English for Specific Purposes 30 (2011) 101–112

Contents lists available at ScienceDirect

English for Specific Purposes

journal homepage: www.elsevier.com/locate/esp

How a corpus-based study of the factors which influence collocation

can help in the teaching of business English
Crayton Walker ⇑
Centre for English Language Studies, Westmere House, University of Birmingham, Birmingham B15 2TT, United Kingdom

a r t i c l e i n f o a b s t r a c t

Article history: In this paper I use two case studies to show how corpus linguistics can be used to help in
Available online 22 January 2011 the teaching of business English. Senior managers in global companies often find them-
selves having to do their job in a foreign language. Given that language is one of the key
Keywords: tools of management, the senior managers are normally very keen to develop a sophisti-
Collocation cated command of what they perceive to be key lexis. Consequently, the business English
Business teacher may be required to answer searching and demanding questions in a very precise
Corpus
and accurate way. In this paper I show how a corpus-based investigation of the colloca-
Semantic prosody
tional behaviour of key lexis can be used to answer many of these questions. For example,
a study of this kind can reveal the different senses of a word and show how it may be asso-
ciated with a particular semantic prosody. By studying the collocations associated with a
group of so-called synonyms it is often possible to identify slight but significant differences
in the meaning of the words in the group.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction

A review of some of the more recent research concerned with the use of corpora in ESP shows how there are basically
three main types of study. The first typically uses one or more corpora to examine aspects of the lexis associated with
the different domains and disciplines of ESP.1 Examples include Nelson (2006), who studied the semantic associations
(e.g., collocations and semantic prosodies) of words taken from the domain of business English, Gledhill (2000) who used
a corpus-based methodology to establish the type of collocations characteristic of scientific writing and Martinez, Beck,
and Panza (2009) who extracted the academic words from a specialised corpus of agricultural research articles. Durrant
(2009) examined the feasibility of compiling a list of frequent collocations which are common to a variety of different aca-
demic disciplines, while Hyland (2008) looked at the use of lexical bundles2 across a range of academic disciplines and found
that there were significant differences in their use.
The second type of study is one which compares academic writing produced by non-native speakers (NNS) of English with
that produced by native speaker students. Nesselhauf (2003, 2005), for instance, found significant differences in the
collocations used in English language essays written by German native speaker undergraduates compared with the colloca-
tions produced by native speakers. She concluded that the first language had some influence on the way in which the
German students collocated translation equivalents in English. Granger (1998) has also examined the use of collocations
and other prefabricated phrases in NNS writing and found that the NNS writers tended to ‘‘cling onto certain fixed phrases

⇑ Tel.: +44 121 414 3533; fax: +44 121 414 3298.
E-mail address: [email protected]
1
In this paper I include the teaching of business English under the heading of ESP.
2
Lexical bundles are groups of words (typically three, four or five words) which frequently occur together e.g., in order to, on the other hand, it should be noted
that. For more information see Biber, Conrad, and Cortes (2004).

0889-4906/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.esp.2010.12.003
102 C. Walker / English for Specific Purposes 30 (2011) 101–112

and expressions’’ (Granger, 1998, p. 156). See also Granger and Paquot (2010) and Schmitt (2004) for more examples of stud-
ies of this sort.
The third kind of study uses a corpus-based approach to assess the authenticity of the language used in ESP and EAP
coursebooks. Unfortunately for those involved in teaching these subjects, many of these studies (Candlin, Bhatia, & Jensen,
2002; Hyland, 1994; Paltridge, 2002; Skorczynska, 2010; Swales, 2002; Williams, 1988) seem to have reached the conclusion
that much of the published material fails to reflect the findings of corpus-based research into academic or business discourse.
Williams criticised textbooks for their choice of unnatural sounding linguistic exponents and found that real business meet-
ings show a high degree of linguistic complexity which is not reflected in the published texts. More recently, Skorczynska has
found that there were significant differences in both the type and use of metaphors which appeared in a published business
English text books (MacKenzie, 1997) compared with those which were found in a specialised corpus of written business
English.
The current study belongs to the first of the three types. It used two corpora to examine the collocational behaviour of
groups of semantically-related verbs (e.g., head, run, manage) and nouns (e.g., system, process, procedure) taken from the do-
main of business English. Results from the study show that much of the collocational behaviour exhibited by these items can
be explained by examining the linguistic features and processes which influence the way collocations are formed. The choice
of items was prompted by queries raised by advanced learners which teachers may have difficulty in answering, and the pur-
pose of this paper is to show how corpus evidence can be used to answer some of those more complex and searching ques-
tions and to confirm (or disprove) teachers’ intuitions about the collocational behaviour of these semantically-related verbs
and nouns.

