Chapter 9 Results and Discussion Part One

9.1 Introduction

This thesis started with two central questions – what is Business English lexis, and how does the lexis of Business English published materials compare to it? It is the purpose of this chapter to look in more detail at both the questions themselves and at the answers found. In doing so, this chapter is divided into four distinct, but overlapping parts.

The first section will, after briefly reiterating the key hypotheses and questions involved in the study, focus on the linguistic aspects of the study of business lexis found in the BEC.
This will lead to the second section, where the linguistic aspects of the analysis of the PMC will be discussed.
In the third section, linguistic analysis of the BEC and PMC will be related to the pedagogical implications the research has for both materials design and, to a lesser extent, classroom practice.
The final, fourth section will present incidental findings gained, a critique of the methods used and proposals for areas of future study.

9.2 Hypotheses and questions

The hypotheses stated that a) the lexis used in Business English is significantly different from general English, and b) the lexis used in published Business English teaching materials is significantly different from the business language that is found in real-world business life. These hypotheses gave rise to two umbrella questions that cover the whole thesis: a) is there a lexis specific to Business English and, if there is, what is it? and; b) can significant lexical differences be found between the language used in published Business English materials and the language actually used in business? The methods of analysis employed in the thesis have been to a large extent quantitative, in that the description and definitions of business lexis and the lexis of business language materials have been arrived at through computer-based, corpus linguistic techniques. However, the work is also exploratory in nature and falls under the heading of descriptive linguistics. This descriptive aspect is stressed again here, as one aim of the study has been to map out the key lexical features of Business English. For this element of the research, qualitative methods have been employed. The first section of this chapter which now follows, concerns itself with the first of these umbrella questions – is there a lexis specific to Business English and if there is what is it?

9.2.1 Research questions relating to Business English lexis

This first main question, concerning just what Business English lexis is, subsumes a number of further, more detailed questions. These questions form the structure for the whole of the first section. They are therefore presented briefly here, followed by a more detailed treatment.

1. Is there such a thing as Business English lexis? and 2. If there is such a thing as Business English lexis, what is it made up of? In discussing this issue the concept of key words (Scott 1997, 1999) is utilised and the key words found for Business English are presented and discussed. Pickett (1986b) suggested that business language was ‘a linguistic world of ‘forms and frameworks’ of conventionalised transactions, governed by the courtesies and formalities of business life’ (1986b:2). Discussion here centres on the possibility of presenting a lexical ‘linguistic world’ of Business English. Scott’s (1997) paper on key words and how they can represent changes and attitudes in society is here applied to Business English with further reference to Williams’ (1978) earlier definition of key words.

These key words form the basis of further analysis into Business English lexis which examines how Business English lexis is defined at a variety of linguistic levels.

3. Can the concept of semantic prosody be found in Business English and if so are there business-specific prosodies? The idea of semantic prosody is interesting, but still largely untried in terms of large-scale collocational analysis. Semantic prosody has been defined both in terms of positive/negative prosodies (Sinclair 1991, Louw 1993), and in terms of lexical/semantic sets (Stubbs 1995, Tribble 1998, Hoey 1997, 2000), and this section looks at how semantic prosody has operated in the BEC.

4. What colligational patterns can be found in Business English and can grammatical patterning typical to Business English be identified? Sinclair (1991), Stubbs (1993, 1996), Hunston et al. (1997), Hunston & Francis (1998), Hoey (1997, 2000) and Hargreaves (2000), have all pointed to the link both between grammatical patterning and lexical meaning, and the fact that grammar and lexis are inseparable concepts. This view is accepted in this thesis and analysis considers what colligational/meaning patterns were discovered in the BEC. In this section the results found are related to the discussion seen in Chapter 3 on the nature of ESP and special languages (Hutchinson & Waters 1987, St John & Dudley-Evans 1998) and sub-technical language in general (Cowan 1978, Trimble 1985, Yang 1986, King 1989).

5. How are words distributed across Business English macro-genres and can they be divided along the ‘knowing’-‘acting’ axis of Pickett (1988)? Pickett (1988:91) suggested that Business English language could be divided into knowing and acting Business English. This thesis has looked at the possibility of dividing lexis into its likelihood of being used when talking about business (knowing), and when actually doing business (acting). Analysis takes advantage of the BEC, divided into macro-generic categories, and utilises the Dispersion Plot function of WordSmith 3.

6. What kinds of clusters can be found in Business English and do business-specific clusters exist? A discussion on the clusters of lexis found in the BEC centres around whether larger clusters can be found to be more genre-specific and less frequent, and smaller clusters more multi-functional and more frequent. Additionally, the doing business – talking about business axes are utilised to give a categorisation of the clusters that is broader than genre.

7. How do words associate with each other in Business English? Here Scott’s (1997, 1999) concept of associates is briefly discussed and the associates found for a selection of key words are presented and discussed.

8. Business English – an overview: Just as important as all these questions is the what of what Business English lexis is, as it has never been analysed on this level before. Thus the words found are just as important as any analysis carried out on them. This last section will summarise the answer to the question posed at the start of the chapter – what is Business English lexis? This will lead to Section 9.4 where the words found in the corpus of Business English materials, the PMC, have been analysed in a similar way.

9.3 Linguistic features of Business English lexis

The following eight sections now address each of the questions asked above in turn.

9.3.1 Is there such a thing as Business English lexis?

The answer to this question must be ‘yes’, but it must also be a qualified ‘yes’. Business English does not exist as a separate entity of its own using entirely its own lexis. It is, like all specialist lexes, tied to the general language that goes to form the most frequent words used in the language. In the lemmatised list of the 100 most frequent words found in the BEC, only seven words can be found that could be thought of as business-related. These are shown in Table XXXII on the next page. In total, when added together, their frequency in the BEC amounts to 14,489 instances, representing only 1.41% of the total corpus. It may look on the basis of these results that, in terms of frequency of use, there is no readily definable lexis of Business English.

TABLE XXXII: BUSINESS-RELATED WORDS FOUND IN THE TOP 100 MOST FREQUENT WORDS IN THE LEMMATISED BEC

N	WORD	BEC Freq.	BEC %	LEMMAS
38	COMPANY	2 934	0.29	companies(1092)
41	BUSINESS	2 837	0.28	businesses(287)
54	MARKET	2 336	0.23	markets(469),marketing(469),marketed(10)
56	WORK	2 234	0.22	works(226),worked(134),working(680)
84	SERVICE	1 461	0.14	services(641),servicing(43),serviced(5)
89	PRODUCT	1 385	0.14	products(644)
94	PRICE	1 302	0.13	prices(417),pricing(69),priced(20)

However, whilst pure frequency of lexis can be useful in linguistic analysis (Francis & Sinclair 1994) it is not the only criterion by which a specialist variety of language can be identified. It is argued here that a more accurate picture of both specialist languages in general, and Business English lexis in particular, can be gained by analysis of words that occur significantly more often or less in a particular linguistic area, in comparison to general language usage, rather than by looking at words that have a high occurrence in terms of overall frequency. These words have been termed key words (Scott 1997, 1999).

9.3.1.1 Key words

The notion of key words has been discussed earlier in this thesis, but a brief repetition will help clarify its use. The concept of a key word is defined as ‘a word which occurs with unusual frequency in a given text’ (Scott 1997:236). Scott’s definition here limited the definition to a single given text, but the notion of using unusual frequency is valid for the analysis of not just one, but a whole range of co-joined texts. Key words were arrived at in this thesis by using the key word function of WordSmith 3 on two large corpora. Here, briefly, is how it was done.

i) All the files that go to make up the BEC were placed together to form a single large computerised frequency word list.

ii) This frequency list was then lemmatised and cleaned up as described in Chapter 7, Step 1.

iii) The frequency list was then statistically compared to a corresponding lemmatised word frequency list created from a reference corpus of general English (the BNC Sampler corpus of 2 million running words). This was done by comparing each word’s frequency in the BEC and its frequency in the BNC, with account taken for the difference in size of the two corpora. The Log Likelihood statistic (Dunning 1993), was used and a very high p value was set at p = 0.000001. This value was set high in order to decrease the possibility of inclusion of erroneous key words. Moreover, by setting the p value so high, it also limited the number of key words obtained which, in turn, aided subsequent analysis of them.

iv) The result of the analysis gave words (key words) that are used in Business English statistically more frequently than in general English. The key words were then placed by the program into an order of ‘keyness’, the most key words at the top of the list. Scott explains:

A word will get into the listing here if it is unusually frequent (or unusually infrequent) in comparison with what one would expect on the basis of the larger wordlist.

(Scott 1999: Key Words Help File)

Positive and negative key words: Key words may thus also be both positive or negative. Positive key words are those described above – words that occur significantly more than in a reference corpus. Negative key words, in contrast, are words that occur significantly less in our corpus than in the reference corpus. Analysis in this thesis has used both positive and negative key words as one of the bases for a description of Business English. By this it is meant that it is possible to say as much about Business English by what is not found there – using the negative key words – as by what is found there – using the positive. The process of gaining the key words is shown in the diagram on the next page.

Business General

Language Language

Business British

English National

Corpus Corpus

Computerised Comparison

by WordSmith 3

Positive key words Negative key words

Unusually high frequency Unusually low frequency

in Business English in Business English

Fig. 31 The process by which Business English key words were arrived at

9.3.1.2 Positive key words in the BEC

Tribble (1998:7) noted that ‘The first feature you notice when comparing a keyword list with a frequency list for the same data is that they are usually very different’. This is true of the key word list gained from the BEC. The top 100 positive key word list for the BEC wasshown in Chapter 8 in Section 8.2.4. It is presented again below, this time excluding the non-business lexis and showing only the 49 clearly business-related words:

TABLE XXXIII: TOP 100 BEC POSITIVE KEY WORD LIST – BUSINESS-RELATED WORDS ONLY

N	Word	BEC Freq.	BEC %	BNC Freq.	BNC %	Keyness Log L.
1	BUSINESS	2 837	0.28	542	0.03	3 557.7
2	COMPANY	2 934	0.29	782	0.04	3 118.6
3	MARKET	2 336	0.23	831	0.04	2 056.1
4	CUSTOMER	1 199	0.12	147		1 763.0
6	PRODUCT	1 385	0.14	412	0.02	1 377.2
7	SALE	1 210	0.12	343	0.02	1 239.4
9	MANAGEMENT	973	0.10	279	0.01	989.6
10	PRICE	1 302	0.13	586	0.03	941.5
11	FINANCIAL	780	0.08	237	0.01	765.0
12	BANK	940	0.09	379	0.02	749.0
14	SERVICE	1 461	0.14	916	0.05	728.7
15	STOCK	889	0.09	350	0.02	722.5
16	ORDER	1 224	0.12	681	0.03	709.0
17	EXECUTIVE	529	0.05	86		707.3
18	CONTRACT	656	0.06	183		678.3
19	CLIENT	535	0.05	126		607.4
21	CONTRACTOR	326	0.03	16		582.3
23	MANAGER	742	0.07	317	0.02	562.4
25	SELLER	298	0.03	12		546.6
26	INVESTMENT	577	0.06	185		546.2
27	SHARE	1 148	0.11	762	0.04	528.8
29	COST	1 127	0.11	747	0.04	520.2
33	PROFIT	799	0.08	429	0.02	482.0
34	SELL	789	0.08	419	0.02	481.8
42	CORPORATE	277	0.03	33		410.6
44	BUYER	292	0.03	42		407.9
45	CREDIT	392	0.04	110		403.8
46	INDUSTRY	712	0.07	404	0.02	402.8
47	SUPPLIER	288	0.03	44		393.9
49	BUDGET	437	0.04	152		390.7
52	ACCOUNT	859	0.08	593	0.03	373.4
54	DISTRIBUTOR	218	0.02	16		363.5
55	DELIVERY	291	0.03	56		363.4
56	CASH	384	0.04	124		361.7
58	COMPANY’S	263	0.03	45		344.7
63	DIRECTOR	541	0.05	289	0.01	328.2
68	SHAREHOLDER	286	0.03	73		311.1
72	INVESTOR	248	0.02	51		300.7
74	EMPLOYEE	307	0.03	94		299.5
78	INVOICE	182	0.02	17		287.8
82	PAYMENT	321	0.03	115		280.8
83	TAX	629	0.06	427	0.02	280.3
84	TRADE	696	0.07	509	0.03	276.3
86	OFFICE	651	0.06	461	0.02	272.1
88	ENGINEER	368	0.04	163		269.9
89	MEETING	739	0.07	575	0.03	264.1
90	FIRM	466	0.05	265	0.01	262.9
91	FINANCE	298	0.03	117		242.7
96	PURCHASE	289	0.03	117		229.4
97	EXPENSE	236	0.02	73		228.7

Key: Log Likelihood statistics shown in ‘keyness’ column p=0.000001

There are at least two striking differences between the BEC frequency list and the key word list – number and content. In the top 100 frequency list only seven words were found that were clearly business-related. In the key word list, 49 words were found to be business-related. The content of the lists also differs: the frequency list is full of function words (e.g. the, and, but, because) and delexicalised verbs (e.g. get, go, know) with only a very small number (nine) of lexical, meaning-carrying words (company, year, business, market, people, service, product, price, system). The key word list, in contrast, displays almost totally lexical, meaning-carrying words. In addition to these words with a high business-related meaning, a group of words – sub-business words[158] – excluded from the list as being not pure business words, are words that could be intuitively expected to be found in a business environment, for example, fax, billion, global, project, performance, year, rate, agreement, group, offer and growth. The difference between the frequency and key word lists is summarised in the table below:

TABLE XXXIV: DIFFERENCES IN THE TOP 100 FREQUENCY/KEY WORD LISTS OF THE BEC

Top 100 frequent words in the BEC	Top 100 key words in the BEC
7 pure business-related words	49 pure business-related words
virtually no sub-business lexis	some sub-business lexis
higher number of function words	lower number of function words
abundant delexicalised language	little delexicalised language

Key words thus sit on top of, and are supported by, general language. They can be seen as a separate, but not independent, set of lexis bound to the situations and activities of business. The position they hold in comparison to general English is represented in Fig. 32 on the next page. It will be noticed that this diagram is reminiscent of Pickett’s (1986a:4) diagram (p.63) representing the specialised language of a particular business. The main difference here is that Pickett’s work was based purely on intuition: this time the key word concept statistically indicates that Pickett’s notion of the relationship of specialist language to general language is probably correct.

St John (1996) noted the lack of a core lexis of Business English and it is proposed here that the key words generated from the BEC represent lexis that is core to Business English. They are words used in business, by business people, to an unusual level of frequency in comparison to general English. However, it must be stressed that whilst these words can be seen as core, they should not be seen as the only possible core words of Business English – the BEC presents one picture of Business English and another Business English corpus may show a slightly different picture. This warning aside, the overall semantic homogeneity of the key words computed from the BEC suggests that this core lexis is accurate and reflects the lexical world of business with a good degree of accuracy.

Business Key Words

unusually high frequency

lexical meaning-carrying words

General English

non-business function words

increased delexicalised language

Fig. 32 The relationship of Business English key words and general English

The next section: Thus far it has been shown that key words for business exist and that they tend to be lexical and meaning-carrying in content. It has been argued that these key words represent lexis that is central to Business English. The next stage further analyses both positive and negative key words in order to semantically define the ‘world of business’ and also lay the groundwork for later pedagogical application.

9.3.2 If there is such a thing as Business English lexis what is it made up of?

Initially, it was planned to analyse only the most ‘key’ key words found in the BEC, i.e. those words that are presented above as being the most ‘key’ words of Business English. However, a superficial analysis was enough to see that the words could be categorised into semantic sets – that is, the words seemed to fall mainly into a limited number of semantic categories.[159] It was therefore decided to assign all the key words gained into these semantic sets.[160] There were in total 1,611 key words remaining after the clean-up process. Of these, 925 were positive key words and 686 were negative. Whilst this was a manageable number for manual analysis it was felt that the semantic groups the words could be assigned to would become too large to handle easily, so the positive, and then negative key word sets were first assigned to their appropriate word classes. This categorisation followed the model of Ljung (1990), who used the following word class categories: noun, verb, adjective, noun/verb, noun/adjective, verb/adjective, noun/verb/adjective and -ly adverbs. These categories are not ideal, but owing to the automatic nature of the lemmatisation process it was not possible to assign all the words to completely separate word classes. Ljung noted that ‘In the absence of…manual editing, the resulting lemmatized lists are naturally a halfway house between a simple list of word types and real lemmatization’ (Ljung 1990:5). For example, in the BEC, fax can act as both a noun and a verb, executive as both a noun and an adjective. The categories represent, therefore, a compromise between time, accuracy and the resources available.

9.3.2.1 Positive key word analysis

This section of the discussion analyses the semantic sets found for the four largest word class categories found – as seen in the table below: nouns, verbs, adjectives and noun/verbs.

TABLE XXXV POSITIVE KEY WORDS GRAMMATICAL CATEGORISATION

a) Nouns

As can be seen from the table, there were 440 positive key nouns, and it was possible to divide them into ten semantic groups. The semantic categories identified are shown in the diagram below (for a full picture of the word class groups and the semantic categorisation of positive BEC key words, please turn to Appendices 2 and 3 in Vol. II).

companies &

people events institutions activities

technology/computers Nouns places

money/finance things states & qualities measures & amounts

Fig. 33 Semantic noun categories of BEC key words

Four initial points can be made about these key nouns: they displayed a remarkable business focus; they were tangible as opposed to abstract; they tended to be impersonal rather than personal; and they were positive in connotation as opposed to negative. Now each category is discussed in detail.

1. People: The most key examples from this category included customer, management, executive, contractor, manager, seller, buyer, supplier, distributor, director and shareholder, and the homogeneity of the group was quite remarkable. Out of the 67 people or groups of people referred to in the key word list, only nine (13% of people found) could be considered as not being directly involved with the business world: allies, attendees, loser, guy, cavalry, prosecutors, chef, neighbours, maker, and some of these would not be seen as particularly out of place in business (chef, loser, attendees and perhaps, unfortunately, prosecutors). It must be stressed here that all the people found in the key word list were categorised here, not just the business-related people. The list thus quite clearly delineates the people, or groups of people, involved in business from those not.[161] This will become more apparent when analysis of the negative key words is completed.

