forms are the frequent and more appropriate verb collocates of the noun from consulting BNCCOCA, including the frequent usages of verb noun collocations in both British and American English. *Do mistakes, occurring 26 times in TOEFL 11, should be corrected as make mistakes. The correct verb for
*do specialization and *do graduation is have. And the more appropriate verb for *do many adventures and *do such enjoyment can be experience.
The possible explanation for learners to make these errors is because the verb do is a delexical verb, which does not have a specific meaning or special meaning compared to other lexical verbs. Thus, learners in TOEFL tended to overgeneralize the meaning of do and combine it with any nouns (Zinkgräf, 2008). In addition, do is the high-frequency vocabulary, which often includes a high degree of polysemy (Viberg, 1996), leading to learners’
erroneous assumption of a delexical verb’s meaning (Liu, 1999) and causing collocation errors.
2. The deviant use of loose in collocations
In Group 2, there are 15 types of the incorrect usage of loose, such as *loose time, *loose
some money, *loose everything, *loose something and *loose your interest, with 73
occurrences in total. Except Chinese learners, learners of the other L1s all made this kind of mistakes, as revealed in Table 4.5. The correct usages of collocations should be lose time,money, everything, something and interest. As explained in Section 4.1, learners confused
about these two words probably because these two words share a certain extent of meanings and their forms and pronunciations are similar.Table 4.5 The deviant use of loose
NO. Miscollocations
Suggested verbs
ARA DEU FRA ITA SPA HIN JPN KOR TUR TEL ZHO Frequency
Shared
reputation
*loose patience lose 0 0 0 0 0 0 0 0 0 3 0 3
Total Frequency 1 10 9 8 7 7 6 3 17 5 0 73
3. The deviant use of experiment in collocations
Group 3 contains the errors of experiment, such as *experiment new things (F=25) and
*experiment a different culture (F=3), which are shared by learners of seven L1s, including
French, Spanish, Hindi, Korean, Turkish, Telugu and Chinese EFL learners, as illustrated in Table 4.6. Observed from the concordance “Success depends on experimentingnew things even though they are risky (TEL),” the verb experiment should be corrected as experience, like experience new things and experience a different culture.
The causes of these errors could be discussed through their meanings and forms. According to Longman Dictionary of Contemporary English, experiment is an intransitive verb and its meaning is “to try using various ideas, methods etc. to find out how good or effective they
are” while experience is a transitive verb and “if you experience a problem, event, or situation, it happens to you or affects you.” Both words convey the meanings of trying
something while the former one is emphasizing whether those things are effective and the latter is focusing on how people are affected by the things they have tried. Therefore, the meanings might be confusing for learners. On the other hand, from the forms of these two words, it can be seen that the forms of these words are similar, which causes the confusion and learners’approximation for those two words (Liu, 1999), which is similar to the case of “loose and lose”
as discussed above.
Table 4.6 The deviant use of experiment
NO. Miscollocations
Suggested verbs
ARA DEU FRA ITA SPA HIN JPN KOR TUR TEL ZHO Frequency
Shared
4. The deviant use of try in collocations
In Group 4, the incorrect use of try in collocations, with 26 occurrences, were produced by learners of nine L1s, including German, French, Italian, Spanish, Hindi, Japanese, Korean, Turkish, Telugu and Chinese learners. Errors such as *try new risks, *try new challenges and
*try new ventures were made by these learners. The most frequent type of errors is *try new
risks, which occurs 17 times in TOEFL 11, as shown in Table 4.7.
Table 4.7 The deviant use of try
NO. Miscollocations
Suggested verbs
ARA DEU FRA ITA SPA HIN JPN KOR TUR TEL ZHO Frequency
Shared
The more correct verb for risk and challenge is take, as take risks and challenges. For
*try new ventures, the more appropriate verb would be undertake. Take risks occurs 1160
times, take challenges occurs 41 times and undertake ventures occurs 16 times in BNCCOCA while none of them combined with try occurs in BNCCOCA. The nouns here,risks, challenges and ventures, means something that is uncertain, dangerous or novel. EFL
learners tends to combine these uncertain and novel things with the verb, try, which might be influenced by the combination “try something new.” The phenomenon of these errors could be explained with the ideas of false analogy (Liu, 1999), that is, this kind of errors might be learners’ false analogy of “try something,” making learners feel that it is acceptable to add any nouns, especially something uncertain or new, after try.5. The deviant use of reach in collocations
The universal collocation errors of reach were found in Group 5, including *reach the
dream, *reach happiness and *reach fame, with 14 occurrences in total. Arabic, German,
French, Spanish, Hindi, Japanese and Korean learners made these errors in their TOEFL writing, as shown in Table 4.8. The more appropriate verb for them should be achieve, such as achieve the dream, happiness and fame according to BNCCOCA. Reach and achieve are synonyms, which makes learners tend to use these two words interchangeably with collocations since learners usually resort to Open Choice Principle (Sinclair, 1991) where learners will add up the meanings of each individual word. In this way, substituting one synonym with another in collocations will make no differences for learners. This phenomenon corresponds to one learner strategy, analogy, proposed by P. Howarth (1998), which is the process of substituting one lexical item for another known element in a collocation. This kind of L2 intralingual error is the erroneous use of synonyms in collocation errors pointed out by Liu (1999).Table 4.8 The deviant use of reach
NO. Miscollocations
Suggested verbs
ARA DEU FRA ITA SPA HIN JPN KOR TUR TEL ZHO Frequency
Shared
After the discussion of the universal errors concerning deviant verb usages, the following section is the deviant prepositions of verbs among universal errors.
4.3.2 Deviant prepositions of verbs
27 types of collocation errors out of 61 universal errors were deviant prepositions of verbs, categorized as seven groups sharing similar kinds of errors in each group and containing the incorrect usages of agree, specialize, travel, listen, search, go and participate. The causes of these errors are all due to the lack of prepositions after verbs.
1. The deviant use of agree in collocations
The first deviant prepositions of verbs are the use of agree and disagree (Frequency
=120). 9 groups of EFL learners made the errors, such as *agree this idea, *agree the opinion,
*disagree the opinion and *disagree that topic, as shown in Table 4.9. The preposition with
is neglected by learners. Theincorrect
usage of agree/disagree was also found in the studies of Lin (2010) and Liu (2013). Both studies analyzed Chinese and Taiwanese learners’collocation errors. After finding the misuse of agree/disagree, both Lin (2010) and Liu (2013) attributed these errors to L1 interferences, indicating that since in Chinese syntax, there are no structures for agree+prep, learners may have more troubles in using the verb agree or
disagree. However, the present study found that these errors are not only made by Chinese
learners. Learners, such as French, Italian, Spanish, Japanese and Korean, also made the errors in their written production, revealing that the misuse of agree/disagree may be also influenced by intralingual factors.Table 4.9 The deviant use of agree
2. The deviant use of travel in collocations
The second group of deviant prepositions of verb is the use of travel (Frequency=53),
The second group of deviant prepositions of verb is the use of travel (Frequency=53),