2. Collocation and semantic prosody

It is clear from the literature that the term collocation is defined in a variety of ways, and that these different definitions
reflect differences in approach, the only common denominator being that the term is used to refer to some kind of syntag-
matic relationship between words. However, it is possible to group the different definitions into two broad categories, those
which use what I call a lexical approach to collocation (e.g., Carter, 1987; Cowie, 1998; Howarth, 1996, 1998) and those
which use a frequency or statistically based approach (e.g., Moon, 1998; Nesselhauf, 2003, 2005; Sinclair, 1991).
Those studies which follow a lexical approach use lexical criteria to decide whether a particular combination can be clas-
sified as a collocation or not. According to this approach, understanding a collocation will typically exhibit a degree of ‘fixed-
ness’ and/or ‘a lack of transparency’ in meaning. There is a tendency with this type of approach to create categories based on
the lexical characteristics exhibited by different combinations. Carter (1987), for example, proposed four categories of col-
locations (unrestricted, semi-restricted, familiar, and restricted collocations: Carter, 1987, p. 63) which seem to be largely
based on the criterion of commutability. Studies which use a frequency or statistical approach, on the other hand, typically
consider a collocation to be a co-occurrence of words within a certain distance of each other. Collocations are seen as being
co-occurrences that are ‘‘more frequent than could be expected if words combined randomly in a language’’ (Nesselhauf,
2005, pp. 11–12). Frequency based approaches are often associated with the work of Sinclair, whose own notion of colloca-
tion was, in turn, influenced by the work of Firth (Firth, 1957, 1968).
The studies which use a frequency or statistical approach do not regard collocations as belonging to a distinct linguistic
category. Collocations are viewed more in terms of probability, where the strength of a particular collocation is assessed on
the basis of how frequently it appears in a large representative sample of discourse. This means that there are virtually no
impossible collocations but that some collocations are much more likely to occur than others. However, as Halliday (1966)
has pointed out, there is a need for at least one cut-off point in order to eliminate combinations which are simply the result
of a random distribution of items within the discourse. Sinclair, writing in the OSTI3 report first circulated in 1970, used the
term ‘significant collocations’ to refer to combinations which co-occur more frequently than ‘‘their respective frequencies and
the length of the text in which they appear would predict’’ (Sinclair, Jones, & Daley, 2004, p. 10).
Sinclair also introduced three very useful terms for any discussion of collocation; span, node and collocate.
We may use the term node to refer to an item whose collocations we are studying, and we may define a span as the num-
ber of lexical items on each side of a node that we consider relevant to that node. Items in the environment set by the
span we will call collocates. (1966, p. 415)
In a paper published in 1974, Sinclair and Jones proposed a span of four words on either side of the node word. The fol-
lowing nomenclature is normally used to describe the positions in the span; node 4 to 1 to describe the four positions to
the left of the node and node +1 to +4 to describe the positions to the right. Although there is some statistical basis for using a
span of four words (Mason, 1997, 1999), the distance between a collocate and a node will normally depend on both lexical
and grammatical factors. For example, the distance between the node and the collocate(s) will normally be greater in the
case of verb/noun collocations compared with adjective/noun or noun/noun collocations and consequently it may be neces-
sary to use a slightly wider span when examining verb/noun collocations.

3
The original OSTI report was circulated among a small group of researchers in 1970. It has recently been re-published by Continuum in the Studies in
Corpus Linguistics series. This new edition, entitled English Collocation Studies, is edited by Krishnamurthy (2004).
 C. Walker / English for Specific Purposes 30 (2011) 101–112 103

Arguably, the frequency/statistical approach is more suited to a corpus-based methodology as it enables large quantities
of spoken or written discourse stored on a computer to be analysed by software programmes (concordancing packages)
which can extract the most frequent or the most statistically significant collocates for a particular node. These programmes
can be used to rank collocates according to their frequency or statistical significance for each of the different positions within
the span. It is also possible to specify a cut-off point in order to eliminate combinations which could have resulted from ran-
dom distribution.
Sinclair (1966, p. 418) limits the term collocate to a word which occurs above a purely random level of frequency. How-
ever, it is difficult to attach a precise degree of significance to a list of words ranked according to the number of times they
occur (raw frequency) together with the node. For this reason statistical measures such as t-score4 are frequently used in
corpus linguistics in order to assign a more precise level of significance to each co-occurrence. For example, any collocate
with a t-score of 2.00 or above can be regarded as significant; that is, the way that it combines with the node is not simply
the result of chance. It should be noted that all the collocates discussed in this paper have t-scores which are well above the
2.00 level. Given that there have been a number of reservations expressed about the use of statistical measures in corpus
research (Clear, 1993; Stubbs, 1995), both t-score and raw frequency data have been included in the tables shown below
with the collocates being ranked according to their t-score values.
I would like to introduce the concept of semantic prosody5 here in the introductory section as it will be discussed later in
this paper (Section 4.2.2). The term was used by Louw in a paper published in 1993 where he credits Sinclair with having pro-
vided him with both the idea and the term in ‘‘a personal communication’’ (Louw, 1993, p. 158). Sinclair examined the collo-
cational behaviour of the phrasal verb set in and found that most of the subjects associated with it referred to ‘‘unpleasant states
of affairs’’ (1991, p. 74). Louw claims that semantic prosody is the result of a diachronic process whereby meaning has been
transferred from one word or words to another and defines semantic prosody as a ‘‘consistent aura of meaning with which a
form is imbued by its collocates’’ (1993, p. 157).
In 1995, Stubbs reported the results of a corpus-based study of the collocational behaviour of CAUSE6 which showed that
although the majority of instances of CAUSE in the corpus exhibited a negative prosody, it would be inaccurate to claim that all
instances of CAUSE exhibit a negative prosody. Hunston (2007) also examined the collocational behaviour of CAUSE using a cor-
pus of articles from the journal The New Scientist. She found that not all uses of CAUSE are associated with unpleasant or negative
things and that it is only when CAUSE is associated with a human agent (or another animate entity) that it exhibits a degree of
negative prosody. Hunston concludes by warning against over-generalizing by making statements about a word (as opposed to
the different meanings of the word) having a semantic prosody. This conclusion is supported by results from the current study
(see Section 4.2.2) where it was found, for example, that not all senses of the word process exhibited a negative prosody. It may
be the case with a low frequency item such as set in that there is only one meaning and therefore, in this situation, it may be
possible to ascribe a particular prosody to the item as a whole. However, in the case of a word with a higher frequency such as
CAUSE which is associated with a number of different meanings, it may only be possible to ascribe a semantic prosody to one or
more of its meanings.
The term semantic prosody is also used by some writers (Nelson, 2006; Sinclair, 1996, 1998, 2004; Stubbs, 2001, 2009) in
a wider sense to describe the way in which a lexical item can develop one of a range of different prosodies such as ‘‘‘some-
thing nasty’ or ‘something worrying’ or ‘disturbing’ [. . .] ‘something magnificent’, ‘socially appropriate’ and ‘positively con-
structive’’’ (Sinclair, 2004, p. 173). Whitsitt (2005) has argued that when the term is used in this wider sense, it is simply a
reflection of the connotative meaning of the word, and for this reason I have chosen to limit the use of the term in this paper
to Louw’s original notion of a lexical item having either a positive or negative prosody.