2. Companies & institutions: There were 36 references (8%) amongst the key word nouns to companies/institutions representing a distinctly recognisable semantic group. Key examples from this category were company, industry, organisation, airline, telecom, plc, EU, subsidiary, Inc and consultancy. Two points need to be raised here. Firstly, the significance of this category being found at all – it is clearly indicative of the business bias of the key words and, secondly, the relative homogeneity of the institutions found – the institutions computed are mostly business-related. The only totally non-business word found in the list was faculty, and that was one of the least ‘key’ key words in the set. The abbreviation EU could also be considered to be non-business, but it has been at least partially mentioned in the corpus in relation to business in Europe. A sample can be seen below, showing EU surrounded by business-related language:

3. Activities: There were 28 words assigned (6% of nouns) to this category and the most key activities found included business, delivery, transmission, development, production and communication. In this category there were few purely business activities – business, administration and takeover being the only clear examples (10% of the sample) – but the remainder (25 words – 90%) were words referring to activities that could be thought of as being in the background of running a business, for example, delivery, development, competition, administration.

4. Things: This was a large category (151 words – 34%) and displayed a very strong emphasis on tangible items. The number of tangibles was by far the largest (127 words – 84% of the category sample) and included words such as product, auto, vehicle. The more abstract words were much fewer (24 instances – 16% of the category sample), and included opportunity, culture, scope, solution and basis.[162] This is an interesting category in that it can be seen as a category that shows the what of what can be found in the business world, both in concrete and abstract senses. It included words that could intuitively be expected to be in the background of running a business, e.g. opportunity, segment, sector, information, unit and capability, whilst words found here that could not be related to the business world in any reasonable way were pets, membrane (actually used in the corpus when describing a product) and souvenir.

5. States & qualities: The states and qualities found here were overwhelmingly positive, including examples such as growth, skill, leadership, competence, excellence, commitment, improvement, stability, success, strength and efficiency. This was contrasted by a much smaller number of negatives – the only overtly negative words were debt, loss, liability, inflation, slowdown, downside and insolvency. Thus, out of 43 nouns in the category, only seven were negative in meaning (16% of the group), whilst 36 (84% of the group) were either overtly positive, or at least neutral. Once again, the relative homogeneity of the words found is striking – they are all words that could be expected to be found used in a business environment.

6. Measures & amounts: The words in this category are words that refer to quantities and measurement, for example, of money – billion, million, trillion – and time – year, month and week. It is a small category, only 12 instances (2% of key nouns). The category indicates a focus in business on large numbers, for example, in terms of money, only the very high-end numbers are included as being key (million, billion, trillion).

It could be deduced from this, that lower-end numbers – the tens and hundreds – are used equally in both the business and non-business world, and the business world deals more with very large numbers. Examples of these high-end numbers in relation to money, taken from the BEC, are shown below:

7. Places: Sixteen ‘places’ (3% of key nouns) were included in this category, showing the key places where business takes place. The majority of the words (9 instances – 56% of category) refer to places belonging to the business world, e.g. office, premises, department, division, boardroom, depot, marketplace, and the remainder point to more general aspects, e.g. street’s, country, world’s, but need not be considered out of place in a business context. The word hotel is also featured here.

8. Events: In all, 23 events[163] (5% of nouns) were included in this category showing events that are central to business life. They included a high number of purely business or work events: sale, merger, bankruptcy, transaction, arbitration, demerger, promotion, privatisation, deregulation and several less directly business-related events, e.g. appraisal, valuation, termination, retention. This group is notable for the positive/neutral aspects of its lexis, with only one overtly negative word – bankruptcy – identified.

9. Money/finance: This was a relatively large category[164] (40 words – 9% of the key noun sample) and was notable for the impersonal quality of the lexis found. The words were centrally concerned with the financial aspects of running a business, e.g. expense, earnings, revenue, margin, salary and equity. There is, therefore, little in this group relating to the individual person, with words being largely descriptions of financial aspects of companies, for example, payroll, cashflow, turnover and maturities.

10. Technology/computers:Words found here (21 words – 4% of key nouns) reflect the importance of computers and technology in modern business life – Internet, PC, software, browser, web – and also that fact that there are many high profile businesses that deal in these areas.

b) Verbs

The verbs found in the positive key words were quite few in number (93), and so can be reproduced (in Table XXXVI on the next page).

TABLE XXXVI: SEMANTIC CATEGORISATION OF POSITIVE KEY VERBS IN THE BEC

Negative	Personal & Inter-personal	Neutral	Work/ Business	Money	Running a business
complicates	please*	include	sell	liquidate	receive
incur	advise	regard	manage		provide
frustrate	agree	continue	deliver		operate
	discuss	underlie	confirm	Technology	send
	expect	combine	enclose		require
Positive	announce	tend	invest	automates	develop
	relate	follow	restructure		generate
excite	motivate	identify	underwrite		consolidate
achieve	inform	compare	compete		certify
improve	propose	alleges	merge		notify
		exist	advertise		maintain
		accelerate	customize		integrate
		totaled	employ		establish
		await	publish		involve
			despatch		implement
			earn		enhance
			buy		consult
			shelve		ensure
			promote		expand
			decentralize		specify
			downgrade		attach
			designate		authorize
			negotiate		create
					perform
					acquire
					submit
					revise
					transmit
					approve
					recycle
					participate
					locate
					strive
					allot
					inquire
					orientate
					tighten
					comply

* Included as a verb in this study despite its more common adverbial/pragmatic function.

As can be seen from the table above, the verbs have been divided into eight categories, with the three largest being neutral, work/business and running a business. In the work/business section the verbs included were verbs intuitively closely associated to business life, e.g. sell, manage, manufacture and invest. Those verbs included in the running a business section were those that could intuitively be expected to be found in a business environment, e.g. receive, provide and operate. When these latter, business-related categories, were added together, they amounted to almost 66% of the verbs, showing their strong business focus.

Further, it will be seen later that verbs such as receive, though not directly ‘business’ verbs, are in fact lexically integrated into the business world and imbued with a business sense by collocational and semantic prosodic relations. The verb receive, for example, has strong prosody with money, stocks and shares (32.29% of all instances) – a brief example from the BEC showing how receive collocates with money-related lexis is given below:

Thus a network of meaning is created by not only the verbs themselves, but by the strong business-related collocative environment in which they are found and in which they operate. A further feature of note with regard to these verbs is the fact that they are to a large degree lexical verbs, that is, they carry meaning. This points to the fact that all language use includes a limited number of extremely frequent delexicalised verbs (Sinclair 1991) such as get and have that are equally common to both business and non-business language use. Delexicalised verbs, therefore, are not found as key words because their usage is similar in both.

This imbalance of lexical and delexical verbs in the BEC is illustrated by examining the 100 most frequent words in the BEC. It was found that there were only four business-related, meaning-carrying verbs in this group of words[165] (market, work, service, price), with a total frequency of 7,333 occurrences. Conversely, there were 23 de-lexicalised verbs, totalling 92,032 occurrences. This relationship of lexical and delexicalised business verbs in the BEC is summarised in the diagram below:

Fig. 34 Relationship of business to delexicalised verbs in the 100 most frequent words of the BEC

c) Adjectives

Based on the results of analysing the adjectives found amongst the key words of the BEC, the business world seems to describe itself in overtly positive but impersonal terms. It was possible to divide the adjectives into nine groups: size/speed, places, positive, negative, neutral, work/business, money, technology and time. Of these, the largest was that of positive adjectives and the list is given in full here:

Positive adjectives found in the BEC Key Word list:

new, mobile,[166] best, successful, available, independent, relevant, appropriate, responsible, outstanding, systematic, effective, exclusive, flexible, important, sustainable, strong, comprehensive, consistent, exceptional, kind, favorite,[167] efficient, innovative, dynamic, accurate

By contrast, only three openly negative adjectives could be found: defective, critical and aggressive and, as West (2000, personal communication) pointed out, aggressive has taken on positive connotations in business. This comment is borne out by some examples from the BEC:

The adjectives noted here, though distinctly positive in content, are also largely impersonal and non-emotive in essence, e.g. independent, relevant, appropriate, systematic and comprehensive.

Work/business adjectives: The second largest group of adjectives was that related to work or business life: corporate, strategic, competitive. This group was made up of a high percentage (8 instances) of -al ending adjectives (47% of the group), e.g. internal, organizational, promotional, industrial, operational, managerial, confidential, entrepreneurial.

Size/Speed adjectives: This semantic group emphasised a feature of the business lexical environment already seen with the nouns – a focus on the fast or high-end of description: high, big, large, rapid, multi, fast, broad, mega, unlimited. Conversely, there was much less slow or low-end description with low and slow being the only examples. The group included words that can also act as nouns (high, low) but are almost always used as adjectives in the BEC.

Places adjectives: This group contrasted the near with the far in business terms: domestic, local versus overseas, continental and stressed the geographical reach of business: global, international, worldwide.

The remaining three groups of interest were concerned with money, e.g. financial, monetary, fiscal, leveraged; technology, e.g. digital, technical, electronic; and time, e.g. monthly, quarterly and daily.

d) Noun/verbs

Ten categories of differing sizes were arrived at for the noun/verbs and are presented in the diagram below:

places & people communication positive negative neutral

time Noun/Verbs running a business

work/business money technology

Fig. 35 Semantic noun/verb categories of BEC key words

The categories represent a compromise – when a word is used as a noun or a verb it takes on, of course, different meanings and functions within the language. For example, in the case of the noun/verb defect – there is no semantic link between to defect and a defect. Often, however, a semantic link between the word classes does survive, for example, to dispute and a dispute, to fax and a fax. This being said, the accuracy of assignation of words into semantic sets in this category suffered as a result of this duality of word class. However, as the category was the second largest, after nouns, some categorisation was essential. Therefore, the words have been placed intuitively into groups using a principle of ‘most likely to belong’ to a group. If one uses the examples given earlier of defect, in the context of a Business English corpus, it is more likely to refer to a defect in, for example, a product, rather than to a person defecting to another country. Along these lines, the words were allotted the following groups:

Places & people: These were all either directly or indirectly business-related including examples such as partner,[168] board, staff, boss and rep.

Communication: This was an interesting group in that the words concerned with communication found amongst the positive key words of the BEC seem to have both a verbal and nominal function, and stood out here as a clear semantic set – fax, mail, telephone, meeting, report, telefax, email, cable, phone and telex. The instruments used for communication seemingly generate the verbs by which the act of communication becomes known, in the same manner that biro and hoover have become embedded in the English language.

Positive & negative: These were very few in number – four positive: experience, increase, gain, progress and three negative: dispute, defect and risk. Thus, whereas the pure nouns displayed a large number of positive attributes in business, the noun/verb class did not.

Work/business & running a business: These two categories were virtually equal in size (work/business 37 words and running a business 34) and there must be a certain degree of overlap between these two groups as they have been separated intuitively. At the pure business end there are words such as outsourcing, retail, market, export, franchise, wholesale, and at the other end business-related words such as plan, value and test, which can equally be used in non-business life as in business. There is then a middle ground with words such as offer and guarantee which could perhaps be placed in either category, i.e. a noun/verb connected to work/business or a noun/verb commonly used in the running of a business. This is shown in the figure below:

Running a business >>> Pure business

plan <<< offer >>> outsourcing

value guarantee export

Fig. 36 The sliding scale of business-related key noun/verbs in the BEC

The two categories combined represent a large portion (35%) of the noun/verb category, indicating once again the business emphasis of the key words.

Money: As with the communication category, words connected in some way to money and finance have both a nominal and verbal function, for example, price, bank, cost, profit, budget, cash, invoice, tax and finance. This category also included items that have separate meanings dependent on word class, e.g. interest, stock and bond.

The remainder of the noun/verb category was made up of neutral noun/verbs, six related to technology and a small section related to time, e.g. date, schedule, due and delay.

As a result of the above analysis it can be seen that the positive key words found in the BEC fall, to large degree, into a limited number of recurring semantic groups. These semantic groups span grammatical word classes, and though distinct in their extremes, are fluid at their boundaries with words sometimes being able to appear in one group or the other. In all, the lexis of Business English found here presents a clear picture of the world of business – its people, its institutions, its activities and events. At the same time, however, ‘neutral’ lexis is also used, pointing to the fact that Business English should not be considered only in these limited terms, but also as being linked to non-business language. Thus Business English lexis can be seen as one of a number of pools of specific lexes, fed by the broad river of general English, but also possessing a diversity and eco-system unique to itself. These ideas can be seen in Fig.37 on the next page.

It is also possible at this stage to relate the findings on Business English to previous work carried out on specific language. Biber (1988), working on the linguistic variation across genres, identified six factors by which they could be distinguished from each other in terms of their textual characteristics. Biber stated that

The interpretation of the factors is based on the theoretical assumption that these co-occurrence patterns indicate an underlying communicative function shared by the features; that is, it is assumed that linguistic features co-occur frequently in texts because they are used for a shared set of communicative functions in those texts. (Biber 1988:101)

People Communication Positives

customer phone successful

manager fax available

supplier email outstanding

shareholder report effective

Money Technology Business

interest digital market

price electronic service

budget graphical corporate

invoice multinational

Things Business English Places

product office

agreement premises

property department

equipment boardroom

Running a business Time Events

plan monthly sale

forecast quarterly merger

handle daily bankruptcy

book transaction

Institutions Activities Measures

company business million

subsidiary delivery billion

corporation development trillion

conglomerate production

GENERAL ENGLISH

Fig. 37 The main semantic groups that go to form key Business English lexis

Biber’s work was, therefore, concerned with genre, and it is not suggested here that Business English is a genre per se,[169] yet the factors identified by Biber can be of use in analysing the results of the BEC. The first of Biber’s factors – informational versus involved production – noted essential differences between texts that are information-bearing (a stress on nouns, longer words) and texts that are representative of involved production, i.e. with ‘affective, interactional, and generalized content’ (Biber 1988:107) – examples of the latter feature being high relative use of first and second personal pronouns, contractions and private verbs. For the time being, the only observation necessary to make is that the Business English positive key words under examination displayed a distinct lack of the features relating to involved production. Thus, whilst Biber’s factors are only briefly mentioned at this point, they will be returned to in relation to both negative key words and the key words of the PMC.

Analysis of Business English, therefore, does not stop here. The positive key words gained from the BEC have allowed us the opportunity of seeing what Business English lexis is. The negative key words further mark out the lexical territory of Business English by showing what it is not.

9.3.2.2 Negative key word analysis

Negative key words represent those words that were found to occur with an unusual infrequency in the BEC. They were, as with the positive key words, first divided into word class groups along the lines of Ljung (1990). The results are shown in Table XXXVII on the next page.

Each category was then further divided into semantic and lexical sets. These will now be discussed in more detail. It is important to remember here that the term negative key word does not mean that a key word has negative connotations, but rather that it occurred in the BEC significantly less than would be expected (to a pre-set statistical requirement) when compared to general English. A full list of all negative key words and their semantic groupings can be found in Appendices 4 and 5 in Vol. II.

TABLE XXXVII: NEGATIVE KEY WORDS GRAMMATICAL CATEGORISATION

a) Negative nouns

The differences between the lexical groups found in the positive key nouns and the negative were stark. In all, eleven semantic categories were identified, and despite the fact that several of these eleven categories found were the same as those for the positive – people, institutions, activities & events, things, states & qualities, and places – the content of the categories was radically different. Additionally, five new categories were identified – food & drink, house & home, time, earth & nature and parts of the body. The semantic groups are shown below, with the new groups in italics:

people institutions activities & events things

states & qualities places

Key Negative Nouns

parts of the body earth & nature

food & drink house & home time

Fig. 38 Semantic noun categories of BEC negative key words

Each of the categories will now be discussed and compared to lexis found in the positive key words.

People: In this category – the largest – a clear divide between the business and the non-business world is statistically marked out by the key word analysis. Forty words were found that were related to people (20% of all negative key nouns)[170] and of these, certain sub-semantic sets could be distinguished. Firstly, there are many words concerned with family – mum, mother, child, dad, son, brother, daughter, sister, daddy, husband, wife, aunt – and there was a smaller set linked to people – man, child, boy, girl, lady, baby, lad. Other groups focused on royalty/aristocracy: lord, king, queen; on religion: god, christian, soul; and on professions: soldier, artist, officer, student, waiter, MP, secretary. Whereas the positive group of peoplewas virtually all business-related, this group had no obvious business connections except, perhaps, secretary. Even the professions that were included here were non-business, e.g. soldier, artist, officer, MP. The semantic divide is shown the diagram below:

Negative Key Words: People Positive Key Words: People

man God executive contractor shareholder

mum mother management customer

child dad daddy supplier distributor director

baby Queen wife manager seller buyer

Fig. 39 People featured in positive and negative key words

Institutions: There were only seven instances here (3% of negative key nouns) church, army, hospital, council, EC and also business-related institutions union and FT.[171] The abbreviation EC is now less used than EU, which was a positive key word, and thus appeared more in the older BNC (i.e. the BEC is more modern than the BNC Sampler corpus and so uses the abbreviation EU more frequently than the older EC).

Food & drink: A small (10 instances – 5% of negative key nouns) but distinct semantic set – fig, egg, chocolate, tea, bread and dinner are some examples.

Activities & events: The activities and events found here (10 instances – 5% of negative key nouns) mostly relate to everyday life – birthday, Christmas, marriage, election and prayer being examples. Intuitively they feel like those events and activities that affect people personally, and occur outside their work. Again, they contrast with the strong business focus of the positive key words such as business, delivery, sale and production.

House & home: Continuing with the personal theme is this group (15 instances – 7% of negative key nouns), in which the words are all related to the house, and things found in the home: curtain, bed, door, garden, home, bedroom, toilet, bathroom and kitchen being prime examples.

Things: This contained the most diverse group of nouns (29 instances – 14% of negative key nouns) and there were no business-related words. The group included abstract nouns such as expression, intelligence, behaviour, character and preference and concrete nouns such as ball, sprinkler and railway.

States & qualities: In the positive key word section, this category contained words mostly positive in connotation such as growth, skill, leadership and competence. The words found here (only 11 in all – 5% of negative key nouns) in the negative key words seem to mostly connect to important ethical questions – those in some way relating to the meaning of life – death, life, war, peace, truth,[172] age and faith. The other words found here were motion, length, unity and joy.

Earth & nature: This category further lexically divides the non-business and business worlds with a stress on substances, e.g. diamond, ash and mercury, on animals, horse and cat, and on nature, e.g. tree, river and leaf.

Two further small categories included were time (8 instances) and parts of the body (11 instances). In the time category, which included words such as night, morning, century and summer, the most interesting find was that only two days were mentioned – Saturday and Sunday – the two days of the week when the business world traditionally rests. Thus the key word function statistically confirms what would ordinarily be only intuitively assumed. The parts of the body category had 11 words including leg, mouth, blood, skin, nose, tooth and heart. The final category – places – contrasts with the corresponding category in the positive key words as can be seen in Fig. 40 below. The negative key words are linked to town: library, ward, district; countryside: bay, forest, cottage, hill, sea; and even above: heaven.

Negative Key Words: Places Positive Key Words: Places

town county village office premises department

ward palace library division boardroom depot

opera prison castle marketplace

Fig. 40 Places featured in positive and negative key words

It can be seen from the comparison of the positive and negative key word nouns that a firm divide is established between the business and the non-business world, in terms of the lexis and related concepts that are found there and the lexis that is not. In the next section, verbs, the contrast is equally marked.

b) Verbs

There were very few negative key verbs so they can be presented in full below:

TABLE XXXVIII: SEMANTIC CATEGORISATION OF NEGATIVE KEY VERBS IN THE BEC

Negatives & Modals	Neutral	Personal	Interpersonal
didn’t	get	know	say
can’t	gonna[173]	see	tell
couldn’t	do	eat	hear
haven’t	come	remember	lie
wasn’t	put	die	marry
could	read	pray	elect
won’t	wear	sit	listen
weren’t	hang	think	ask
wouldn’t	burn	wanna	pretend
isn’t	fetch	feel
hasn’t	have	suppose
	seem	born
		tire
		forget
		condemn
		reckon
		observe

It can be seen immediately that they differ in both form and content from the verbs found amongst the positive key words. Four groups were identified: negatives & modals, neutral, personal and interpersonal. The negatives and modals group contained just that, e.g. didn’t, can’t, couldn’t and haven’t. The two other categories of interest – personal and interpersonal – consisted of verbs referring to things that people can do or have either individually – know, see eat, remember, die, pray – or with other people – say, tell, hear, lie, marry, elect, listen. Included were verbs of negative connotation – die, tire, condemn, lie and pretend. Additionally, the negative key word verbs are indicative of Biber’s factor of involved production discussed briefly at the end of the positive key word section. Features of involved production – i.e. interactive and affective situations – are use of private verbs (Biber gives think and feel as examples of private verbs) and contractions. Biber noted that these verbs are used for ‘the overt expression of private attitudes, thoughts, and emotions’(Biber 1988:105). In the BEC negative key words this is seen in words such as know, see, say and tell.

This emphasis on the personal and the interpersonal activities of people suggests that the lexis of business is divorced from the emotion that these negative key words display. It may be slightly surprising to see verbs such as say, tell, hear and ask less used in business lexis, but many of the other verbs have an emotive quality that is seemingly out of place in business life, e.g. hang, burn, die, pray, marry and born.

The positive key word verbs are in contrast with the negative key verbs for at least two reasons. Firstly, the grammatical composition of the words – the negative contracted verb forms found here were missing in the positive key word category – the verbs in the positive category were all positive and uncontracted. Secondly, when considering the verbs from a semantic aspect, it can be seen that the majority of the positive key word verbs were practical, business and action-related verbs, for example, sell, manage, manufacture, deliver, confirm, enclose, invest, restructure, underwrite and compete. These contrast with the negative: can’t, won’t, couldn’t; the emotive: hang, burn, die; and the personal; marry, born and pray, found in the negative key word verbs.

c) Adjectives

Once again, these are relatively few in number and so are presented in full below. As can be seen, six sets were used to categorise the negative key word adjectives. The largest group (17 instances) is that of adjectives with a negative connotation. There is even a sub-group of these adjectives that refers to an extreme of unpleasantness – terrible, horrible, awful. This contrasts with a group of adjectives of positive connotation which refer to a positive extreme – beautiful, wonderful and great.

TABLE XXXIX: SEMANTIC CATEGORISATION OF NEGATIVE KEY ADJECTIVES IN THE BEC

Size	Positive	Negative	Neutral	Age	Colour
little	nice	bloody	rural	old	green
thin	lovely	dead	political	young	white
tall	beautiful	dark	wet	ancient	red
tiny	funny	cold	loud	elderly	black
	bright	blind	northern		pink
	wonderful	terrible	east		yellow
	true	strange	asleep		grey
	calm	horrible	soft		blue
	great	sorry	alive		silver
		sad	quiet
		awful	agricultural
		fat	religious
		boring	heavy
		guilty
		stupid
		mad
		angry

These adjectives are, as with the other word classes seen so far, in marked contrast to the corresponding category in the positive key words. In the positive key word list, the largest group of adjectives was adjectives of positive connotation, and these tended to refer to positive, non-emotive, and often non-personal attributes, for example, new, available, relevant, appropriate, systematic and effective. The adjectives of positive connotation found amongst the negative key words are firstly much fewer in number, and secondly, are words which tend to be used to describe people, for example, nice, lovely, beautiful, funny, bright and wonderful. It could be suggested, therefore, that the two sets of adjectives are essentially being used to describe different things: the positive key word adjectives are used to describe business, products and the business world, and the negative key word adjectives to describe people and emotions. They can be seen on the cline shown in Fig. 41 below.

It must be stressed here that this representation is very fluid – the positions of the words on the scale are not absolute, but simply attempt to represent the essential differences and partial convergence of the two groups of adjectives.

Positive Key Words >>>>>> Negative Key Words

BUSINESS/ >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> PERSONAL

PRODUCT

available successful bright nice

relevant outstanding true lovely

systematic flexible beautiful

effective exceptional funny

efficient calm

comprehensive

innovative

> > > > > >

Fig. 41 The sliding scale of key word adjectives

d) Noun/verbs[174]

In all, nine semantic categories of key negative noun/verbs were determined, the largest of which was that of neutral words. However, several interesting categories were recognised: people, war, food & drink, body, negative, animals, places and personal. The category of noun/verbs is represented in the diagram below:

people war food & drink

personal Negative Key Words body

noun/verbs

negative neutral animals places

Fig. 42 Semantic noun/verb categories of BEC negative key words

People: This semantic category continued the trend shown amongst pure nouns in the negative key words by being totally non-business: police, father, doctor, captain, minister and kid being examples (9 words – 4% of negative noun/verbs).

Food & drink: A distinct group (13 words – 6.7% of negative noun/verbs) was found that referred to food or drink, to the act of eating/drinking or the utensils/crockery used to eat/drink, for example, water, cup, milk, bowl, drink, cook, cream, salt and sugar.

Negative: This category was very small (5 words – 2% of negative noun/verbs), but very negative – murder, rubbish, sick, fear and hurt.

Three further small categories were found: animals (e.g. dog and fish), places (e.g. house, school, plan, park, jail) and war (e.g. kill, gun, fight, troop, bomb, shoot). The final category, personal, was larger (25 members), and consisted of noun/verbs that relate to people and what they do, for example, like, love, look, watch, walk, laugh, shout, cry and smile.

The negative key word noun/verbs follow the same pattern as the other negative key word categories in their non-business focus and a preoccupation with personal and societal attributes. The analysis so far has, to a large extent, been practical and descriptive in nature. The next section attempts an overview of the findings in relation to the literature and takes a more theoretical approach.

9.3.2.3 Key words and the ‘world of business’

Key words – the approach used in this thesis to identify core business lexis – is based on ideas that go back to the 1930s and J.R. Firth. Stubbs (1996) refers to Firth’s (1935/57) notion of focal or pivotal words, and both Stubbs (1995) and Scott (1997) refer to later work on key words and the meaning of vocabulary done by Williams (1976). Williams’ interest in vocabulary had been aroused by his experience of returning to Cambridge University after a prolonged spell in the army, serving in World War 2. On his return he found it difficult to fit in, and when discussing this situation with an old colleague who was in a similar position, both men commented that the new people at the university ‘don’t speak the same language’.[175] This was not meant literally, of course, but that, as Williams noted, ‘we have different immediate values or different kinds of valuation, or that we are aware, often intangibly, of different formations and distributions of energy and interest. In such a case, each group is speaking its native language, but its uses are significantly different’ (Williams 1976:9). Williams was, therefore, interested in the meanings of words and how they have changed over time and the differing connotations they have. He believed that these keywords helped to define culture – the changes in their meanings embodied the changes in society. Williams’ starting point was his own intuition and he did not have available the computerised corpora now in use to back up his ideas. His belief was that the keywords pointed to social attitudes – these keywords illuminated and shed light on the cultural world. In this thesis, the key words, albeit arrived at by different means, point not to culture in general, but to the semantic environments of business – they are indicative of and embody the ‘culture’ of business.

This idea of a ‘world of business’ is not a new one. Pickett (1986b), noted earlier, had suggested that business language was ‘a linguistic world of ‘forms and frameworks’ of conventionalised transactions, governed by the courtesies and formalities of business life’(1986b:2). The key word analysis has found that this ‘world’ is clearly marked out in terms of the lexis used, i.e. what lexis is used and, conversely, what lexis is not used, or used to a much lesser extent. The table below shows how the world of Business English lexis is distinguished from the lexis of the everyday world:

TABLE XXXX: BUSINESS LEXIS VS NON-BUSINESS LEXIS: POSITIVE AND NEGATIVE KEY WORDS

Business Lexis (positive key words)	Non-Business Lexis (negative keywords)
1. People: from the business world: customer, contractor, manager, seller, buyer	1. People: family, royalty, domestic relations: man, mum, wife, dad, baby, Queen
2. Institutions: Companies and business institutions: company, industry, airline, telecom	2. Institutions: Societal: church, army, hospital, council
3. Things: business-related, concrete: product, property, equipment	3. Things: diverse: horse, cat, diamond, glass, river abstract: expression, intelligence, preference
4. Places: Business-related: office, department, boardroom	4. Places: House and home: curtain, bedroom, bathroom, kitchen Countryside: bay, hill, sea, forest
5. Days of the week: Not key	5. Days of the week: Saturday, Sunday
6. States & qualities: business-related and positive: growth, stability leadership, competence	6. States & qualities: ethical questions / meaning of life death, life, war, peace, truth, age, peace
7. *Dynamic public* verbs*: sell, manage, manufacture, deliver, confirm*	7. *Personal and interpersonal private* verbs*: know, see, pray, feel, die, lie, marry*
8. Positive impersonal adjectives: new, best, successful, available, relevant	8. Positive and negative personal adjectives: nice, lovely, beautiful bloody, dead, dark
9. Money: focus on money/finance: cashflow, VAT, GDP, capital, earnings	9. Money: No/little mention of money/finance – only mentions: quid and pound
10. Activities: business-related: investment, payment, development, production	10. Activities: personal, family related: birthday, Christmas, marriage, prayer

It is important when viewing the demarcation of business lexis from non-business lexis, however, that the key words are seen in terms of tendencies rather than absolutes. The key word analysis shows which words occur with unusually high frequency, not that they are the only words that occur. This thesis, therefore, does not attempt to say that the lexis found amongst the positive key words is the only lexis of Business English, but rather that these words are much more likely to be used in a business environment than other words. That being said, the semantic delineation formed by the key word statistic between the business and non-business world is sharp and can, in general terms, be seen along a series of dichotomous axes: business vs society, positive and shallow states & qualities vs conflicting and more philosophical states & qualities, emotive vs non-emotive and dynamic actions vs reflection. These are presented in the chart below:

High frequency in Business lexis Low frequency in business lexis

business ——————-> society

positive ——————-> positive & negative

shallow ——————-> deep/philosophical

dynamic ——————-> reflective

non-emotive ——————-> emotive

Fig. 43 The axes that delineate Business English lexis

A fuller discussion on the lexical world created here will take place later in the chapter when the implications of the next sections of analysis can be added to the discussion.

9.3.2.4 The next section

The key word analysis has presented lexis at the level of the word: it has told us which words are found with an unusually high frequency and which words appear with an unusually low frequency in Business English in comparison to general, everyday English. Further analysis now turns its attention to lexical relations beyond the single word level, to look at how words typically collocate in Business English, forming networks of meaning via the concept of semantic prosody.

9.3.3 Can the concept of semantic prosody be found in Business English?

The concept of semantic prosody, which ‘occurs when a word associates with a particular set of meanings’(Hoey 2000:232) was discussed in full in Chapter 4. Although the idea of semantic prosody is interesting, and stands up well in the few articles in which it has been mentioned, it has not so far been used in any large-scale lexical analysis. Furthermore, knowledge of semantic prosody with regard to Business English has so far been very limited.[176] This thesis has used the concept of semantic prosody to gain an overview of the lexical relations that business words habitually form.

Semantic prosody is here utilised as part of a wider analysis which makes use of some of the five questions asked by Hoey (1997:1) of any word at the centre of a concordance line. These were laid out in Chapter 7 but are repeated here. The five questions were:

1) What lexical patterns is the word part of? (this refers to collocation)

2) Does the word regularly associate with particular other meanings? (semantic prosody)

3) What structure(s) does it appear in? (colligation)

4) Is there any correlation between the word’s uses/meanings and the structures in which it participates?

5) Is the word associated with (any position in any) textual organisation?

In this study, questions 1-4 are investigated in terms of business lexis. Questions 1 and 2 are joined together, in that by using semantic prosody, a broader and more structured picture of collocations is provided.[177] The following section will now take examples of words analysed in depth for semantic prosody in the BEC and discuss the broader implications of the findings.

9.3.3.1 Analysis of business lexis by semantic prosody

With nearly one thousand key words to choose for analysis, the temptation to choose just the most ‘key’ was great. However, as was noted in the previous section, the way the key words seemed to naturally fall into semantic categories suggested a more meaningful and ultimately pedagogical approach. To this end, five semantic categories were chosen to embody the key words found: people in business, business activities, business actions, business description and business events and entities. For each of these categories, ten words were chosen for analysis. This choice was based on frequency – those words that occurred more frequently would offer more fruitful examples for analysis – and on intuition – those words that seemed to be interesting in themselves. Thus the criteria were both statistical and intuitive.

A further point worthy of mention was that the words chosen were, in fact, lemmas. The words chosen came from a lemmatised corpus and so they each actually also represent same-class grammatical inflection, e.g. business > businesses, customer > customers. In the analysis carried out, only the head-words have been analysed, that is, business has been analysed but not businesses; customer, but not customers. It is acknowledged that a full lemmatised analysis would have been better, but to do that would have at least doubled the amount of data to be analysed, if not tripled it. This was not felt a reasonable task for a single researcher to carry out. The resulting analysis is a compromise, but a compromise that has still yielded a significant amount of new data concerning business lexis.

Once the words had been chosen, they were concordanced using WordSmith 3 and all concordances for each word were printed out and stored. Then, the concordance lines for each word were manually analysed one by one, semantic categories for each word were identified, and then each instance for each category was counted and further given as a percentage of the total amount of instances. It was decided that rather than simply present semantic sets that associated overall with the word in question, they would be sub-divided into ‘left of the node word’ and ‘right of the node word’ groups. An example of the left of the node word prosodies can be seen below in Table XXXXI, using the word customer. The table shows the semantic prosodies attached to customer, the frequency of each prosody, that frequency expressed as a percentage of the total number of occurrences of customer, and examples of the prosodies. The prosodies, in fact, could easily be divided into left and right prosodies, i.e. different prosodies were found on either side of the head word, though sometimes they transcended position. The analysis carried out for each of the five main groups is now discussed and further linguistic implications are noted.

TABLE XXXXI: LEFT OF NODE WORD SEMANTIC PROSODIC ANALYSIS OF THE WORD CUSTOMER

semantic prosody	frequency/613 & %	example
positive	23 – 3.75%	the dream customer a really good customer
size / number of customers	15 – 2.44%	second largest customer
type of customer (characteristics or line of business)	42 – 6.85%	focus on the business customer the professional-type customer
negative	4 – 0.65%	a really bad customer

a) People in business

As with each of the five sections designated for analysis, ten words were chosen. In this case they were customer, manager, supplier, distributor, shareholder, employee, staff, partner, boss and management. Each word was analysed in the way described above, and prosodies were gained for all of them. Significantly, it was found that whilst several of the words each had their own unique semantic prosodies, they also had prosodic categories that were common to several of them. For example, the word customer had a unique prosody concerning company-customer relations, whilst the words manager, staff, and partner all shared the same semantic prosody related to company hierarchy and status. Additionally, some of the prosodies that were common to several words were more in evidence in relation to one word rather than another – manager displayed the semantic prosody to company hierarchy and status in only 3% of its occurrences, whereas a very similar prosody accounted for 25% of the instances of the word partner. Thus even common prosodies varied in their strength from word to word. A full and detailed analysis of all the words can be found in Appendix 6 in Volume II, and so a straight repetition of this is neither appropriate nor necessary here. However, several points of interest can be presented to show how the concept of semantic prosody was found to operate, and a selective sample of words is now discussed.

Customer: Customer displayed a unique prosody – the division of collocating lexis into words concerned either with company to customer relations, or customer to company relations. These consisted of relatively large groups, 34% of the sample for company-customer relationships and just over 12% of the sample for the customer-company relationship. Below, an example of the company-customer group is shown using customer-driven – the company is attempting to look after its customers, and the lexis refers to the company:

The second group – customer-company – describes what a customer has, does or is and the lexis refers to the customer:

Manager: Manager had a very strong semantic prosody with titles which made up 70.6% of left of the node collocates:

The prosody itself is not surprising, but perhaps its dominance is. Additionally, there were three small groups found (totalling together just over 7% of the sample) one of which was positive adjectives, e.g. forthright, excellent and good manager, confirming the findings of the key words that business lexis stresses the positive over the negative.

Supplier: One of the problems of using corpus-based analysis surfaced with this word.