3. Details of the current study

The main corpus used in the current study was the Bank of English (BoE)7 corpus which is a large corpus of general English
consisting of 450 million words. A second more specialised corpus of business English was also used in order to check that the
results obtained from the corpus of general English are also valid in the domain of business English. The second corpus, which
was made up of commercial and financial data files from the British National Corpus,8 contains 6.3 million words. In the current
study this second, more specialised, corpus is referred to as the British National Commercial Corpus (BNCc).
The first stage of the research consisted of establishing a collocational profile for each of the selected items using a corpus
of general English, in this case the BoE. This involved identifying the most frequent collocates for each of the positions within
a span of four words on either side of the node. The easiest way to do this with the BoE was to use the picture function which

4
T-score is a statistical instrument which is used to measure distribution, or more specifically how the distribution of something deviates from what is
standard. For more information regarding t-score please refer to Hunston (2002) and Barnbrook (1996).
5
Refer to Stewart (2010) for a very comprehensive and critical account of semantic prosody.
6
Capitals are used in this study to indicate that reference is being made to all members of the lemma. For example, the verb CAUSE has been written in
capitals in order to show that all forms of the lemma were studied (i.e., cause, causes, causing, caused).
7
The Bank of English (BoE) corpus is jointly owned by HarperCollins Publishers and the University of Birmingham. During 2003–2006, when most of the
research was carried out, the corpus contained 450 million words. http://www.titania.bham.ac.uk
8
The British National Corpus (BNC) is a 100 million word corpus developed in the 1980s. It is maintained and distributed by the Oxford University Computer
Service (OUCS). http://www.natcorp.ox.ac.uk.
104 C. Walker / English for Specific Purposes 30 (2011) 101–112

identifies and ranks the most frequent collocates for each of the positions within a specified span. Once all the data had been
recorded they could then be examined using more quantitative techniques in order to identify features such as groups of
semantically-related collocates, frequently occurring fixed or semi-fixed phrases or any signs of a particular semantic
prosody.
The aim of the second stage of the research was to establish a collocational profile for each of the selected items using a
corpus of business English. By comparing the two profiles (i.e., the profile obtained from the BNCc and the profile obtained
using the BoE) it was possible to establish whether there are any significant differences in the way the selected items are
used in a business domain compared with a more general one.9

4. Introduction to the case studies

A manager working in a foreign language will usually need to develop a very sophisticated knowledge of the target lan-
guage in order to be able to perform successfully. The manager may already be very competent in his or her own first lan-
guage and will often be striving for the same level of competency in the foreign language. Because this type of learner is
already aware of many of the nuances and complexities of their own first language, they will often ask the teacher very
searching questions in order to discover the same degree of complexity in the target language.
So-called synonyms can have slight but significant differences in meaning and these differences can often be important
for communication in situations where precision is required; in business for example. In this paper I use two case studies to
illustrate the way in which corpus linguistics can inform the teaching of business English. The detail included in each case
study is based on real students whom I taught during my 20 years as a business English teacher in Germany. Each study con-
tains one or more questions which were asked by the students at the time and these rather searching questions will be an-
swered by examining data drawn from the BoE and the BNCc.

4.1. Case study one

4.1.1. The learner and the teaching situation

Doctor T is a senior manager working in a large manufacturing company with its head office in Germany. He is 47,
German and responsible for Human Resources and Training throughout the company. The German company has recently
acquired a British auto-parts manufacturer and Doctor T has been asked to visit this newly acquired subsidiary to meet
the management and address the workforce. In order to prepare for the visit Doctor T has booked 40 hours of one-to-one
training with an experienced English teacher. He wants the teacher to help him prepare the presentation he will give to
the native speaker workforce. The presentation is about the structure of the German parent company, its Human Resources
policies and ethos, and its management style. The German company prides itself on having a very ‘flat’ structure and an
inclusive decision-making style.
Doctor T already has a high standard of English but is very concerned about getting the right tone in his presentation.
While working on the initial part of the presentation, Doctor T turns to his teacher and asks ‘‘should I say I run, manage
or head, the Human Resources Division, or should I use a phrase like I am responsible for, or in charge of instead?
How should the teacher answer Doctor T’s question bearing in mind that he is very concerned about creating the right
impression amongst the native-speaker audience? Would the teacher be correct in saying, for example, that these are simply
different ways of introducing one’s position in the company, or are there slight differences in meaning which reflect different
styles of management? Is it the case that items like run or in charge of are more frequently associated with a top-down
management style, while responsible for or manage are associated with a more collaborative style?

4.1.2. Evidence from the Corpora

Table 1 shows some of the most frequent collocates of RUN, MANAGE and HEAD. The word business, for example, was
found to occur 946 times in the node +2 position10 with the verb RUN (e.g., run a business, ran the business, running a business)
in the Bank of English; whereas it only occurred 164 times with MANAGE and 14 times with the verb HEAD (e.g., head the busi-
ness). It can be seen that RUN is more frequently associated with nouns which refer to non-human entities such as businesses,
companies or firms. MANAGE is associated both with nouns which describe such entities, and with nouns which refer to people.
HEAD, on the other hand, occurs more frequently with nouns which denote structures consisting of people, such as committees
or panels.
A more detailed examination of the most frequent node +2 nominal collocates of RUN (see Table 2) shows how it is reg-
ularly associated with the notion of power. RUN, for example, frequently occurs with country and government.