Its inclusion as a key word was partly due to high occurrence of the word in manuals and legal contracts, thus skewing the data to a certain extent. However, this allowed the chance to look at a genre-specific semantic prosody. This was a prosody of obligation, expressed through high use of modality[178] and came from both the legal agreements and manuals sections of the corpus:

Other semantic prosodic groups found for supplier not linked to a specific genre included size and significance, e.g. leading supplier, the world’s second largest supplier and what is supplied and to whom, e.g. supplier of business solutions, supplier to this segment. Similar findings in terms of genre-specific prosodies were also noted for the word distributor, which again was found in high occurrence in legal agreements in the corpus. Here modality was used to specify distributor rights, e.g. distributor may terminate.

Shareholder: Shareholder displayed a genre-specific prosody related to financial and annual company reports – value and returns – with examples being total shareholder return and improving shareholder value.

Employee: Analysis of this word showed the potential usefulness of semantic prosody in both linguistic analysis and, potentially, in the language classroom – less common collocates can be subsumed into a larger semantic set and thus become more readily recognised/categorised. A clear group of ‘right of the node’ collocates all concerned with employee benefits was noted, e.g. in lines 75, 76 and 78 in the example shown below:

Collocations such as employee’s holiday entitlement may not be common, yet when added together with examples such as employee benefits and employee restaurants, a clear semantic association can be seen between the employee and benefits they get from the company. It is argued here that this kind of relationship would have been impossible to establish without the use of computerised corpus linguistic techniques.

Boss: Boss displayed only very small semantic prosodies with companies and institutions on both sides of the node. However, it was interesting to note that no positive adjectives were used with boss, but three negative adjectives were noted – America’s meanest boss, old-fashioned boss and idealistic boss (used negatively). A larger sample size would be needed for safe conclusions, but it seems that in the BEC, bosses are referred to more in negative than positive terms.

Shared prosodies: In addition to these word-specific prosodies, a number of semantic sets in a common associative relationship to a number of different words was found. These associations are summarised in Table XXXXII below – where a word associates with a semantic group it is noted with an asterisk and where there are no associations to any semantic prosodic group the word is shaded. Page references for where each word can be found in Vol. II are given.

The table shows, for example, that customer, manager, staff and management all displayed a positive prosody, and that five semantic prosodic categories – shown on the horizontal axis – recurred throughout nine of the ten words. This finding is significant as it extends the definition of Business English lexis beyond the information already gained from the key words. The key words showed that business lexis is largely formed from a limited number of semantic groups. The analysis by semantic prosody not only shows the fact that these semantic groups connect and co-occur with each other, but it also shows how they connect and more specifically, which words are attracted to which semantic sets. Analysis will now continue showing that the same was found for the four other groups of words.

TABLE XXXXII: PEOPLE IN BUSINESS: TABLE OF SEMANTIC PROSODIC RELATIONS

WORD ê SP è	Page Ref.	positive	size/ number /significance	titles	company/ institution	company hierarchy /status
customer	665	*
manager	668	*		*		*
supplier	671		*
distributor	674
shareholder	677		*
employee	679				*
staff	681	*	*	*		*
partner	684					*
boss	687				*
management	690	*			*	*

b) Business activities

The words chosen for this section were: business, investment, delivery, payment, development, production, communication, competition, takeover and distribution. Six recurring semantic groups were found, covering eight out of the ten words in the group. This group then, was less homogenous than the previous one, where nine words shared at least two semantic groups, but here the diverse activities can still be seen to co-associate with recurring semantic groups quite widely. The table on the next page shows the co-associations.

Business: Business collocated with a wide variety of semantic groups, the largest of which (9.86%) was concerned with the line of business. An example is shown below:

Notable also about the word business was its positive prosody. There were only 50 positive adjectives referring to business found, e.g. successful, sound, strong making 1.96% of the sample. However, when compared to the fact that only seven negative adjectives were found (e.g. unviable, boring) it could be considered significant. Business typifies the comment made earlier regarding smaller prosodic groups – clear semantic prosody is displayed, but the groups found are relatively small.

TABLE XXXXIII: BUSINESS ACTIVITIES: TABLE OF SEMANTIC PROSODIC RELATIONS

WORD ê SP è	Page Ref.	prod/ serv	positive	size/ number /significance	company/ institution	time/ timing	money/ value	people
business	694		*	*	*		*	*
investment	707		(*)	*	*	*		*
delivery	711					*
payment	714					*
development	717		*		*			*
production	721	*	*
commun- ication	724		*
competition	727
takeover	730						*
distribution	732	*

A positive semantic prosody was also noted with the words investment and development. In the case of investment, one semantic set – results of investment – though relatively small (9% of the sample), consisted of virtually only positive results (95% of the semantic set). Similarly, with development, the category nature of development showed 30 positive instances, e.g. value-enhancing development (5.8% of the sample), to only one negative instance (slow development). The positive stress in business indicated in the key word analysis earlier, is therefore further confirmed by the proliferation of positive semantic prosodies found in the BEC.

One word did not associate with any of the semantic categories shown in the table: competition, where a vaguely negative prosody was found to a group of words denoting the toughness of the competition – unbridled, fierce and aggressive competition being examples.

Distribution had prosodies with four semantic groups shown in the diagram below:

product/service place

indoor TV aerial UK

controlled air global

Distribution

nature of distribution (how) infrastructure

automated channel

electronic system

network

Fig. 44 Semantic prosody for ‘Distribution’. All show left collocation except for ‘infrastructure’

Production had a notable prosody (7.98%) with products, i.e. what is produced, and examples are shown here:

c) Business actions

This section consisted of ten verbs: sell, manage, receive, confirm, provide, send, develop, discuss, achieve and improve. Of the five sets of lexis chosen for analysis in the thesis, this group displayed both the lowest level of semantic prosody and the most diverse prosodic groups with which the words associated themselves. The overt semantic relations between the words, however, were replaced by strong colligational links that will be discussed in the next part of the chapter. Despite the above, three recurring semantic groups were identified, though it must be stated that the links were not as strong as in the other 4 semantic groups under study. The recurring groups are shown in the table below:

TABLE XXXXIV: BUSINESS ACTIONS: TABLE OF SEMANTIC PROSODIC RELATIONS

WORD ê SP è	Page Ref.	positive	money	products
sell	734			*
manage	738
receive	741		*
confirm	743
provide	746	*		*
send	749
develop	753
discuss	756
achieve	759	*	*
improve	761

It can be seen that six words were found to have none of the recurring themes of the others. Additionally, two words (confirm and develop) were found to have no specific prosodies at all. Develop, however, did collocate with a wide range of words that had a general business theme, as the examples below show:

Provide: This word had four clear semantic groups with which it collocated: business services, products, information and the most notable aspect of the word: a very strong positive connotation (35.82% of instances):

Send: Send had very predictable collocates: on the left of the node requests and on the right documents:

Later in this section, the prosodies gained from these words in the BEC are compared to prosodies found for the same words in the BNC. It will be seen that send forms both similar and additional categories of prosody in general English.

Discuss: The only prosodic group found here was a group of expressions referring to future contact or further discussion (12% of instances of discuss):

It will also be noted that these examples come from the category of correspondence from which the majority of examples for this prosodic group comes. This once again points to the possibility of a genre-specific prosody.

Achieve: Achieve had an overwhelmingly positive semantic prosody related to business outcomes, amounting to 100% of the sample. The collocates are concerned in some way with the positive outcome of business goals and targets:

A further semantic prosody was found for achieve, referring to money, costs or prices (17% of the sample). Examples can be seen below:

d) Business descriptions

Ten adjectives were chosen for this group: those concerned with size or speed: high, big, low; those concerned with place: global, international, local; and those related to the

business world: competitive, corporate, strategic and financial. In contrast to the verbs of the previous group, the adjectives[179] chosen here, quite predictably, had a high number of semantic prosodies. Two aspects of special interest can be noted with this group. Firstly, the semantic prosodies found for a number of words – high, low, competitive, corporate – are concerned with high-level and extremes. Secondly, and more importantly, is the fact that those adjectives that can be considered the most general in meaning, or more specifically, those adjectives that are very open-ended collocationally, such as big, high and low, all actually demonstrated clear semantic prosodies related to business. This is, of course, not surprising in itself. When discussing business matters one’s use of the word big, for example, could be expected to relate back to the business world. However, it is important to note that in the business environment words that are normally very open in their collocative potential become at least partially fixed. This must have consequences for both teaching and teaching materials and these aspects will be discussed in more detail later in the chapter. In the table below, the recurring semantic prosodic groups are shown for the ten words. As can be seen, eight recurring semantic prosodies were identified and, in addition, several semantic prosodies unique to individual words were noted. These are now briefly discussed.

TABLE XXXXV: BUSINESS DESCRIPTIONS: TABLE OF SEMANTIC PROSODIC RELATIONS

WORD ê SP è	Page Ref.	bus activ	extremes	comp/ inst.	economic/ financial indicators	money	products/ services	size/ signif- icance	people
high	764		*			*
big	769			*		*			*
low	773		*		*
global	777	*		*	*		*		*
international	780	*		*				*	*
local	784			*			*		*
competitive	787		*			*
corporate	790	*	*			*			*
strategic	793	*
financial	795			*					*

High: This had a neutral prosody in that the amount of positive and negative collocations were more or less equal. High was one of the words that associated with semantic prosody of extremes[180] and examples using the word extremely itself can be seen below:

The largest prosodic group of high was a group of words/phrases relating in some way to money (11.36% of the sample):

Big: Three semantic prosodies were found for big: companies & institutions, money and people. Of these, by far the largest was that of companies & institutions (23.85% of the sample). These prosodies combined, amounted to 31% of the sample and it will be seen later in this section that big used in general English (i.e. in the BNC) was much more open in its collocates, with semantic prosody only accounting for 12% of instances.

The words global and international both displayed a very strong tendency to be used as the names of companies, with international being more popular than global. The words also shared three semantic prosodies in common shown here:

TABLE XXXXVI: COMPARATIVE PROSODIES OF ‘GLOBAL’ AND ‘INTERNATIONAL’

WORD ê SP è	business activities	company/ institution	people
global	17.59%	16.66%	5.55%
international	9.39%	34.76%	4.94%

Despite the similarities, there were clear differences. Global collocated more with business activities than international, and international more with companies & institutions than global. Both these words were in contrast to local. The words collocating with local included a large number of non-business-related words and this can be seen most clearly in the companies/institutions category. All three words, global, international and local shared this semantic prosody with companies/institutions, but the institutions collocating with local were noticeably more often of a distinctly non-business nature:

Of the four business-related adjectives – competitive, corporate, strategic and financial – the findings were both expected and unexpected. The word corporate was found to be in association with five semantic groups, the largest of which was money/finance, e.g. corporate bonds, corporate earnings and corporate insolvency. Corporate also associated with extremes (8% of the sample), e.g. almost $8 billion worth of corporate debt.

Strategic had three prosodies, perhaps the most interesting of which was a group of words relating to aiming for the future (18% of sample). Examples are shown here:

These findings could well be expected. An unexpected finding was the slight negative prosody of the word financial. Examples could be found in both pre-nodal and post-nodal position. Together they made only approximately 9% of the sample, but there was no real corresponding positive prosody to balance it. Below are some examples of the negative collocates of financial – crisis and trouble:

The largest prosody for financial was macro-level demarcation, which constituted on its own just under 10% of the sample. An example is shown here with the collocate sector:

e) Business events and entities

The lexis referring to events and entities chosen for this section was sale, merger, trade, package, export, service, market, earnings, performance and product. As with the other groups of words, they displayed both common and unique prosodies. In this case, eight semantic categories were found to collocate with the words:

TABLE XXXXVII: BUSINESS EVENTS AND ENTITIES: TABLE OF SEMANTIC PROSODIC RELATIONS

WORD ê SP è	Page Ref.	time	comps/ inst.	Posi- tive	macro- level demarcation	money/ finance	products/ services	size/ signif- icance	people
sale	802		*			*	*
merger	805	*	*			*		*
trade	808		*		*				*
package	812			*		*		*
export	815								*
service	818			*					*
market	823			*	*	*	*
earnings	832	*
performance	834			*		*			*
product	837				*		*

It can be seen that the most frequent semantic prosodies are concerned with people, money & finance and positive attributes. In addition to these broader categories, several unique prosodies of interest were found.

Sale: A unique prosody was found for this word, that of availability – attached to 26.82% of all examples of sale in the corpus. Examples are shown below:

The link between collocation and colligation can also seen to be at work here with availability often denoted by phrasal verbs, e.g. put up for sale, put on sale.

The distinct positivity of business lexis has been pointed to throughout this chapter. Examples in this section of positive semantic prosodies included service, which was found to have slight positive prosody with right of the node collocates such as attentive, superior and innovative. Similarly, market displayed 53 overtly positive collocates, e.g. good, strong and bull (3.79% of the sample) but only 19 negative, e.g. bear, cut-throat (1.36%). Market also had prosodies related to products and services (17.47%), and a company’s position in relation to other companies or the market (10.45%). Examples of this latter unique prosody are shown below. Whilst these collocations (shown below) are common, the semantic umbrella under which they fit would not have been possible to detect without the use of the concept of semantic prosody and the process of computerised concordancing.

Performance also had a slight positive prosody, e.g. world-beating, perfect (8% of the sample) and only two examples of negativity: dire and poor (0.4% of the sample).

This is not to suggest, however, that no negative prosodies were found. The word trade, for example, was found to have a slight negative prosody (admittedly only just under 5% of the sample), collocating with words such as deficit, dispute and sanctions. The word export had a semantic prosody with limitations and control (6% of the sample) that could be considered partially negative. Example collocates included export controls, export restrictions and export regulations.

The final word in the group, product, surprisingly showed very limited prosodic potential. It had half the number of prosodies of market, for example, and those prosodic groups it did display were relatively small. Product, however, displayed a grammatical and semantic sense in Business English not found in COBUILD and only marginally represented in teaching materials contained in the PMC. This will be dealt with in detail in the next section which considers the question of semantico-grammatical patterning in Business English.

9.3.3.2 Semantic prosody: conclusions

Can semantic prosody be found in Business English? It was noted at the beginning of the section on semantic prosody that it is a largely untried concept in terms of lexical analysis. The question was asked as to whether semantic prosody could be found in Business English or not. In answer, the following points can be made.

Of the 50 words examined here, 48 displayed from one to eight semantic prosodies of varying sizes.
It was found that business lexis displays a tendency to collocate with recurring semantic prosodic sets but also retains the potential of collocating with prosodic sets unique to individual words.
The prosodic categories found, however, should be seen in terms of tendency rather than absoluteness. Words had a tendency to associate with a certain semantic set or sets, but the tendency often represented only a small percentage of the overall uses of the word. There were only a limited number of words where semantic prosodic coverage was high, e.g. achieve. In the diagram below, the most frequent semantic prosodic sets occurring with the 50 words under analysis are shown, along with the number of words that associated with them in semantic prosodic relations:

positive 14 company/institutions 15 money 13 size

hierarchy 4 Business Lexis extremes 4

products/services 10 people 15 business activities 5 time/timing 8

Fig. 45 The most frequent semantic sets collocating with business lexis and the number of words found to associate with them in semantic prosodic relations /50

The findings of this analysis are significant as they closely match the findings of the key word analysis. The key word analysis showed that the business world revolves around recurring semantic sets: people, institutions, places and money, for example. The analysis in terms of semantic prosody found that business words are not only members of these, largely similar, semantic sets, but also these words, as members of these semantic sets, interacted and regularly collocated with the other sets. There is, therefore, a loosely integrated and interlocking lexical world orchestrated by meaning.

The diagram above shows the most common recurring prosodies that business lexis is attached to, but it is important to repeat the point that this cannot be seen in absolute terms. Business lexis in general displays a distinct tendency to collocate with these semantic groups, but it would be a mistake to think that business lexis is bounded by them. Individual words have their own unique prosodies and, moreover, a large amount of language does not fall into any significant semantic boundary and indicates the openness of business language to the world outside. For example, taking the first five positive key verbs of the BEC, semantic prosody accounted for a varying amount of the total collocational occurrences of words:

TABLE XXXXVIII: PERCENTAGE OF COVERAGE BY SEMANTIC PROSODY

Word	% of total instances covered by semantic prosody
sell	62%
manage	69%
require	92%
confirm	0%
provide	64%

Thus semantic prosodic groups, added together, totalled 92% of all occurrences of require, but there were none found with confirm. This shows a range of semantic prosodic potential in the words and the fact that semantic prosodies often form only a limited part of the collocating lexis of a given word.

Business-specific semantic prosody? This section on semantic prosody has concentrated on the types and relations of these prosodies found in Business English. The fact that the phenomenon exists can be in no doubt. However, it is a much greater step to say that Business English has semantic prosodies unique to itself. Tribble (1998:11) distinguished between what he called global and local semantic prosody. This means that a word may have a general semantic prosody in ‘everyday’ language, but a particular semantic prosody in a given language environment. Tribble gave the example of experience, noted earlier, which was found to be imbued with a prosody typical to his corpus of EU Phare project proposals, where experience took on the sense of professional capital.

In order to see if the same could be found in the BEC, five words already analysed in the BEC were chosen: send, manage, big, global and package. The same words were then analysed in the BNC and the results compared to the same words in the BEC (for the full analysis see Appendix 9 in Vol II). The words were chosen for their generality – send, big and global – and for possible business connections – manage and package. They also come from separate word-class categories used in the analysis, verb, adjective and noun/verb. None of the words is a hard-core business word with the exception, perhaps, of manage, but that too has several senses. This choice of words was made in order to see if quite general words would behave differently in terms of their semantic prosodies in the Business English environment when compared to general English. The results showed small but clear differences in terms of the semantic prosodies found, and in the amount of coverage the semantic prosodic groups afford each word in the two different environments. These are seen summarised in the table below:

TABLE XXXXIX: COMPARATIVE OCCURRENCE OF SEMANTIC PROSODY BEC/BNC

word	SP in BEC	SP% BEC	SP % BNC	SP in BNC
send	sending documents sending requests	46.84%	34.97%	sending documents sending people
manage	what is managed (largely business -related)	69.51%	0%	–
big	companies money people	31.68%	12.87%	people companies/organisations/ institutions
global	business characteristics/qualities business activities products companies people economic/financial indicators	56.45%	26.46%	climate people
package	size/value finance positive computers	36.4%	25.26%	size/value computers

Three main points can be made:

1. The first point returns to one made earlier in this section – in the Business English environment, collocates become more fixed. This is true of both words that could be considered ‘business-related’ e.g. manage, and also of general words such as big. In each word above, the percentage of coverage by semantic prosodic sets is greater in the BEC than in the BNC, indicating a wider collocative potential in the BNC.

2. Some words had the same semantic prosodies in both the BEC and the BNC, for example, send – documents; big – people; global – people and package – computers. This fact is partially misleading as there is a difference in size and content of the prosodic groups. In the case of send, for example, in the BNC the kind of documents sent that are referred to are different from those in the BEC. In the BEC they are business-related, whereas in the BNC the documents are much more varied and general, e.g. postcards, sae, and non-business letters. In the BEC big collocated with three clear (though in two cases small) semantic prosodic groups, companies/institutions, money and people. Of these, companies/institutions was the biggest (23.85% of BEC sample). In the BNC the largest group for big is that of people with only 38 instances of companies/ organisations/institutions (4.61% of BNC sample).

3. In addition to these differences in content, there were also unique business-related semantic prosodies. This can be seen, for example, with global and package.

Global: The sample size of global in the BNC (34 instances compared to 324 in the BEC) made any in-depth analysis very difficult, but whereas global was found to be very rich in terms of semantic prosody in the BEC, in the BNC it was limited to only two semantic sets. A larger sample size would be needed for more accuracy, but it might be found, with further investigation, that there are strong business-related prosodies for this word, e.g. products, economic indicators, whilst the more general prosodies are concerned with climate – global warming, global climate changes – and people – global consumer, global viewer.

Package: In the BEC package collocated with three groups, the largest of which being financially-related lexis (18.51% of the sample in the BEC). In the BNC, this group was much less evident (4 instances 5.33%). Additionally, the positive group found in the BEC e.g. competitive package, excellent package, effective package, was not evident in the BNC.

In answer to the question regarding business-specific semantic prosody the following can be said:

Analysis by semantic prosody has shown that words become more collocationally fixed in the Business English environment.
There is enough evidence to show that words can associate with business-specific semantic prosodies when in the Business English environment, but that they also associate with semantic prosodic groups that are the same in both Business English and general English.
When the words associate with semantic groups that are the same in Business English and general English, the contents of the groups can differ, as in the case of send above, to provide a business slant to the group even if the heading is the same. Thus the concepts of global and local semantic prosody are indicated here. However, a much wider analysis would need to be carried out in order to make definite statements as to what the differences are and where they occur.

The next section will now answer Hoey’s (1997) questions 3 and 4. These asked of a word being examined by concordancing: what structure does it appear in? – a reference to colligation – and is there is any correlation between the word’s uses and meanings and the structures in which it participates?

9.3.4 What colligational and grammar/meaning patterns can be found in Business English?

In Chapter 4, the notion of colligation – how words typically behave grammatically – and further, how grammar and meaning interrelate, was discussed (Sinclair 1991, Hunston et al. 1997, Hunston & Francis 1998, Hoey 1997, 2000 and Hargreaves 2000). In line with these previous works, this thesis has attempted to discover the typical grammatical patterns of a selection of business-related lexis. In addition, the meanings found in this lexis were related to the grammatical forms they took. The primary reasons for both these analyses were at the same time pragmatic and pedagogical. The pragmatic aspect was driven by an desire for an applied element to the work – the results of the analysis should be directed to the language classroom via materials creation. The pedagogical aspect concerned the question of the accuracy of these materials. Firstly, Business English students need to be exposed to probable language rather than possible language, and presenting language in its most typical grammatical setting goes a long way towards achieving this. Secondly, students need to understand that the varying grammatical forms and patternings a word takes influences meaning, and they need to know what these changes and meanings are.

To this end, the same fifty words that were analysed for semantic prosody were also analysed for both colligation and meaning/grammatical form relationships (please turn to Appendix 6 in Vol. II to see the full analysis). This analysis was essentially descriptive in nature, but allowed several conclusions to be drawn about the behaviour of lexis in Business English – its special patternings, its restrictedness of meaning and the special, business-related meanings words display. Before discussion of the analysis, the concepts of colligation and grammar/meaning relations are both briefly re-visited in turn.

Colligation: Colligation was discussed in a recent article mentioned briefly earlier, by Hoey (2000), where he took a group of words relating to professions – actor, actress, architect and carpenter – from a large corpus and studied the colligational patterning they displayed. He found that despite their surface similarity, i.e. as ‘profession’ words, there were significant differences between the way the words most commonly occurred, e.g. in terms of use of articles, classifiers, apposition and parenthesis (Hoey 2000:234-235). Hoey thus stressed the grammatical context in which the words are most typically found. Hoey then examined how typical colligational patterns were represented in teaching materials and found a disparity between real-life patterning and patterning found in, for example, English Vocabulary in Use (McCarthy & O’Dell 1994) – which he regarded as the best book in its field at the moment.[181] His analysis showed the dangers of using intuition in creating language examples for materials and he concluded that ‘intuition, even the intuition of the best lexical applied linguists, is likely to be flawed’ (Hoey 2000:237). Thus it is argued here that – in line with Hoey – analysis of colligation should best come from observable corpus data.

The following sections present findings from the BEC with regard to colligational patterning. For each of the five semantic groups under analysis – those that were analysed for semantic prosody – colligational patterning is discussed first, followed by analysis of grammar/meaning with reference to the COBUILD (1995) dictionary. In order to assist the recognition of colligational patterning, the corpus was part-of-speech (POS) tagged, using Autasys (Fang 1998). However, it was not felt reliable enough to rely on entirely for this analysis[182] so the concordance lines of each word were printed and analysed manually, and typical patterns were identified and noted. A brief example is shown here for the word competitive showing a typical grammatical patterning in which it was found to occur:

i) -ly adverb + competitive + noun (group)

12 instances 6.15%

extremely competitive market place, highly competitive US market

Grammar/meaning relations: The relationship of meaning and grammatical form was a more difficult prospect, and previous studies had to be relied upon for the analytical process. This study utilised the COBUILD (1995) dictionary, which was chosen for several reasons.

First and foremost, the COBUILD dictionary is based on corpus evidence of general English, and so represents English from a corpus-based perspective. Secondly, via the work discussed by Hunston et al. (1997) and Hunston & Francis (1998), the dictionary operates on the principle of grammar/meaning assignation, whereby each entry shows the relationship of meaning to grammatical patterning. In this way COBUILD shows how the typical grammatical patternings in which words appear, also determine meaning. Additionally, COBUILD gives the most common meanings of a word in order of frequency,[183] and thus it is simple to see how the meanings gained from Business English match or differ in frequency from those of general English. This approach further facilitated the possibility of looking at sub-technical language (Trimble 1985), and a means of seeing how meaning varied from general to Business English.

For this thesis, each instance of the 50 words was assigned to a COBUILD meaning/sense[184] category. The instances allotted to each meaning of a word were then totalled, and a percentage value, in relation to all instances of a given word, was noted. Examples of each meaning/sense were also given (shown in italics). To clarify this process, an example is given below showing the analysis of the word competitive:

COMPETITIVE

COBUILD Sense 1 (situation or activities where people or firms compete with each other)

183 instances 93.84% of sample

Patterns: Adjective

In today’s competitive business climate

we need to be competitive in the next century

COBUILD Sense 2 (a competitive person)

1 instance 0.51% of sample

Patterns: Graded adjective

A formidable woman and she was and is very competitive

COBUILD Sense 3 (goods and services at a competitive price)

11 instances 5.64% of sample

Patterns: Graded adjective

France’s nuclear power plants generate electricity at competitive prices

Are they reliable and do they have competitive prices?

It can be seen that the word competitive conforms to general English in terms of its most frequent sense (Sense 1), but that Senses 3 and 2 are found in reverse order. The words under analysis will now be discussed and conclusions drawn from them.

9.3.4.1 Colligation and grammatical form/meaning relations in business lexis

a) People in business

i) Colligation

The words in this first category of people in business differed noticeably in their relationships to noun/verb phrases, compound adjectives, modality and prepositional usage.

Phrasal aspects: Some words commonly formed part of pre- and post-modified verb and noun phrases and compound adjectives: customer, shareholder, manager, staff, partner, and management – whilst the other words in the group – supplier, distributor, employee and boss did not display this tendency to the same extent.[185] Examples are shown below:

TABLE L: EXAMPLES OF NOUN/VERB PHRASES AND COMPOUND ADJECTIVES RELATED TO ‘PEOPLE IN BUSINESS’

word	Three-word	pre-modified	post-modified
customer	–	–	compound nouns: customer-base, customer-care, customer confidence, customer-consciousness, customer needs, customer orders, customer requirements, customer satisfaction, customer service, customer support compound adjectives: customer-led, customer-driven, customer-focused, customer-friendly
shareholder	noun phrases: total shareholder return, significant shareholder value verb phrases: increase shareholder value, create shareholder value, improving shareholder value	leading shareholder minority shareholder	shareholder return shareholder value
staff	–	agency staff, service staff	staff discount, staff forum, staff levels, staff performance, staff recruiting, staff training
management	senior management team total quality management strategic management system	asset management credit management group management project management	management accounts management consulting/consultancy management contractor management team

The words management and staff are ‘outsiders’ in this group as they do not refer to a person per se, but rather to a group of people and, indeed, a function or activity of a group of people. They were chosen for this group, however, because they were both key words and relatively frequent. It is seen above that management, for example, showed a tendency to be part of 3-word noun phrases. It also was found 29 times in a 3-word noun phrase co-joined by ‘and’:

Modality: The words supplier, distributor and employee were notable for the high degree of modality used with them – for example, modality was evident in 30% of all instances of supplier. This was mentioned in the previous section on semantic prosody, and is a result of the high frequency of the words in agreements and contracts. Thus, general conclusions about the relationship of these words to modality should not be drawn, but rather a genre-specific colligation is indicated, i.e. the obligations of the participants in the agreements are designated by modality. An example is shown below, displaying usage of shall:

Prepositional usage: Three items, manager, staff and partner, stood out for their linkage to a range of prepositions:

Manager:

Prepositional usage: (determining the relationship of the manager to the company or the job)

44 instances 10.6% of sample.

at, for, in, of, with, along (with)

Group Manager at Aerosystems, must designate a manager for the project

Staff:

Staff + of + number:

5 instances 1.28% of sample

a staff of just 150 inspectors

prepositional use: (says where the staff works)

18 instances 4.63% of sample

staff at the Samsung Group, staff from other countries, staff in the hotel, staff of the UK ‘s top organizations

noun + of + staff:

27 instances 6.95% of sample

member of staff, number of staff, calibre of staff, development of staff, the drift of staff away

Partner:

Prepositional usage: (shows where the person is a partner, with whom and for what purpose)

17 instances 14.78% of sample

a senior partner at the McKinsey firm

a marketing partner for the drug

a partner in PW

managing partner of Andersen consulting

ii) Grammatical form/meaning

All the words in this group save one, partner, followed the frequency distribution of COBUILD, that is, the meanings of a word given in order of frequency by COBUILD followed the same frequency distribution here. However, a reduced number of meanings was noted, pointing to the specificity of language used in ‘special’ ESP type situations (West 1997, St John & Dudley Evans 1998), with a reduced variety of meanings in operation. The table below summarises the words in this group in terms of a) how limited

their use was in the BEC as compared to definitions given in COBUILD;[186] b) if the order of the most common meanings given in COBUILD was different in the BEC, and c) if there were any new meanings specific to the BEC identified:

TABLE LI: BUSINESS-SPECIFIC USAGE OF WORDS RELATED TO ‘PEOPLE IN BUSINESS’

Word	Limited use	Change of Order	New Meaning
customer	Yes: – 1/2
manager	Yes: – 2/3
supplier
distributor	Yes: – 1/2
shareholder
employee
staff	Yes: 3/4
partner	Yes: 3/5	Yes: 3-4-1	Yes: partner in business
boss	Yes: 2/4
management	Yes: 2/3

It can be seen that seven out of ten words displayed reduced meaning, that is, of the meanings given for each word in COBUILD, there were less meanings found in the BEC. In terms of change of order, one word did not follow the COBUILD frequency patterning – partner – where its most usual meaning in general English of life or sexual partner, Sense 1 (here only 0.86% of the sample), was superseded by COBUILD Sense 3 – the people who share ownership of a business (56.52% of the sample). Thus, a business-specific aspect of the word is stressed. Partner was also of interest, in that a sense not mentioned in COBUILD was detected. In the BEC sample, partner was found to be used as a verb, colligating with the preposition with, meaning a new sense can be presented:

ADDITIONAL sense: similar to COBUILD Sense 5 (to partner someone in a dance or game).

There are 6 instances of ‘partner’ (5.21% of sample) used as a verb in the sense of ‘partner with another company for potential mutual benefit’.

The pattern found is:

verb + with + noun: partner with us, we’re willing to partner with the administration

Examples of variation in meaning according to grammatical form was in evidence. For example, the word management, when used as an uncount noun, referred to the control and organising of a business (e.g. this style of management). When used as a variable collective noun it referred to the actual people (e.g. management shake-up, management buy-out).

A further grammar/meaning patterning of note, though small, was that found with the word supplier. This word was found in the following pattern:

genitive/article + positive adjective + supplier + preposition + (product/company)

This pattern occurred 24 times (13.33% of the sample of supplier) and indicated the size of the supplier in predominantly positive terms, for example, UK’s leading supplier of PC accountancy software, a leading supplier of naval radars and the biggest supplier of Twinings.

b) Business activities

i) Colligation

The words in this section, displaying a high level of nominalisation, were business, investment, delivery, payment, development, production, communication, competition, takeover and distribution. It was found, as with the words in the previous section, that typical grammatical patterns could be found for all of them, and that these patterns, whilst shared by several words, also varied between them.

Investment displayed a distinct tendency to be formed into 3-word noun phrases, e.g. realized investment gains, return on investment and chief investment officer. Development was also often found in 3-word noun phrases co-joined by and, for example, research and development, training and development. Likewise delivery:

Delivery, development, competition and distribution all showed strong colligational links to prepositions. Obviously, prepositional usage also affects meaning and this can be seen clearly in the example pattern delivery + of (17.55% of the delivery sample), where delivery is followed by of , which is followed by the ‘what is to be delivered’:

Production stood out in this group in its tendency to be in colligation with plural nouns (15.2% of the sample):

Sentence positioning: Hoey’s question 5 concerned the positioning of a word in a given text. As was noted earlier, this analysis has not been carried out in this study, but sentence positioning, rather than textual positioning, has been noted to some extent. Thus, the position a word typically appears in a sentence can be seen as part of its overall grammatical patterning. It was found that the word business showed a marked tendency to appear in sentence-end position (10.46% of the sample), whilst the other words in this group did not display this same tendency.

ii) Grammar/meaning

Similar to the previous category of people in business, it was found in this section that the full range of meanings given by COBUILD were not present. For example, with the word production, COBUILD Senses 3 & 4 were missing completely. With competition, COBUILD Sense 1 (a general competitive state) was completely missing, making COBUILD Sense 3 (firms competing for business – the state of competition) the most common meaning of the word in the BEC (91.33% of the sample). COBUILD Sense 5, which refers to a competition, i.e. something one can enter and win, was only found to occur 4 times (1.57% of the sample).

A summary of differences can be seen in the table below, where six words showed limitation in meanings compared to COBUILD definitions:

TABLE LII: BUSINESS-SPECIFIC USAGE OF WORDS RELATED TO ‘BUSINESS ACTIVITIES’

Word	Limited use	Change of Order	New Meaning
business	Yes: 15/22	Yes: 1-3-5-2-12	Yes: conversational filler
investment	Yes: 3/4
delivery	Yes: 2/4
payment
development		Yes: 3-1-2-4-5
production	Yes: 3/6
communication
competition	Yes: 4/5	Yes: 3-2-4-5 (no number 1)
takeover	Yes: 1/2
distribution

There were clear examples of the interrelationship between grammatical patterning and meaning. For example, Sense 2 of competition (the people you are competing with – the competition) has a pattern of singular noun + the, e.g. more differentiation from the competition is needed, whereas Sense 3 (the state of business competition), e.g. competition for market position, is found in an uncount noun form.

Communication displayed a grammatical pattern related to positive attributes. The pattern – positive adjective + communication + noun (group) – occurred 11 times (8.66% of the sample of communication), e.g. excellent communication skills.

A further point of note concerning this category was the vast number of senses of the word business. COBUILD noted 22 major senses of the word, of which 15 were found to be present in the BEC. An additional sense was also found, though of very low occurrence:

ADDITIONAL sense:

Conversational filler:

2 instances 0.07% of sample

all this sort of business

that sort of business, you know. The usual story….

Links were found between definite article usage and COBUILD Senses 3 & 5 – both senses of business were commonly preceded by a definite article, for example:

Sense 3 – (an actual business/firm): The minimum target for the business to survive is £250 million

Sense 5 – (line of business): How did you come into the business then originally, the car business?

There was also a link between indefinite article usage and COBUILD Sense 3, where the indefinite article can precede business, for example: The skills that are needed to start a business.

These points were found to be true except when business was used as part of a noun group, e.g. the business community, the business mix for 1993 was 65.5% personal, a business magazine.

c) Business actions

i) Colligation

Analysis was carried out on the following verbs: sell, manage, receive, confirm, provide, send, develop, discuss, achieve and improve. Of these verbs, sell, provide and send showed distinct colligation with a variety of modal forms. Sell colligated with modals in 17.3% of the sample, provide 24.87% and send 26.3% of the sample. Thus a key aspect of these verbs is their relationship to likelihood, necessity and ability. Examples of modal + send are shown as an example below:

Notable was the low frequency of phrasal verbs, though provide for made up 3.48% of instances of provide, and send had 13 instances of send out (3.56% of the sample):

Two verbs – develop and sell – showed typical colligation with and and another verb, and collocated with predominantly dynamic verbs.

Develop + and + verb: Develop was commonly followed by a secondary verb co-joined by and (11.28% of the develop sample). The second verbs were notable for their dynamic meaning:

Verb + (noun) + and + sell: In the case of the verb sell, the co-joining verb preceded it, but was again linked by and (9.61% of the sell sample):

Adverb + improve: The verb improve stood out in this sample as a verb that was pre-modified by -ly adverbs. Whilst the sample was small – only 6.06% of the instances of improve – similar patterns were not noted with the other verbs.

ii) Grammar/meaning

Of the verbs in this group, six verbs followed the frequency/meaning categories assigned to them by COBUILD and four differed: sell, manage, confirm and develop.