9
The results showed that there were very few differences in the way the items discussed in this paper (i.e., RUN, HEAD, MANAGE, system. process and
procedure) are used in the two domains and, as a result, much of the data used in this paper have been taken from the BoE because, as the larger of the two
corpora, it is liable to yield more reliable results.
10
As a general rule, the node +2 and node +3 positions were found to yield the most nominal collocates when dealing with verb/noun collocations (e.g., RUN a
company, MANAGE the finance department) whereas the node 1 position is more fruitful when examining adjective/noun or noun/noun collocations (e.g., legal
system, learning process).
 C. Walker / English for Specific Purposes 30 (2011) 101–112 105

Table 1
A comparison of the most frequent node +2 collocates of RUN, MANAGE and HEAD. The data are taken from the BoE.

RUN MANAGE HEAD

t-score Freq. t-score Freq. t-score Freq. Node +2
30.73 946 12.80 164 3.73 14 BUSINESS
24.48 601 9.73 95 7.00 49 COMPANY
11.12 124 5.74 33 3.87 15 FIRM
0.00 0 5.99 36 0.00 0 PEOPLE
0.00 0 5.10 26 0.00 0 STAFF
0.00 0 3.16 10 0.00 0 EMPLOYEES
0.00 0 0.00 0 7.40 55 COMMITTEE
2.45 6 0.00 0 3.71 14 BOARD
0.00 0 0.00 0 3.66 13 PANEL

Table 2
The 10 most frequent node +2 nominal collocates associated with RUN. The data
are taken from the BoE.

Node +2 t-score Freq.

1. Risk 34.14 1187
2. Business 29.62 946
3. Country 28.40 874
4. Company 22.93 601
5. Show 22.09 539
6. Campaign 13.83 215
7. Club 13.50 226
8. Government 13.20 276
9. Place 13.18 239
10. School 12.21 202

The following concordance lines taken from the BNCc data illustrate the way in which RUN is frequently used with the
collocates country and government.

1 there are more efficient ways to run a country. No European nation can
2 land. But I don’t know how you can run a country of frothy money, and I
3 their carefree attitude to running the country is beginning to have an
4 ontinue as long as big borrowers run the government. The Bank has to ret
5 little prospect of being able to run a government, Mr. Shariff set about
6 group. Civic Forum, which is running the government in the meantime is abo

The data in Table 2 also show how RUN is frequently associated with phrases such as run the show and run the place. The first
phrase run the show does not refer to an actual show but rather to those in power, whereas the second run the place does refer
to those who are in charge of a particular locality, organisation or business. These two examples taken from the BNCc illus-
trate this difference.

7 They abandoned long ago an illusion to which most executives cling – that only the people at the
top have the wherewithal to run the show. The best way to exercise power, the AES founders argue, is
to give it up.
8 He refused to defer to the board of directors, most of whom he saw as country-clubbers who golfed
while he ran the place.

The fact that RUN is frequently associated with power, and that overt expressions of power are often avoided when they
relate to having power over people, may explain why RUN is more frequently used with nouns (in the object position) which
refer to non-human entities rather than to people.
An analysis of the data shows that MANAGE is associated with nouns in the object position which refer to both a team of
people and also to non-human entities such as a club or a business. The figures in Tables 1 and 3 show that, although
MANAGE is frequently used with both people and non-human entities, the non-human considerably outnumber the people.
The list of the 10 most frequent node +2 nominal collocates (see Table 3) also shows how MANAGE is frequently associated
with financial items such as the economy, money, investment, finances and so forth.
106 C. Walker / English for Specific Purposes 30 (2011) 101–112

Table 3
The 10 most frequent node +2 nominal collocates associated with MANAGE. The
data are taken from the BoE.

Node +2 t-score Freq.

1. Business 12.24 163
2. Economy 11.13 128
3. Money 10.75 128
4. Affairs 9.79 98
5. Club 9.45 98
6. Investment 9.28 92
7. Portfolio 9.25 87
8. Company 8.98 95
9. Team 8.72 87
10. Finances 8.52 73

Table 4
The 10 most frequent node +2 nominal collocates associated with HEAD. The data
are taken from the BoE.

Node +2 t-score Freq.

1. Team 11.43 144
2. Government 9.34 112
3. Investigation 9.27 88
4. Inquiry 9.06 84
5. Department 8.87 84
6. Group 8.54 88
7. Committee 7.06 55
8. CIA 6.60 44
9. Commission 6.48 47
10. Company 5.72 49

Table 5
The 10 most frequent node +2 nominal collocates associated with responsible for.
The data are taken from the BoE.

Node +2 t-score Freq.

1. Deaths 15.07 228
2. Actions 13.08 172
3. Death 12.16 153
4. Shareholders 10.02 101
5. Loss 9.70 97
6. Attack 9.31 90
7. Policy 9.26 91
8. Development 9.15 88
9. Murder 9.12 85
10. Safety 8.61 76