Confirm stood out as stressing a particular business-related sense. The most typical meaning assigned by COBUILD was confirm in the sense of confirming what one believes, suspects or fears. In the BEC, this meaning only constituted 3.29% of the sample. COBUILD Sense 3 (to confirm an arrangement or appointment) was, instead, the largest group, forming 72.52% of the sample. Both senses had similar grammatical patterning. Sense 1 had a typical patterning of verb-noun, e.g. confirm prejudices. Sense 3 had the same pattern but could also appear simply as a verb: I can confirm that I will visit Cork. Again, key changes are noted in the table below:

TABLE LIII: BUSINESS-SPECIFIC USAGE OF WORDS RELATED TO ‘BUSINESS ACTIONS’

Word	Limited use	Change of Order	New Meaning
sell	Yes: 6/11	Yes: 1-2-5-4-6
manage	Yes: 4/7	Yes: 1-4-3-2
receive	Yes: 1/7
confirm	Yes: 3/7	Yes: 3-2-1
provide	Yes: 1/2
send	Yes: 4/7
develop	Yes: 4/12	Yes: 1-7-10-4
discuss
achieve
improve	Yes: 3/4

d) Business descriptions

i) Colligation

The words studied in this group were: high, big, low, global, international, local, competitive, corporate, strategic and financial. Colligationally, this was a very interesting group of words, as the words displayed lexical richness through their very strong tendency to bond with others to form compound nouns or longer noun phrases, with the phrases being post- rather than pre-modified. Examples are shown below:

Three-word noun phraseswere also common, the examples shown here are with global in central position:

High and low: It was noted in the section on semantic prosody earlier that both high and low, whilst being used predominantly in adjectival form, also performed a nominal function. In the BEC it was found that there was a difference in the frequency of occurrence in the noun forms of these two words. High and low have similar uses and patterns in Business English, simply representing different ends of the scale, but what has been found here is that low used the noun form approximately twice as often as high. There were 26 instances of a high (3.88% of the high sample), whilst a low had 21 instances (7.72% of the low sample). The number of occurrences is approximately the same, but low is much less used than high, so a low gains a higher (double) percentage of the sample. Some examples showing how low is used are given below indicating a typical patterning of indefinite article + low + of + numeral/%:

Financial: This word stood out in the section for the variety of characteristics it displayed. A negative semantic prosody was noted in the previous section, with the negative word coming in post-adjectival position (59 instances – 7.56% of the sample) for example, financial meltdown and financial difficulties. Financial was also observed to fall in the middle of noun groups connected by and:

Financial displayed a similarity to the word business in that it was found to occur in end of sentence position (14.23% of the sample) and, additionally, at end of clause position (also 14.23% of the sample):

ii) Grammar/meaning

Four words from this group displayed differences in common meaning when compared to COBUILD meaning frequency as a result of the Business English environment. It was found that high, big, low and competitive all shifted in terms of central meaning in the BEC. This also correlated with the semantic prosodies noted earlier, where big, for example, had a strong semantic prosody with companies and institutions.

TABLE LIV: BUSINESS-SPECIFIC USAGE OF WORDS RELATED TO ‘BUSINESS DESCRIPTIONS’

Word	Limited use	Change of Order	New Meaning
high	Yes: 8/23	Yes: 4-12-9-6-7 (no number 1)
big	Yes: 6/12	Yes: 3-4-1-2
low	Yes: 5/22	Yes: 4-6-2-8-18 (no number 1)
global	Yes: 1/2
international	Yes: 1/3
local	Yes: 2/4
competitive		Yes: 1-3-2
corporate
strategic	Yes: 2/3
financial

High: COBUILD Sense 1 (something that extends a long way from bottom to top) was completely absent from the sample. In contrast, Sense 4 (high in amount or degree) constituted 62.92% of the sample. Both senses, however, share the same grammatical patterning of graded adjective.

Low: Likewise, Sense 1 of low (something a short distance from bottom to top) was also missing from the sample, with Sense 4 (low in amount or degree) making up 89.7% of the sample – again, both senses shared the same grammatical patterning.

Big: In this case, Sense 3 (increase in degree, extent or importance) made up the largest group (52.47% of the sample) whilst Sense 1 (a big person or thing) came to only 8.51% of the sample.

Competitive: There were slightly more instances (11) of competitive Sense 3 – goods and services at a competitive price – than Sense 2 (1 instance) – a competitive person. The most common sense of competitive was the same as in COBUILD (situation or activities where people or firms compete).

It was noted in the section on semantic prosody that even common, delexicalised words, such as high, low and big, have limits placed on their ability to collocate in the Business English environment. This analysis further shows that in addition to an increase in collocational fixedness, there is often also a limitation in terms of the base meaning placed on these words.

e) Business events and entities

i) Colligation

The words in this group were sale, merger, trade, package, export, service, market, earnings, performance and product. As with all the other groups, this group showed distinct differences in the grammatical company the words keep. The main differences are highlighted briefly below.

Prepositional usage: Two words, merger and performance, displayed a strong tendency to appear in colligation with prepositions, and the words naturally altered meaning, depending on the prepositions used. Taking the example of merger, grammatical collocation with three main prepositions was observed:

Merger + between and merger + of: (these patterns usually name both parties of the merger or at least imply the presence of the two parties). There were 33 instances of these patterns, making 19.64% of the sample (11 between, 22 of). Examples were:

merger between UniChem and Alliance

merger between Waterstone’s and Dillon’s

a merger of the two…

Merger + with:(both parties named) which accounted for 8.92% of the sample:

Performance not only displayed the same tendency, with prepositional colligation accounting for 12.8% of the sample, but it also showed a slight tendency to be formed into a noun phrase with performance preceded by a noun and linked by and (3.8% of the sample):

Two words, export and market, showed strong tendencies to be formed into pre- and post-modified noun groups/compound adjectives. This was most noticeable with the word market. Market was a very lexically rich word in the manner in which it attached itself to business lexis to form a variety of meanings. For example, it frequently formed into compound adjectives that were both pre-modified:

and post-modified:

Market also combined to form the compound noun market leader in the typical pattern: article + compound noun + in, e.g. the market leader in digital phones. These combinations of words are potentially very useful for students to know for forming accurate chunks of business language, and this area will be discussed further in the third part of this chapter.

ii) Grammar/meaning

The meanings of some words noted in this section conformed to those found in general English, whilst others differed, sometimes quite remarkably. Merger, export, earnings and product all displayed the same frequency/meaning attributes as presented by COBUILD, whilst the other words all deviated from the norms presented in COBUILD. Further, four words – sale, package, export and product – displayed meanings either specific to the business environment, or not noted in COBUILD. An overview can be seen in the table below:

TABLE LV: BUSINESS-SPECIFIC USAGE OF WORDS RELATED TO ‘BUSINESS EVENTS AND ENTITIES’

Word	Limited use	Change of Order	New Meaning
sale	Yes: 5/10	Yes: 1-7-10-8-4	Yes: point of sale
merger
trade	Yes: 4/7	Yes: 1-3-2-4
package	Yes: 3/5	Yes: 3-1-4	Yes: collection of services bought by a company
export	Yes: 3/4		Yes: Export file ‘Export’ the dpt.
service	Yes: 11/27	Yes: 3-5-4-8-1
market	Yes: 9/12	Yes: 3-2-7-5-11 (no number 1)
earnings
performance	Yes: 2/7	(no number 1)
product			Yes: product as uncount noun

Some examples are now given of words that changed in meaning frequency when in the Business English environment.

Package: The most frequent meaning in the BEC for package was Sense 3 (a set of proposals to be accepted or rejected) making 56.79% of the sample. This contrasted with the most common meaning in general English – a package or parcel – which was only 13.58% of the BEC sample.

Market: The most usual meaning of market, as in the market where you go to buy goods, was absent from the BEC and the most frequent meanings found were Sense 2 (the market for a commodity/product – 33.23% of the sample) and, even larger, Sense 3 (total amount of product sold each year) at 37.96% of the sample. These two different meanings (Senses 2 & 3) were also colligationally marked. Where market meant the market for a product, e.g. the market for boxed brands, the access equipment market, the most typical pattern was noun + for or noun + in. Here a specific market was referred to. When Sense 3 was adopted – where the amount of product sold was referred to – the most typical pattern was the + market. Examples of this latter group are shown below:

Service: Sense 1 of service (a service provided by the government in an organised way) was overshadowed by Sense 3 (a company providing a service), which totalled 79.17% of the sample.

Performance:With performance, the most usual sense (a performance, e.g. at a theatre) was displaced by Sense 2 (how successfully something is done), which made up 99.0% of the sample.

Additional senses: In addition to the changes in frequency ranking brought about by the meanings noted above, some words also displayed meanings not noted in COBUILD at all. Sale had a small number of examples (3) of point of sale – referring to a place in an outlet where goods were on sale:

Package also displayed a meaning not noted by COBUILD, but this time of much greater frequency:

Package

FURTHER definition: (a collection of services a product/item can collectively offer its users, e.g. a desk top publishing package)

47 instances 29.01% of sample

Patterns: Count noun

The Demo CD package, 3 dimension modelling and design package, an interesting and comprehensive tuition package

This could not be tied to the BEC per se, as a similar meaning was found in the BNC Sampler corpus, but it is definitely business/product-related.

Export: Export displayed two meanings not found in COBUILD. One of the meanings was directly linked to the business world – Export as in the Export Department, e.g. Export receive signed PODs from Dublin. The second new meaning, an export file, e.g. Auto-CAD export file, is related to computers and so only indirectly to business.

Product: Perhaps the clearest example of a business-specific usage of a word that was also reflected in its grammatical form was found with the word product. It is noted in the analyses found in Appendix 6 as follows:

Product

ADDITIONAL Sense: not noted in COBUILD: products denoted collectively

106 instances 14.3% of sample

Patterns: Uncount noun

We are on the point of having to source product externally.

‘Product’ used in this way acts as a replacement for ‘products’ in the plural, and cannot be used in conjunction with an article. In all of these instances product is not post-modified as in product group or product development. Examples are shown below:

It was noteworthy that this form, relatively common in the BEC (14.3% of the sample), is not mentioned by COBUILD at all, and is almost completely lacking from the Business English materials gathered for the PMC (2 instances only). It is thus a word that displays a unique sense in Business English, separated both semantically and by grammatical patterning. Students need to be made aware of just this kind of business-specific lexis and it is obvious that in this case it has not so far been done.

9.3.4.2 Discussion: business-specific grammatical patterning?

The question was asked at the beginning of this chapter as to whether any grammatical patterning typical to Business English could be found. Differences have been noted in three areas – limitation of use, change of meaning order and the observation of meanings/patterns special to the business environment.

Limitation: What this section has shown is that in the business lexis studied, word meanings are often restricted – 35 out of the 50 words displayed a limitation of meaning in the BEC – and, therefore, the grammatical patterning attached to those meanings is used, as opposed to others.

Change of order: Additionally, 18 words out of the 50 stressed meanings that were different in order of frequency than noted in COBUILD. This further predisposes the lexis

of business to form into certain grammatical patternings more often than is found in general English. For example, analysis of the word service showed that Sense 8 (one’s service in a company), marked by being an uncount noun, e.g. length of service, was the fourth most frequent sense of the word. The most common meaning in general English (service provided by the government in an organised way), which is denoted by being a count noun, was relegated to fifth position.

Special senses: The words partner, business, sale, package, export and product displayed senses not found in COBUILD – though often quite few in terms of occurrences. Each new sense displayed a typical grammatical formation: partner – verb + with; business – (conversational filler) this/that sort of + business; sale – (compound noun) point of + sale; package – count noun; export – adjective/uncount noun; and product – uncount noun. Again, these grammatical forms become more used in the business environment than in general English, but the forms themselves cannot be seen as unique to business.

In order to further identify business-specific grammatical patterning, in addition to the analysis of the 50 words discussed above, the five words from the BNC analysed for semantic prosody (send, manage, big, global and package) were also analysed for typical grammatical patterning. In the BNC much wider grammatical patterning was in evidence, and this was in turn related to wider semantic possibilities. For example, in the BNC, the word send showed a much greater tendency (8.64% of the sample) to be linked into a wide variety of phrasal verbs (send for, send off, send out). In the BEC, phrasal verbs only constituted 3.56% of the sample and only one was found (send out). Manage in the BNC included many examples of the negative form (can’t manage, could not manage) notably lacking in the BEC. Big also showed different patterning. In the BNC it was much more likely to be used with an indefinite article than a definite article:

Big in the BNC:

preceded by a – 238 instances 28.91% of sample

preceded by the – 126 instances 15.3% of sample

This differs from the sample in the BEC where ‘big’ was only slightly more likely to be used with an indefinite article than a definite:

Big in the BEC:

preceded by a – 160 instances 27.25% of sample

preceded by the – 130 instances 22.14% of sample

This may be a result of more first-time, direct reference to tangible items being found in the BEC:

In both corpora, big showed a tendency to form post-modified compound nouns. In the BEC examples were: Big Blue, big business, big money, big players, Big Bang, the Big Three, big-time – most of these being in some way business-related. In the BNC these were both less frequent, and those that did occur were much more general in their range, e.g. Big Mac, the Big House, Big Kath, Big Ben – with many examples including a person’s name, e.g. big Eleanor, big Fergus, big Richard.

In summary it can be said that the words analysed in the BEC and BNC showed that most words had patterns that appear to be typical of, but not unique to, the business environment. The choice of certain word senses over others created a tendency for some grammatical patterns to be used more than others, but these are mostly patterns that can also be readily found outside the Business English environment. Conversely – as with the words product, partner, supplier and communication – meaning/pattern distinctions seemingly specific to Business English can be found, but it must be assumed, without the benefit of further study, that these make up the minority of instances.

9.3.4.3 Sub-technical language and Pickett – a ‘footnote’

It has been established that word senses change within the Business English environment. This fact was discussed in Chapter 3, where the work of Pickett (1986a,b, 1988, 1989) was examined in detail. Pickett suggested that Business English was an ergolect – a work language of its own, displaying situational- and topic-specific lexis. Pickett’s other important idea of poetics – the way in which words are brought into Business English from general English in a scale of transparent meanings to opaque – was also discussed in Chapter 3. This section takes a brief look at one specific language environment in order to give empirical confirmation of Pickett’s ideas. A meeting (Meeting 2 in the BEC – 8,015 words) between the owner of a sportswear shop and a sales rep was recorded, transcribed and the key words of the meeting were obtained using Log Likelihood, with a p value of p = .000001, using the BNC as the reference corpus. These key words can be seen in full in the table below:

TABLE LVI: KEY WORDS COMPUTED FROM MEETING 2 IN THE BEC

n	word	meeting freq.	meeting %	bnc freq.	bncsamp %	keyness Log L.
1	shoe	59	0.74	21		558.0
2	gel	29	0.36	26		243.5
3	it’s	127	1.58	7 410	0.38	170.6
4	we’ve	49	0.61	1 179	0.06	137.6
5	heel	14	0.17	7		127.5
6	forefoot	10	0.12	0		110.1
7	yeah	110	1.37	8 825	0.45	97.5
8	rubber	14	0.17	33		97.2
9	mean	51	0.64	2 353	0.12	86.9
10	ds	8	0.10	1		81.8
11	sold	16	0.20	146		72.9
12	sole	10	0.12	27		67.1
13	carbon	9	0.11	24		60.6
14	product	15	0.19	193		58.9
15	selling	12	0.15	101		56.4
16	durado	5	0.06	0		55.0
17	trastic	5	0.06	0		55.0
18	know	71	0.89	6 204	0.32	54.9
19	neutral	7	0.09	11		53.1
20	trail	7	0.09	13		51.3
21	pillar	5	0.06	2		46.7
22	shoes	10	0.12	93		45.2
23	foot	11	0.14	135		44.2
24	marco	4	0.05	0		44.0
25	headingley	4	0.05	0		44.0
26	tartha	4	0.05	0		44.0
27	adidas	4	0.05	0		44.0
28	sales	12	0.15	181		43.7
29	medal	6	0.07	16		40.4
30	price	14	0.17	336	0.02	39.3
31	rear	7	0.09	37		38.8
32	but	90	1.12	10 615	0.54	38.2
33	technically	6	0.07	20		38.1
34	clearance	5	0.06	8		37.8
35	blown	6	0.07	21		37.6
36	laces	4	0.05	2		36.4
37	toe	6	0.07	25		35.8
38	racer	3	0.04	0		33.0
39	gallileo	3	0.04	0		33.0
40	medial	3	0.04	0		33.0
41	orthoses	3	0.04	0		33.0
42	nike	3	0.04	0		33.0
43	durados	3	0.04	0		33.0
44	asics	3	0.04	0		33.0
45	duo	4	0.05	4		33.0
46	we’re	22	0.27	1 173	0.06	32.4
47	because	35	0.44	2 733	0.14	32.2
48	lot	23	0.29	1 342	0.07	30.7
49	etc	7	0.09	71		30.5
50	that	163	2.03	25 240	1.29	29.8
51	sell	9	0.11	165		29.6
52	cushioning	3	0.04	1		28.5
53	thermo	3	0.04	1		28.5
54	breathable	3	0.04	1		28.5
55	running	11	0.14	306	0.02	28.0
56	erm	58	0.72	6 471	0.33	28.0
57	better	16	0.20	749	0.04	26.8
58	quid	8	0.10	148		26.2
59	retail	5	0.06	32		26.0
60	soccer	4	0.05	13		25.6
61	hundreds	6	0.07	70		24.6
62	than	28	0.35	2 265	0.12	24.4
63	more	37	0.46	3 539	0.18	24.4
64	certainly	11	0.14	372	0.02	24.4

The key words were then divided – using the concordancer to establish meanings – into the three categories shown below: general English, sub-technical lexis and specialist lexis:

TABLE LVII: DIVISION OF KEY WORDS INTO THREE CATEGORIES OF LEXIS

General English	Sub-technical lexis	Specialist lexis
shoe/s	gel	durado/s
heel	sole	trastic
forefoot	neutral	tartha
rubber	trail	adidas
product	pillar	gallileo
selling	rear	medial
foot	clearance	orthoses
sales	blown	nike
medal	racer	asics
price	duo	thermo
technically	cushioning
laces	breathable
toe
sell
running
retail
soccer

In general terms, Pickett’s idea of the grading of language from the general to the opaque can be quite clearly seen:

1. General words such as foot, sales and price.

2. Sub-technical words that have a special meaning in this field e.g. pillar, blown and cushioning. Examples of usage shown below:

pillar;

blown;

and cushioning:

3. Specialist words, such as medial, orthoses and thermo.

A clear example of Pickett’s notion of poetics was also found in one specialist word that did not qualify as a keyword. In the textile/sportswear business the word colourway is used to denote the colour scheme of a garment or shoe – examples shown below:

Thus, from two ‘everyday’ words colour and way, a specialist, opaque meaning is formed. More work would need to be done in this area, but this brief example lends empirical credence to Pickett’s ideas, and this could be one area for later research work. Reference back to the work of Pickett continues in the next section, with discussion on his notions of knowing and acting in business.