The evidence from both corpora seems to show that HEAD (see Table 4) is frequently associated with nouns which belong
to one of two categories. The first category refer to structures made up of, for example, panel, team, committee, commission
and the government, whereas the second refers to activities which involve people, for example, investigation, inquiry.
The corpus data for responsible for (see Table 5) seem to show that the phrase has two distinct but related meanings. The
first describes the responsibility that someone (e.g., a manager) or something (e.g., a department) has for a particular part or
function of a company, business or similar type of organisation, while the second meaning is more concerned with the
answerability that someone or something has for a particular action or event. In the case of the second meaning, the event
or action is normally negative in nature, as can be seen from the collocates listed in Table 5. Death, deaths, actions, loss, attack,
murder all relate to the second meaning, while shareholders, policy, development and safety are associated with the first.
A comparison of the data for in charge of (see Table 6) and responsible for (see Table 5) shows how the two phrases have
one meaning in common, that is, the responsibility for managing an element of a company or organisation (i.e., the first
meaning of responsible for). In charge of is not associated with any sense of answerability for a negative action or event
and all 10 of the most frequent node +2 nominal collocates associated with the phrase refer to non-human entities, with
the possible exception of team. However, in only three out of the 68 occurrences is team followed by the preposition of,
which indicates that, like RUN, in charge of is associated more with the non-human attributes of the word team. The fact that
country is also one of the most frequent collocates of in charge of supports the argument that, like RUN, the phrase is
frequently associated with the notion of power and therefore tends to be used with nouns which refer to non-human
 C. Walker / English for Specific Purposes 30 (2011) 101–112 107

Table 6
The 10 most frequent node +2 nominal collocates associated with the phrase in charge
of. The data are taken from the BoE.

Node +2 t-score Freq.

1. Case 8.77 80
2. Policy 8.59 76
3. Affairs 8.44 72
4. Investigation 8.33 70
5. Project 7.99 65
6. Team 7.76 64
7. Country 6.83 51
8. Operations 6.65 45
9. Operation 6.40 42
10. Development 6.26 41

entities. In other words, the evidence seems to show that we prefer to speak about being in charge of something rather than
someone.

4.2. Case study two

4.2.1. The learner and the teaching situation

Jürgen is 34, also from Germany but is currently working in the UK. He is a purchasing manager working for a large
German car manufacturer and has recently been asked to head a team of 12 administrators/purchasers working in the com-
pany’s UK offices near Coventry. He has a very good level of English but still visits a local language school twice a week for
some individual tuition to help with his business English.
Jürgen wants to introduce a new way of handling customer invoices at the UK office and would like to know from his
English teacher whether he should refer to this as a new system, process or procedure. He wants to know whether the three
terms mean the same or whether they have different connotations. Jürgen feels that he has a very inclusive style of manage-
ment and wants the members of his team to discuss the new plan for handling invoices and to actively participate in its
implementation. He feels that he should avoid the term procedure as this might imply something which is prescriptive
and ‘set in stone’. It might be better to use the term process, or so he thinks, as he is referring to a series of actions rather
than to something that has a physical form. What advice should his teacher give, remembering that Jürgen wants to give
the impression that he has a very inclusive management style and does not want to be seen as the German manager impos-
ing a German way of doing things on the UK team?

4.2.2. Evidence from the corpora

An examination of the most frequent node 1 collocates shows that system (see Table 7) is used to refer to a range of
different types of systems and it is therefore possible to group these collocates according to the type of system they describe.
The corpus data contained a number of verbal collocates which were found to occur with specific types of system. Verbs
such as INSTALL and ASSEMBLE, for example, were found to be mainly associated with technical systems, whereas REFORM
and RESTRUCTURE were largely used with social systems. The following concordance lines taken from the BoE contain exam-
ples of verbal collocates found to occur with specific types of system.

9 key factor, because that can depress the human immune system.’’ Analyses
10 a chemical transmitter that stimulates the heart, digestive system and
11 lpha-2 blocking agents that act on the central nervous system. With the
12 ld. Designing, building and maintaining the computer system that keeps
13 one of the first brewers to install a cellar cooling system free from
14 his fever, Professor Saito assembled a temporary distillation system
15 has unveiled its plan for reform of the banking system. Treasury offi
16 of worker participation, to restructure the social security system, and
17 this year alone on a major overhaul of the prison system. He said the

Examination of a random sample of 500 concordance lines taken from the BoE showed that system was most frequently used
to refer to social systems (36% of the sample) followed by technical (17%) and political systems (14%). A comparison of the
data for node 1 collocates in the two corpora showed that, based on the average number of occurrences per million
words,11 collocates which refer to business or technical systems (e.g., management, computer) occur more frequently in the

11
The content of the BNCc was examined using a concordancing package called Wordsmith Tools Version 3 (Scott, 1998) which does not offer the option of
calculating the strength of a collocational pairing using t-score. It was therefore necessary to use the average number of occurrences per million words (per
millionages) to compare results from the two corpora.
108 C. Walker / English for Specific Purposes 30 (2011) 101–112

Table 7
The most frequent node 1 collocates associated with the singular form system. The data
are taken from the BoE.

Node 1 t-score freq.

Social systems
Legal 39.31 1554
Education 36.90 1380
Business systems
Management 18.78 461
Accounting 10.56 112
Transport systems
Transport 23.62 565
Rail 14.99 225
Geographical systems
Solar 40.17 1615
Rail 11.04 122
Political systems
Capitalist 18.78 354
Democratic 18.31 341
Technical systems
Computer 40.11 1610
Telephone 17.16 305
Biological systems
Immune 55.43 3074
Nervous 46.67 2200

Table 8
The most frequent node 1 collocates associated with the noun process. The data
are taken from the BoE.

Node 1 t-score freq.

Legal processes
Legal 18.56 348
Judicial 13.76 190
Political processes
Peace 79.38 6303
Political 30.70 952
Production processes
Production 21.02 447
Manufacturing 15.86 252
Cognitive processes
Learning 21.26 453
Creative 19.05 364
Business processes
Decision-making 25.37 644
Negotiating 12.53 157
Natural processes
Healing 17.28 299
Ageing 15.55 242

BNCc, whereas collocates which refer to biological or geographical systems (e.g., immune, solar) occur more frequently in the
BoE. Differences in the frequencies in the two corpora of collocates which refer to social, political or transport systems were
found to be less significant. These findings reflect the difference in the content of the two corpora and it is to be expected that
the BNCc will include more occurrences of collocates which refer to business and technical systems.
An examination of the most frequent node-1 collocates associated with the noun process shows that it is used to refer to a
variety of different types of processes. Table 8 shows the most frequent node-1 collocates grouped according to the type of
process they refer to.
A comparison of the data for node-1 collocates in the BoE and the BNCc, based on the average number of occurrences per
million words, showed that the collocates which refer to production or manufacturing processes occur more frequently in
the BNCc, whereas those which refer to political or natural processes occur more frequently in the BoE. These differences,
like those found with system, reflect differences in the content of the two corpora.
There is evidence from both corpora to show that both the singular and plural forms of process exhibit a negative semantic
prosody in that they are more frequently associated with adjectives which refer to negative attributes rather than to positive
ones. However, this negative semantic prosody is only associated with the single nouns process and processes, and not with
noun phrases containing process or processes (e.g., learning process, manufacturing process, biological processes, etc.). Table 9
 C. Walker / English for Specific Purposes 30 (2011) 101–112 109

Table 9
The left-hand column shows the most frequent node 1 attributive adjectives
associated with process while the right-hand column shows how frequently
their antonyms occur in the node 1 position. The data are taken from the
BoE.

Node 1 t-score Freq.

Long 17.42 304
Lengthy 11.74 138
Slow 16.90 286
Gradual 12.20 149
Complex 13.37 179
Difficult 11.31 128
Painful 12.44 155
Short 1.73 3
Quick 3.60 13
Fast 1.73 3
Simple 9.16 84
Easy 6.56 43
Painless 3.74 14

Table 10
The most frequent node 1 collocates associated with procedure. The data are
taken from the BoE.

Node 1 t-score Freq.

Legal procedures
Criminal 8.70 76
Parliamentary 8.05 65
Recruitment procedures
Selection 10.94 120
Vetting 5.29 28
Complaints procedures
Complaints 11.30 128
Grievance 8.24 68
Medical procedures
Surgical 11.31 128
Medical 9.13 84
Scientific procedures
Testing 6.46 42
Experimental 5.18 27
Occupational procedures
Operating 6.99 49
Administrative 5.18 27

shows the most frequent attributive adjectives which occur with process. The figures show that this negative prosody is not
totally consistent and that, for example, process is also frequently associated with the more ‘positive’ collocates simple and
easy.
Further evidence for this negative prosody can be seen in the pattern adjective and adjective + noun (e.g., long and difficult
process). The adjectives which most frequently appear within this pattern are all ‘negative’, as can be seen in these concor-
dance lines taken from the BNCc.

18 re-learning the painful and difficult process of change itself. Again,

19 however, can be a long and painful process, and often it can be futile
20 this would create a long and wasteful process before Romania recognise
21 embark upon the difficult and risky process of integration, they say
22 expenses is an uncertain and complex process. This estimation process

Like system and process, an examination of the corpus data for the most frequent node 1 collocates of procedure reveals how
they refer to a range of different types of procedure. Table 10 shows the most frequent node-1 collocates grouped according
to the type of procedure they refer to.
Unlike system (see Table 8), the data for procedure contained very few verbal collocates associated with specific types of
procedure. PERFORM and UNDERGO were found to occur with medical procedures, as can be seen in the following concor-
dance lines taken from the BoE data.
110 C. Walker / English for Specific Purposes 30 (2011) 101–112

23 watched one surgeon perform the procedure on a 21-year-old girl with

24 surgeon who performs a surgical procedure incorrectly can be held legally
25 numbers of men undergoing this procedure are small but plastic surgeons
26 to choose to undergo a medical procedure so they can give informed cons
27 ce about undergoing a surgical procedure. Sometimes these professionals

However, as a general rule, procedure occurs more often with higher frequency verbs such as use, change, follow, and
explain which are associated with a range of different types of procedure.
There is evidence from both corpora showing that procedure is frequently used to refer to something which is seen as
being the correct or normal way of doing things. For example, it was found that procedure is frequently associated with a
characteristic set of node 1 adjectival collocates which include items such as normal, standard and correct. The fact that
FOLLOW was found to be the most frequent verbal collocate to occur in the node 2 and node 3 positions also supports
the argument that procedure is used in a more prescriptive way than either system or process. The following concordance
lines taken from the BoE show how the verbal collocate FOLLOW is used together with the adjectival collocates mentioned
above.

28 fear a tone. Following the usual procedure, you repeatedly sound the
29 instead of following the normal procedure of completing a circuit of
30 emains, why follow this standard procedure? Why don’t we simply value
31 rators not following the correct procedure’’, adding that it would take
32 believer in following the proper procedure, came in, I explained the

There are also instances in the data where procedure is used in a more general way to refer to the normal or prescribed way of
behaving in, for example, a law court or in parliament. When procedure is used in this way it is uncountable and therefore
characterised by the lack of a determiner. The following concordance lines taken from the BoE contain instances of this
uncountable form of procedure.

33 rules for particular situations. Procedure is also a means towards the

34 grumble and do as they are told. Procedure can be too rigid. The standa
35 process focuses on questions of procedure rather than guilt, often mak
36 You know, Speaker, the rules of procedure of this assembly and this was

The uncountable form of procedure appears rather infrequently in the data. Out of a random sample of 500 concordance lines
taken from the BoE, only 38 contained the uncountable form (7.6% of the total sample). This uncountable form was also
found to be frequently associated with the pattern noun + of + procedure (e.g., a question of procedure, rules of procedure).