9.3.5 How are words distributed across Business English macro-genres?

One of the central issues when forming the BEC corpus was the division between ‘talking about business’ and ‘doing business’. It was felt that the language used in these two situations would be essentially different. This view on the potential differences in types of language matches Pickett’s (1988) observation that Business English language could be divided into knowing and acting Business English – essentially the difference between the language needed by pre- and post-experience learners. Pickett drew a table to show this division and suggested that ‘we ought to be able to supply words and expressions [in each box] that would not just as easily go in another’ (Pickett 1988:91). One lesser aim of this thesis, therefore, was to examine this claim to see if it could be statistically upheld. This was achieved by examining the range the words in the BEC displayed. In order to do this, the BEC was first divided into 28 macro-generic categories, for example, emails, reports, faxes, letters, phone calls (shown in Chapter 6, Section 6.3.3). Then, using the Dispersion plot function of WordSmith Tools 3, the same 50 words as above were analysed for their range and frequency within the macro-genres designated in the corpus. The full analysis can be found in Appendix 6 in Vol. II, but an example of a dispersion plot, for customer, is shown here:

Fig. 46 Dispersion plot for ‘customer’- each file represents one macro-genre

The dispersion plot computes two main sets of data. Firstly, the macro-genres in which the word customer appears, in order of relative frequency. Thus above, it can be seen that customer appears in the agreements file (reem~1.txt) most frequently (113 hits, 3.83 instances per 1,000 words), then in the miscellaneous (misc.txt) file, then in job advertisements (jobads.txt), and so on. Secondly, the Plot section on the right hand side shows – for each macro-genre – the dispersion of the word. The small black lines denote an occurrence of the word in the samples included in the genre. It can be seen, therefore, if the word had an even distribution or if it was tied to an individual document where it appeared very frequently. Customer appeared very frequently in one agreement – as seen by the block of black lines in the middle of the plot – but in contrast, was fairly evenly distributed in the company brochure category (combrocs.txt).

The dispersion plot mechanism provides a quick and visual method for viewing frequency and range of occurrence, categorised further by the macro-genres that make up the BEC. For this analysis it is important to note that the macro-genres were divided between genres that were concerned with talking about business, for example, UK television programmes, radio programmes, books, newspapers and magazines while the other macro-genres were concerned with actually doing business, for example, emails, reports, quotations, meetings and negotiations. Thus three important aspects could be easily determined:

1. The overall range of the word: Did each word appear only in a few genres, i.e. could it be seen as genre-specific, or was it spread evenly throughout Business English?

2. About vs doing: As each macro-genre was assigned to one of the two groups – either talking about business or doing business – it was then relatively simple to see where each of the words drifted, i.e. could the words be seen as words more often used for doing business or for talking about business?

3. Written vs spoken: The macro-genres also represented either spoken or written genres. Thus again it could easily be seen whether the words tended to be used more in written or spoken situations. Each of these three factors will now be considered briefly in turn.

9.3.5.1 Overall range of Business English lexis across macro-genres

The range of the words across macro-genres is noted in the table below. For each word the number of macro-genres in which it appeared is given – the maximum number being 28 (the number of macro-genres in the BEC):

TABLE LVIII: MACRO-GENERIC DISTRIBUTION OF BUSINESS ENGLISH LEXIS

WORD	N	WORD	N	WORD	N	WORD	N	WORD	N
customer	25	business	28	sell	23	high	27	sale	22
manager	25	investment	21	manage	20	big	22	merger	13
supplier	19	delivery	22	receive	25	low	23	trade	25
distributor	17	payment	24	confirm	19	global	19	package	23
shareholder	12	development	24	provide	27	international	26	export	17
employee	20	production	24	send	25	local	27	service	27
staff	24	commun- ication	21	develop	24	competitive	23	market	25
partner	21	competition	23	discuss	25	corporate	24	earnings	13
boss	11	takeover	5	achieve	22	strategic	19	performance	24
management	25	distribution	23	improve	23	financial	26	product	26

N = number of macro-genres the word was found in (/28 maximum)

The table is represented in graphical form below to show the distribution of the 50 words under analysis, across the 28 macro-genres. The range is from those words found in only five macro-genres (takeover), to those found in the full 28 (business). Most words appear in 20 to 26 macro-genres, with the highest score – eight words – being found in 25 macro-genres.

Fig. 47 Macro-generic distribution of the 50 words taken for analysis

This information on the range of the words across macro-genres is useful, but it needs to be taken further. For example, a word may have a wide range, but may in fact concentrate most of its occurrences to one or two specific genres. The next section will examine how the words were divided across the doing and about business divide and also how words were spread across spoken and written macro-genres.

9.3.5.2 Doing vs about and spoken vs written Business English lexis

For this section of the analysis, very simple techniques were employed. The dispersion plot of each of the 50 words under analysis was analysed. The top ten macro-genres where a word appeared were counted to see how many of the top ten macro-genres came from the written or spoken sections of the corpus and also how many came from the doing business or talking about business sections. This is explained in full below:

a) Written/spoken: the number of written and spoken macro-genres found in the top ten macro-generic occurrences of each word were counted and results noted. For example, the word big appeared in eight spoken genres and two written genres out of the top ten occurrences.

b) Doing/about: similarly, the same words were then analysed to see which macro-genres they belonged to in terms of genres that were about business (e.g. newspapers, radio, TV) and those that were concerned with actually doing business (e.g. faxes, emails, negotiation and meetings). Using the same example, big, six out of the ten genres where it most frequently appeared were about business genres, whilst four were in the doing business categories. Each word was then plotted on the table below. The table is divided into quadrants and it is thus simple to see where the words lie on these two scales.

TABLE LIX: PLACEMENT OF THE 50 WORDS ALONG THE SPOKEN/WRITTEN AND DOING/ABOUT CONTINUA

Spoken/About Written/About

	Spoken						>			Written
About	1	2	3	4	5	6	7	8	9	10
10							takeover
9
8
7					boss				competition
6			big			investment market		business corporate
5						manage	low competitive	high local merger earnings	global sale	strategic
4							shareholder trade	partner achieve financial	improve	distribution
3							distributor staff sell package	manager employee export	confirm international	production develop service performance product
2							send		customer management receive	development communication
1									delivery payment discuss	supplier provide
Doing	1	2	3	4	5	6	7	8	9	10
	Spoken						>			Written

Spoken/Doing Written/Doing

Discussion: Less attention was paid to the written-spoken divide, as it has been known for some time that certain lexical items are more likely to be found in written rather than spoken environments and vice versa.[187]

The results show eight words that fall clearly into the about business quadrants: big, boss, investment, takeover, market, business, corporate and competition. This was at first a somewhat disappointing result as it was expected that the words would be more equally distributed between the quadrants. The results, however, are mitigated by several factors.

Firstly, there are more written macro-genres in the BEC (15) to only eight spoken, making the written side of the corpus slightly bigger (56%) than the spoken (44%). Moreover, the doing business section of the corpus represents 59% of its total to 41% for the talking about business section. Added together, this predisposes the words, to a certain extent, to fall where they did.
The analysis carried out only looked at the ten macro-genres where the words appear most frequently, and not all of them. It is, therefore, unwise to draw absolute conclusions from these results.
Several words showed a greater tendency to be about words than shows up in the table, due to the macro-generic categorisations in use in the BEC. The word business, for example, had in its top ten job advertisements and press releases on the doing business side. These two macro-genres are classed in the corpus as doing business, as they emanate from the company and are considered part of the business process. However, they definitely talk about business whilst at the same time doing it. If these about-type genres had been reclassified, the result would have shown a greater drift towards the about side.
As Biber (1988) has shown, written genres cannot be considered the same in terms of lexical density and some written genres show characteristics of spoken language (private verbs, contractions, first and second person pronouns) more than others, for example emails compared to annual reports. Thus the written/spoken divide cannot be considered as mutually exclusive in terms of the language found there.

Putting the initial disappointment and these mitigating factors aside, it can be seen, though perhaps not to the extent thought, that lexis can be divided along spoken/written and doing/about axes. Based on the evidence above, it cannot be said, as might be inferred from the work of Pickett, that words would only happily fit into one category to the exclusion of the others, as most of the words were found to come from genres belonging to all parts of the quadrants. What can be said is that words display a tendency to fall into one quadrant or the other. Some words, such as delivery, payment, supplier and provide, were heavily linked here to the doing of business, whilst takeover was very much associated with talking about business. Most words, however, fell in between, albeit more on the doing and written sides. Pickett, it must be said, was more concerned with longer pieces of language – functional frameworks and situational language – linked to the differing registers of business. This notion of longer stretches of language leads to the next section, where analysis continues by looking at word clusters in the BEC. It will be seen that these clusters of words, in fact, fit more firmly into one quadrant or the other than single words did.

9.3.6 What kind of clusters can be found in Business English?

The notion of word clusters or multi-word items (MWIs) was dealt with in some depth in Chapter 4. It was noted there that native speakers of English store a vast repertoire of lexical chunks and collocations that speed language processing and ease communication (Nattinger 1980, 1988, Pawley & Syder 1983, Kjellmer 1991, Nattinger & DeCarrico 1992, Lewis 1993, 1997, Fernando 1996, Moon 1997, Williams 1998). It was suggested that these language chunks and collocations are very important for language learners to pick up (Lewis 1993, 1997, Hill 1999, 2000, Morgan Lewis 2000), meaning that analysis of the language chunks found in the BEC was an important element of the study. The field of study devoted to language chunks, however, as could be seen from the review of the literature, is vast and could easily warrant several complete studies in itself. The analysis undertaken on the BEC provides an overview of a limited number of lexical MWIs, but only scrapes the surface in terms of the work that could be done, and this will be pointed out as one area for further study on the BEC. In this study, the analysis took the following form:

1. The most frequent 2-8 word lexical clusters were computed by WordSmith.

2. These clusters were statistically compared to a similar list obtained from the BNC Sampler corpus, and 2-6 word key word clusters for clusters from the BEC were obtained.

3. For each of the 50 single key words analysed, 3-word clusters formed from these words were computed and stored.

4. Additionally, five 2-word clusters and five 3-word clusters were subjected to the same analysis as the 50 key words. The results of the analyses of these clusters can be found in Appendices 7 and 8 in Vol. II.

In this way it was possible to gain both an overview of the clusters as they were found across the whole BEC, and also examine individual clusters to see how they behaved locally in relation to semantic prosody, colligation and distribution across macro-genres.

9.3.6.1 Lexical clusters in the BEC at the macro-level

It was suggested in Chapter 4 that longer MWIs or clusters would be less frequent and more genre-specific. Henry (1996) suggested that ‘there is likely to be a very close link between certain phrases and certain ‘genres’’ (1996:297). Conversely, shorter clusters would be more frequent and multi-functional in purpose (Williams 1998). The results gained from the BEC bear out this suggestion to a large extent, though it will be seen that smaller clusters, too, have a tendency to be limited to certain language domains.

Frequency clusters: The ten most frequent 8-word clusters from the BEC are shown here:

Based on these clusters, Henry’s assumptions seem to be correct. Genre-specificity is shown by the fact that these clusters came from very specific parts of the corpus – faxes, letters and press releases – where the phrases were part of the format of the document. This is clearly seen above, where the most frequent phrase – if you have any difficulty receiving this transmission – comes from fax cover sheets. The low frequency of these longer clusters is also shown here – the most frequent appearing only 30 times in the whole corpus. Indeed a closer look at the full 8-word cluster list – not shown here – shows most 8-word clusters appearing only two times or less in the corpus. A further point to note is that even these long clusters are part of longer clusters. Cluster 1 above, for example, can be continued by please advise at once.

Moving on to 6-word clusters, the same phenomena of genre-specificity and low frequency can still be found:

Here, though, we see the first non-genre-specific cluster, at the end of the day.[188] These longer 6- and 8-word clusters are in marked contrast to the top ten 3-word clusters shown below:

Here the clusters are much more frequent – the most frequent occurring 451 times as opposed to 30 times for the 8-word clusters, and not seemingly tied to any specific genre – thus potentially performing multi-functional roles within the language. A sentence-length cluster, I don’t know, also appears here. The first stage of the analysis, therefore, shows that the most frequent clusters follow the patterns expected of them in the literature, in terms of their length and functionality. The next stage was to examine those clusters that could be considered specific to the business environment by using the same key word analysis carried out on single words.

Key word clusters: For the analysis, 2-6 word clusters were produced from both the BEC and the BNC Sampler and then compared to each other using the Log Likelihood statistic with a value of p = 0.000001. Examples of the most ‘key’ clusters can be seen below. The charts show the frequency of the clusters in the BEC (on the left) compared to frequency in the BNC (on the right – mostly zero) :

6-word:

BEC BNC

4- word:

BEC BNC

3-word:

BEC BNC

The keyness of the words is clearly visible by the fact that none of the BEC clusters, with the exception of if you have any and of the goods, appeared at all in the BNC Sampler corpus. The 4- and 6-word key clusters show remarkable similarity to the corresponding 4- and 6-word frequency clusters seen earlier, which is not a typical finding for the key word/frequency distinction. The key 3-word clusters, though, differ markedly from the corresponding 3-word frequency list. What is striking about the key 3-word clusters is their strong tendency to be genre-specific, rather than the more genre-independent frequency clusters. The most ‘key’ key word clusters, as found with the most frequent clusters above, were those that appeared as part of a document’s format (e.g. registered in England) or those that were repeated many times in the same document for some reason

(e.g. a level # performer). It can be assumed, therefore, that the high frequency, multi-functional clusters are equally used in Business English as in general English and thus cancel each other out.[189] What are left here, with the key clusters, are those clusters that relate specifically to business situations – hence their genre-specificity.

Word-specific clusters: Additionally, as part of the analysis of each of the 50 words, 3-word clusters were computed for each of them, using four occurrences of a cluster as the cut-off point, unless there was either very many, or very few clusters connected to a given word. These clusters can be seen in full in Appendix 6 in Vol II. These will be briefly discussed in relation to the analysis of the PMC later in this chapter.

9.3.6.2 Lexical clusters in the BEC at the micro-level: analysis of individual 2-3 word clusters

The discussion on clusters in the previous section was at the macro-level, with frequency and keyness being central to the analysis. In this section, a sample of 2- and 3-word clusters was chosen for analysis to see how they operated at the micro-level.

2-word clusters: Williams (personal communication 1999) suggested that there was little point in studying 2-word clusters as most of these were nonsense combinations. Whilst this was indeed true, analysis of the 2-word cluster key word list displayed several clusters of interest.[190] Five of these were chosen for their obvious business content: interest rates, cash flow, market share, stock market and Wall Street.

3-word clusters: The key word list for 3-word clusters was problematic, in that many of the most key clusters were actually very low in frequency, making any meaningful analysis difficult. Thus, the three most frequent 3-word clusters were chosen: a lot of, one of the, the end of, and then two key word 3-word clusters that had a frequency reasonable enough to warrant further analysis – in order to and we need to. The full analysis of these clusters can be seen in Appendices 7 and 8 in Vol II, but the results will now be briefly discussed.

a) 2-word clusters

Semantic prosody: All the 2-word clusters showed clear semantic prosodies to semantic sets that were both the same as those noted for the single words, for example, time, positive, negative and companies, but also prosodies unique to the individual clusters. An overview of the semantic prosodies for these clusters is set out below:

down

interest rates time stock market places

containment/control trouble/down

decisions market share obtaining/ losing positive/up

availability good news

cash flow negative Wall Street bad news

helping cash flow people

companies

Fig. 48 Semantic prosodic sets for 2-word clusters

The key finding here was that the semantic prosodies were able to illuminate the process of business thought with regard to the key business concepts they relate to.

Semantic prosodies show how business people think: It was found, for example, that interest rates were more than twice as likely to be discussed when they rise than when they fall (36% rising/high level and 14% falling/low level). Additionally, the surrounding co-text differed noticeably – negative words for rising interest rates and positive for falling. Examples from the BEC are shown below:

The co-text surrounding rising interest rates is often very negative:

Conversely, lower interest rates are accompanied by positive co-text:

A further prosody shows how interest rates are controlled and kept in check:

Business attitudes to interest rates are, therefore, shown clearly, not only in the semantic prosodies, but also in the surrounding co-text in which they are found. Interest rates can be seen, then, as something to be feared and often talked about on the rise (surrounded by negative language – murderously, disastrous), discussed less on the fall (surrounded by positive language – market-favorable, taking advantage), and further talked about when the importance of keeping them under control is necessary (peg, keep in check).

Interest rates was not the only example the ability of semantic prosody to reflect business thinking. Cash flow and stock market both displayed slight negative prosodies (9.09% and 25.71% of the samples respectively), leading to the idea that cash flow is more likely to be discussed when there is a problem with it, e.g. cash flow difficulties, than when it is in order, and similarly the stock market becomes ‘news’ when there are problems, e.g. stock market crash, stock market turmoil. The prosodies attached to Wall Street, on the other hand, were evenly balanced between good news and bad news.

Colligation: As will also be noted with the 3-word clusters, the 2-word clusters displayed both similar and diverse colligational patterning. The phrase interest rates was preceded by the pattern noun + preposition (in) in 8.6% of the sample. This pattern was strongly linked to a description of upward (72% of the pattern’s sample), and to a lesser extent, downward movement (18% of the sample).