5. Discussion and conclusion

Based on the evidence from both corpora, Doctor T would be well advised to stay clear of the word RUN and the phrase in
charge of, both of which would seem to be associated with power (e.g., run the show, in charge of the country) and therefore a
top-down management style. In addition, the data show that RUN frequently occurs with nouns which describe non-human
entities and may give the feeling to the native-speaker audience that their new masters regard them as automatons who
simply have to be told what to do. In fact, this is exactly the impression which Doctor T does not want to give, and so his
teacher would be wise to point him in the direction of the verb MANAGE or a phrase such as responsible for which do not
seem to carry the same connotation of power and are more frequently associated with people.
Given the corpus evidence, and the fact that Doctor T wants to create the impression of a very flat and inclusive manage-
ment style, his teacher should advise him to use utterances such as I am responsible for human resources or I manage the
Human Resources Division and to avoid phrases like I am in charge of human resources or I run the Human Resources Division.
It would seem, based on the data from both corpora, that Jürgen should avoid using the word procedure. Given that he is
trying to demonstrate a very inclusive style of management and encourage his team to discuss the details of the new way of
handling invoices, it seems that procedure is far too prescriptive for Jürgen’s purposes. Equally, if he were to use the word
process, it may create the impression of something which is long, complex and slow in the minds of the native speakers. Sys-
tem, on the other hand, is neither as prescriptive as procedure nor as negative as process and therefore, according to the data
from the corpora, would appear to be the best candidate of the three.
In conclusion, Jürgen’s teacher should tell him that he is right to avoid the term procedure as the data confirms his impres-
sion that it is too prescriptive for the current situation, but equally he should also avoid using the word process as it is
 C. Walker / English for Specific Purposes 30 (2011) 101–112 111

obviously associated with a certain level of negativity. In this case, his teacher should recommend the use of the word system,
this being the most neutral of the three.
The results from the current study show that many collocations are not simply arbitrary or idiomatic combinations of
words. It is the frequent collocates that help to disambiguate the different uses of a word and identify slight but significant
differences in meaning between what might appear to be groups of synonyms such as RUN, HEAD MANAGE or system, process
and procedure. A corpus-based study of the collocational behaviour of a word can identify the different meanings or uses of
that word and whether it is associated with any negative prosody or similar connotational associations; information that is
invaluable to the lexicographer, materials writer, classroom teacher and language learner.
There are many accounts in the literature of how one or more corpora, or large quantities of corpus data, have been used
in the classroom to heighten the learners’ awareness of various features of the language. Gavioli (2005), for example, has
shown how the analysis of smaller specialised corpora, restricted in the range of text-types and topics they contain, can
be used to heighten awareness of key lexical, grammatical or textual issues amongst learners of ESP. Lee and Swales
(2006) describe how NNS PhD students following a course in EAP were given access to specialized corpora of academic dis-
course. The students were encouraged to investigate one or more aspects of the language and to make comparisons between
their own writing and that of more established writers in their fields. It is clear from these studies that a corpus or corpus
data can be used quite successfully in the classroom to heighten awareness but that it may need some form of mediation in
order to help the learners interpret the data. It is clear from the work of Vannestál and Lindquist (2007), for example, that
students often have difficulties interpreting corpus data, particularly when looking at the more complex phraseological as-
pects of English. Flowerdew (2009) advocates a form of mediation which involves the students being prompted, with a work-
sheet for example, to notice particular features of the language, formulate rules and to discuss their observations.
In this paper I have used my experience as a teacher of business English to show how learners can ask very searching and
demanding questions about the lexis they are learning. The senior manager is often very aware of how to use particular
words and phrases in his or her own first language and is trying to develop the same level of expertise concerning what
he or she regards as being the key lexis in the foreign language. Results from the current study show how a corpus can
be used to answer many of these questions and how it is a valuable resource for teachers and learners alike. The corpus
is a very useful reference tool for teachers enabling them to check the many intuitions they may have about the language
before passing them on to their learners and it can also often be used in the classroom to heighten learners’ awareness of
some of the more important lexical features of the language.

References

Barnbrook, G. (1996). Language and computers. Edinburgh: Edinburgh University Press.

Biber, D., Conrad, S., & Cortes, V. (2004). If you look at. . .: Lexical bundles in university teaching and textbooks. Applied Linguistics, 25, 371–405.
Candlin, C., Bhatia, V., & Jensen, C. (2002). Developing legal writing materials for English second language learners: Problems and perspectives. English for
Specific Purposes, 21, 299–320.
Carter, R. (1987). Vocabulary: Applied linguistic perspectives. London: Allen and Unwin.
Clear, J. (1993). From Firth Principles: Computational tools for the study of collocation. In M. Baker, G. Francis, & E. Tognini-Bonelli (Eds.), Texts and
technology: In honor of John Sinclair (pp. 271–292). Amsterdam: John Benjamins.
Cowie, A. P. (Ed.). (1998). Phraseology: Theory, analysis and application. Oxford: Clarendon Press.
Durrant, P. (2009). Investigating the viability of a collocation list for students of English for academic purposes. English for Specific Purposes, 28, 157–169.
Firth, J. R. (1957). Papers in linguistics 1934–1951. London: Oxford University Press.
Firth, J. R. (1968). Descriptive linguistics and the study of English. In F. R. Palmer (Ed.), Selected papers of J.R. Firth 1952–1959 (pp. 96–113). London: Longman.
Flowerdew, L. (2009). Applying corpus linguistics to pedagogy. International Journal of Corpus Linguistics, 14, 393–417.
Gavioli, L. (2005). Exploring corpora for ESP learning. Amsterdam: John Benjamins.
Gledhill, C. (2000). The discourse function of collocation in research article introductions. English for Specific Purposes, 19, 115–135.
Granger, S. (1998). The computer learner corpus: A versatile new source of data for SLA research. In S. Granger (Ed.), Learner English on computer (pp. 3–17).
London: Longman.
Granger, S., & Paquot, M. (2010). Lexical verbs in academic discourse: A corpus-driven study of learner use. In M. Charles, S. Hunston, & D. Pecorari (Eds.), At
the interface of corpus and discourse: Analysing academic discourses (pp. 193–214). London: Continuum.
Halliday, M. A. K. (1966). Lexis as a linguistic level. In C. Brazell, J. Catford, M. A. K. Halliday, & R. Robins (Eds.), In memory of J.R. Firth (pp. 148–162). London:
Longman.
Howarth, P. A. (1996). Phraseology in English academic writing. Tübingen, Germany: Max Niemeyer Verlag.
Howarth, P. A. (1998). Phraseology and second language proficiency. Applied Linguistics, 19, 24–44.
Hunston, S. (2002). Corpora in applied linguistics. Cambridge: Cambridge University Press.
Hunston, S. (2007). Semantic prosody revisited. In R. Moon (Ed.), Words, grammar, text: Revisiting the work of John Sinclair. International Journal of Corpus
Linguistics (Vol. 12, pp. 249–268).
Hyland, K. (1994). Hedging in academic writing and EAP coursebooks. English for Specific Purposes, 13, 239–256.
Krishnamurthy, R. (Ed.). (2004). English collocation studies: The OSTI report. London: Continuum.
Hyland, K. (2008). As can be seen: Lexical bundles and disciplinary variation. English for Specific Purposes, 27, 4–21.
Lee, D., & Swales, J. M. (2006). A corpus-based EAP course for NNS doctoral students: Moving from available specialized corpora to self-compiled corpora.
English for Specific Purposes, 25, 56–75.
Louw, W. (1993). Irony in the text or insincerity of the writer: The diagnostic potential of semantic prosodies. In M. Baker, G. Francis, & E. Tognini-Bonelli
(Eds.), Text and technology: In honour of John Sinclair (pp. 157–176). Amsterdam: John Benjamins.
MacKenzie, I. (1997). Management and marketing. Hove, UK: LTP Language Teaching Publications.
Martinez, I., Beck, S., & Panza, C. (2009). Academic vocabulary in agriculture research articles: A corpus-based study. English for Specific Purposes, 28,
183–198.
Mason, O. (1997). The weight of words: An investigation of lexical gravity. Proceedings of PALC 97 (pp. 361–375). Lodz, Poland: University of Lodz.
Mason, O. (1999). Parameters of collocation: The word in the centre of gravity. In J. Kirk (Ed.), Corpora galore (pp. 267–280). Amsterdam: Rodopi.
Moon, R. E. (1998). Fixed expressions and idioms in English: A corpus-based approach. Oxford: Clarendon Press.
Nelson, M. (2006). Semantic associations in business English: A corpus-based analysis. English for Specific Purposes, 25, 217–234.
112 C. Walker / English for Specific Purposes 30 (2011) 101–112

Nesselhauf, N. (2003). The use of collocations by advanced learners of English and some implications for teaching. Applied Linguistics, 24, 223–242.
Nesselhauf, N. (2005). Collocations in a learner corpus. Amsterdam: John Benjamins.
Paltridge, B. (2002). Thesis and dissertation writing: An examination of published advice and actual practice. English for Specific Purposes, 21, 125–143.
Schmitt, N. (Ed.). (2004). Formulaic sequences: Acquisition, processing and use. Amsterdam: John Benjamins.
Scott, M. (1998). Wordsmith tools: Version 3. Oxford: Oxford University Press.
Sinclair, J. (1966). Beginning the study of lexis. In C. E. Brazil, J. C. Catford, H. A. K. Halliday, & R. H. Robin (Eds.), In memory of J.R. Firth (pp. 410–430). Harlow:
Longman.
Sinclair, J. (1991). Corpus. concordance and collocation. Oxford: Oxford University Press.
Sinclair, J. (1996). The search for units of meaning. Textus 9/1. In R. Carter (Ed.), Trust the Text: Language, corpus and discourse (pp. 24–48). London: Routledge
(Reprinted).
Sinclair, J. (1998). The lexical item. In E. Weigand (Ed.), Contrastive lexical semantics. Amsterdam: John Benjamins. In R. Carter (Ed.), Trust the text: Language,
corpus and discourse (pp. 131–148). London: Routledge (Reprinted).
Sinclair, J. (2004). Lexical grammar. In R. Carter (Ed.), Trust the text: Language, corpus and discourse (pp. 164–176). London: Routledge.
Sinclair, J., & Jones, S. (1974). English lexical collocations: A study in computational linguistics. Cashiers de Lexicologie, 24(2), 15–61.
Sinclair, J., Jones, S., & Daley, R. (2004). In R. Krishnamurthy (Ed.), The OSTI report. London: Continuum.
Skorczynska, H. (2010). A corpus-based evaluation of metaphors in a business English textbook. English for Specific Purposes, 29, 30–42.
Stewart, D. (2010). Semantic prosody: A critical evaluation. London: Routledge.
Stubbs, M. (1995). Collocation and semantic profiles: On the cause and trouble with quantitative methods. Functions of Language, 2, 1–33.
Stubbs, M. (2001). Words and phrases: Corpus studies of lexical semantics. Oxford: Blackwell.
Stubbs, M. (2009). The search for units of meaning: Sinclair on empirical semantics. Applied Linguistics, 30, 115–137.
Swales, J. (2002). Integrated and fragmented worlds: EAP materials and corpus linguistics. In J. Flowerdew (Ed.), Academic discourse (pp. 150–164). London:
Longmans.
Vannestál, M., & Lindquist, H. (2007). Learning English grammar with a corpus: Experimenting with concordancing in a university grammar course. ReCALL,
9, 329–350.
Whitsitt, S. (2005). A critique of the concept of semantic prosody. International Journal of Corpus Linguistics, 10, 283–305.
Williams, M. (1988). Language taught for meetings and language used in meetings: Is there anything in common? Applied Linguistics, 9, 45–58.