Likewise, a smaller semantic prosody related to decisions – i.e. decisions made to raise or lower interest rates – had a typical structure of noun-on-noun or noun-to inf-noun as seen below:

Market share displayed colligation with prepositions (in and of) totalling just over 17% of the sample. These were also semantically linked: in to show where the market share was (market share in the UK) and of to show either the amount or the company (a market share of 13%, the market share of SIMS Graseby). Stock market was integrated into verb and noun phrases co-joined by the preposition on (11.42% of the sample) – floated on the stock market, making huge profits on the stock market – and was also linked to (often plural) nouns to form compound nouns (stock market crash, stock market indices, stock market losers). Finally, Wall Street appeared in 17.46% of the sample in the pattern noun + on + Wall Street:

b) 3-word clusters

Semantic prosody: Four out of five of the 3-word clusters displayed clear semantic prosodies, again both to some of the recurring semantic groups noted for single words, and also to unique prosodies of their own. The semantic relations are summarised in the diagram below. Indeed, these clusters displayed much stronger levels of semantic prosody than most of the single words. Two good examples stand out: the end of, which displayed a strong prosody with periods of time, and in order to, which associated with positive goals and outcomes.

people time

a lot of money/finance/economy the end of meetings

companies/institutions

positive

one of the negative in order to positive goal/result/

numbers outcome

Fig. 49 Semantic prosodic sets for 3-word clusters

The end of collocated with a wide range of words and phrases concerned with time, representing an overwhelming semantic prosody (76.71% of the sample), e.g. days, months, years and dates. Also notable was the phrase at the end of the day, which occurred 33 times, 30 of which were metaphorical, rather than literal usage. This phrase was often preceded by because or but.

The end of also had a minor prosody with meetings, with examples being that’s the end of the meeting or until the end of the meeting. In order to, similar to the end of, had a very strong prosody (71.72% of the sample), this time with a positive goal or outcome. The verbs following this cluster were very diverse – there was no single large number of a particular verb – but the overall intent fell within a clear positive semantic group. Examples are shown below:

A further point about this cluster is that even some seemingly negative verbs, for example avoid, showed positive connotation, in this case pre-empting difficult situations:

The 3-word clusters, then, show the capacity for linking strongly to semantic prosodic groups in the same way that single words did. They also showed both idiosyncratic and shared colligational patterns.

Colligation: A lot of and the end of were both, as could be expected, followed by nouns/noun groups. However, the end of was preceded by a preposition in 92.36% of all instances (at the end of, by the end of, towards the end of) whereas a lot of was preceded by an existential subject (there’s a lot of), in 10.64% of the sample; a modifier (quite a lot of) in 4.43% of the sample; or a personal pronoun (we get a lot of) in 10.42% of the sample. The differences are shown in the table below:

TABLE LX: COLLIGATIONAL DIFFERENCES BETWEEN ‘A LOT OF’ AND ‘THE END OF’

a lot of	the end of
followed by noun/noun groups	followed by noun/noun groups
preceded by prepositions	preceded by existential subject, modifiers & personal pronouns

One of the showed unique colligational patterning, being followed by a superlative (24.01% of the sample):

It was also followed by ordinal or cardinal numbers, for example, one of the first companies, one of the first credit-card banks (7.59% of the sample) and things + (that) + (personal pronoun), for example, one of the things that I had to do and one of the things we are really good at is pricing.

Clusters: Sinclair (1991), Williams (1998) and Lewis (personal communication 1997) all point to the fact that smaller clusters are often, in fact, just part of larger clusters. This was found to be the case with the 3-word clusters chosen for analysis here. Using the 3-word clusters as if they were a single word, 4-6 word clusters were computed for each of them. It was found, as noted above, that the larger 6-word clusters were smaller in number, whilst the 4-word clusters were more numerous and diverse. This can be seen, for example, with the cluster a lot of. The most frequent 6-word cluster was a lot of time and money (3 instances), the most frequent 5-word cluster was we have a lot of (11 instances) and the most frequent 4-word cluster was a lot of people (45 instances). The lengthening of short clusters to form part of longer ones was a recurring pattern with all the clusters under examination.

TABLE LXI: PLACEMENT OF THE 2- AND 3 WORD CLUSTERS ALONG THE SPOKEN/WRITTEN AND DOING/ABOUT CONTINUA

Spoken/About Written/About

	Spoken						>			Written
About	1	2	3	4	5	6	7	8	9	10
10
9
8				Wall Street*
7					one of the		interest rates
6					the end of		cash flow stock market	market share
5		a lot of
4
3					we need to
2									in order to
1
Doing	1	2	3	4	5	6	7	8	9	10
	Spoken						>			Written

Spoken/Doing Written/Doing

* Only found in five genres so this is an estimated positioning

Macro-generic distribution: Using the same spoken/about, written/spoken quadrants as for single words, the distribution of all the 2- and 3-word clusters together can be seen in Table LXI above. It can be seen that the phrases interest rates, cash flow, stock market and market share tend to be used when writing about business, whereas Wall St, one of the and the end of all fall in the talking about business quadrant. The number of clusters is too small to draw firm conclusions from this dispersal, but the significance here is that shorter 3-word clusters, which are included for their high frequency and not keyness, as well as the more business-related 2-word clusters, all drift towards one of the four corners. Therefore, although the functionality and generic diversity of smaller clusters is greater, they are still restricted in their use by higher level boundaries. These boundaries are the sliding scales between spoken and written form, and doing and talking about business. These boundaries represent a higher-level categorisation than individual genre, and can be seen as an over-arching structure around which Business English lexis is constructed. The idea is represented in the diagram below:

Supra-generic level

Doing business > Talking about business

Level of individual genre

Words Clusters

Level of individual genre

Written genres > Spoken genres

Supra-generic level

Fig. 50 Levels of lexical connection in the Business English environment

The words and clusters are interrelated. They are then further linked to individual genre to a greater or lesser degree, and finally the words, genres and clusters are encased by the two continua (doing-talking about – written/spoken), shown in the diagram, forming an integrated, fluid whole, at a supra-generic level. The key point to be made here is the fact that shorter clusters, previously thought to show a high degree of genre-independence, are now found to be assignable to one of the four quadrants, and their use is limited not to individual genre, but rather to a number of genres found in one of the four quadrants.

9.3.7 How do words associate with each other in Business English?

The idea of associates as put forward by Scott (1997, 1999) was discussed in Chapters 4 and 7. Therefore, the concept will only be briefly summarised here by reiterating Scott (1997:238) when he said that associates are ‘words that are key in the same texts as a given key key-word’. Associates, then are words that co-occur with a given key key- word[191] and WordSmith ranks the associates according to the number of files that they co-occur with the key key-word. An example, showing the top ten associates of business, is given below:

Here we can see that, for example, business associated with itself in 111 files, company associated with business in 27 files, we in 20 and customer in 20. Thus, where the word business appears, these other words are likely to be found with it to a greater or lesser degree of likelihood. The concept is akin to collocation. However, in this case, the only criterion is keyness in the same text, not necessarily actual proximity to the head word. Scott (1999, WordSmith Tools Help File) notes that, like collocates, associates can be strong or weak, and one- or two-way.

Analysis of associates was carried out for all the 50 key words previously under analysis, and a full description can be found in Appendix 6 in Vol II. Here only a brief overview of associates is given. In order to give the overview, the associates for the top ten key key-words of the BEC were analysed. The key key-words were business, company, customer, financial, management, market, price, product, sales and stock. The associates of all these words were computed and then compared using the Detailed Consistency Analysis function of WordSmith 3, with a low-level cut-off point of co-occurrence in four files. The results can be seen below. What is seen is that the ten key key-words – shown at the top on the horizontal axis – very often associated with a limited number of other words – the associates shown in the vertical axis – to a greater or lesser degree. This ranged from business that had links to all the associates, to stock that only had links to six of the associates shown.

TABLE LXII: ASSOCIATES OF THE TOP TEN KEY KEY-WORDS IN THE BEC

key key-words

n	word	files	business	comp-any	cust-omer	finan-acial	manage-ment	market	price	product	sales	stock
1	market	10	18	10	7	6	7	45	6	6	10	6
2	business	10	111	27	20	13	19	18	7	8	16	6
3	company	9	27	81	5	12	11	10	5	0	10	6
4	sales	8	16	10	8	0	0	10	6	12	58	7
5	cost	7	11	5	5	0	5	5	0	5	5	0
6	per	7	10	5	0	5	5	9	5	0	7	0
7	systems	7	11	0	6	5	8	5	0	6	6	0
8	companies	7	17	20	5	7	5	13	0	0	6	0
9	customers	7	18	11	15	0	0	9	5	10	15	0
10	customer	7	20	5	43	0	8	7	0	9	8	0
11	reporting	6	9	8	0	5	5	5	0	0	5	0
12	million	6	10	9	0	0	5	5	0	0	5	5
13	year	6	8	9	0	7	5	7	0	0	7	0
14	corporate	6	8	9	0	7	7	7	0	0	5	0
15	group	6	13	8	0	10	7	5	0	0	5	0
16	share	6	11	8	0	8	6	12	5	0	0	0
17	its	6	15	13	0	6	5	8	0	0	6	0
18	we	6	20	10	7	0	0	0	5	6	6	0
19	products	6	11	10	9	0	0	8	0	10	9	0
20	price	6	7	5	0	0	0	6	29	5	6	0
21	product	6	8	0	9	0	0	6	5	36	12	0
22	management	6	19	11	8	8	39	7	0	0	0	0
23	executive	5	7	0	0	6	6	5	0	0	5	0
24	growth	5	8	6	0	5	0	6	0	0	6	0
25	shares	5	6	7	0	6	0	7	5	0	0	0
26	or	5	12	0	7	5	0	0	5	5	0	0
27	our	5	10	6	0	5	7	0	0	0	6	0
28	markets	5	8	5	0	9	0	10	0	0	5	0
29	businesses	5	18	5	0	5	5	0	0	0	6	0
30	profit	5	11	9	0	8	5	0	0	0	10	0
31	stock	5	6	6	0	0	0	6	0	0	7	33
32	financial	5	13	12	0	33	8	6	0	0	0	0

Key: Associates are shown on the vertical axis (n=32), and the key key-words on the horizontal axis (n=10) – cut-off level is associate co-occurrence with the key key-words in 5 texts/files (1 text = 1 file).

To narrow the analysis down further, the way the top ten key key-words associated with each other was also investigated. This is shown in the smaller table below:

TABLE LXIII: ASSOCIATIVE PATTERNING BETWEEN THE TOP TEN KEY KEY-WORDS IN THE BEC

word	business	company	customer	financial	man- agement	market	price	product	sales	stock
sales				NO	NO
market
business
management							NO	NO	NO	NO
company								NO
product		NO		NO	NO					NO
price			NO	NO	NO					NO
financial			NO				NO	NO	NO	NO
customer				NO			NO			NO
stock			NO	NO	NO		NO	NO

It can be seen above that business, and market associated with all the other words and the other words associated with them. However, those words that showed no association seemed to be in a reciprocal non-associative relationship – where one word did not associate with another, it was found to be reciprocal, i.e. where word X did not associate with word Y, word Y did not associate with word X. For example, company did not associate with product and product did not associate with company; management did not associate with price and price did not associate with management. The word ‘NO’ has been used to show where no association occurs.

The sample used here is very small and the possibility of non-reciprocal association in the BEC cannot be ruled out, but the regularity in which the words associate or, more specifically, how they do not associate, points to essentially different semantic settings in which these words must regularly appear. This is also quite a surprising result intuitively, in that no associations were found, for example, between company and product or customer and price. Intuition would suggest that these words would be very closely related – where one was key the other would be expected to be, too. The fact that they are not often found together can have at least two possible explanations. Firstly, it may be that their occurrence is affected by relative frequencies in the corpus – that is, the corpus may be skewed and may not represent the words well enough. However, a second explanation is more likely – simply that in the texts where one is mentioned, it is not common to mention the other. Taking the example of company and customer noted above, analysis of the three annual reports in the BEC showed that, indeed, company has a frequency of 132 (out of sample size of 34,537 tokens), whilst customer occurred only eight times in the same sample. Thus, whilst company is made key in the text, customer is not.[192]

This is clearly an area that needs more investigation and warrants more coverage than falls within the scope of this study.[193] What can be said here is that yet another level of infrastructure can be seen to be at work in Business English beyond the levels so far discussed. Previous discussion on the world of Business English showed that business lexis is linked to semantic groups by collocation and semantic prosody, that it falls into certain grammatical patterning shown by colligation, that it is to some extent ordered by the talking about-doing business continuum, and now, further, that the general lexical backdrop to Business English is painted by a limited number of associated words that frequently co-occur, or have a tendency not to co-occur, when given lexis is used. Two further examples are shown below for company and sales:

Fig. 51 Associates of ‘company’ Associates of ‘sales’

As well as the linguistic implications mentioned above, this associative relationship between words must also have pedagogical implications, and these will be discussed later in this chapter (Section 9.5.8). The final part of this section of the chapter will now summarise the findings so far, and give a definition of what this research has discovered about Business English lexis.

9.3.8 Business English: a summary

To conclude this section of the chapter, we can return to both the hypothesis under discussion and the questions related to it. Hypothesis 1 stated that the lexis used in Business English is significantly different from general English. Based on the analysis carried out in this study the answer is that Business English lexis is significantly different from general English lexis, but is, at the same time, very firmly attached to it. This may appear a somewhat contradictory statement and will now be explained by revisiting the research questions one by one.

1. Is there such a thing as Business English lexis and if there is such a thing what is it made up of ?

Analysis in this thesis has shown that there exists a clearly definable set of lexis that is statistically linked to Business English with a one in a million chance of error. This lexis is not defined by frequency of occurrence, though some of the words are indeed very frequent, but by unusual occurrence – either words that occur far more often in Business English than would be expected in general English, or words that occur far less in Business English lexis than would be expected in general English. When talking about key words it is important to remember that they sit on top of a vast amount of ‘ordinary’ language – the same language, in fact – as general English. Thus, Business English lexis does not exist as a separate entity of its own, but feeds from and to general language. This being said, Business English lexis displays several characteristics typical to itself.

The lexis of Business English is, to a large degree, formed from a limited number of semantic groups that create a ‘meaning world’ for business. This world is populated by business people, companies, institutions, hierarchy, money, business events, places of business and is marked by its positive, shallow, dynamic and non-emotive lexis. It is a world that is concerned with concrete entities, methods of communication and of quantity and measure. Conversely, when considering negative key words, the business world is clearly separated from the personal – it is not about society, family, house & home, personal activities, weekends and distinctly negative states. Lexis that reflects all that is deep, reflective and emotive tends to be pushed away by the business world. It is not a world for philosophical debate, but a world for practical action in relation to concrete entities. All this may possibly be deduced by intuition. What this thesis has done has shown that these intuitive thoughts can be backed up by statistical data, and that the key words gained from the BEC fall into clearly defined semantic categories. A final word of caution, however, needs to be stressed. The semantic categories identified, and the words’ allocation to them, was not rigid and all-inclusive. This thesis does not state these were the only semantic sets found, but that they were prominent and the words displayed tendencies to fall into one or other of the groups.

2. Can the concept of semantic prosody be found in Business English and if so are there business-specific prosodies?

Semantic prosody was clearly on view in the BEC and there were indications that business-specific prosodies were a feature of the lexis. The use of semantic prosody as a means of analysing the collocations related to business lexis showed that the way in which words occur and where they occur is not a matter of chance. Clear patterns were shown to exist between individual lexical items and recurring groups of meaning. Thus, a given word within its role as a member of a given semantic set, for example, employee as a member of the people in business set, collocated regularly with other recurring semantic groups such as companies and institutions. This feature is summarised in the diagram repeated below:

positive company/institutions money size

hierarchy Business Lexis extremes

products/services people business activities time/timing

Fig.52 Semantic groups commonly associating with business lexis

A given word of business lexis is, therefore, likely, to a greater or lesser extent, to collocate with words from the semantic sets shown in the diagram. Again, these groups are not exclusive, but show only strong tendencies. Words were also found to have their own unique individual semantic prosodies.

Analysis by semantic prosody also showed that words which normally have weak collocational ties have a tendency to become much more collocationally fixed in the business environment. Collocates, of course, are very often genre- or area-specific, but here it was seen that words which are normally collocationally open-ended, for example big, narrowed in their application considerably, to collocate more with a limited number of semantic sets.

The possibility of business-specific local semantic prosody was shown by a comparison of the use of words in the BEC to the same words’ usage in the BNC. Distinct differences were found between the content, size and type of semantic prosodies associated to the words. A wider study, however, would need to be done in this area.

3. What colligational patterns can be found in Business English and can grammatical patterning typical to Business English be found?

The colligational analysis of the lexis showed that whilst words of the same grammatical class will naturally have certain things in common, they can also differ widely in their grammatical patterning. It was noted, for example, when discussing the category of people in business, that several of the words displayed a strong tendency to be formed into 2- or 3-word noun or adjectival groups, whilst the others did not. This colligational patterning – the way in which words typically appear grammatically – provides key information for learners that has so far been largely ignored. This will be discussed later in the chapter (Section 9.5). Additionally, it was found that sub-technical language and specific business-related meanings abound. The analysis of business lexis and its comparison to typical frequency/meaning relationships shown in the COBUILD (1995) dictionary showed that the lexis displays not only well-known sub-technical language – where a word has one meaning in ‘everyday’ language and a different meaning in a specific environment (Trimble 1985), for example partner (life partner vs business partner) – but also new meanings completely missed by both the COBUILD dictionary and the teaching materials, as in the case of product and export.

Analysis of the 50 words showed that a reduction in meaning potential was found in 35 out of the 50 words, and that 18 out of the 50 words showed an altered stress in meaning – i.e. the most common meanings for the word in general English were not in the same order in the BEC, and vice versa.[194] This section also showed that Business English lexis has a tendency to choose certain grammatical patterns over others, but that these patterns are, in the majority of cases, not applied prescriptively and uniquely and can also be found in general English. There are, however, many patterns attached to words that would be of great value in the classroom – a point that will be returned to in the third part of this chapter.

4. How are words distributed across Business English macro-genres and can they be divided along the ‘knowing’ – ‘acting’ axis of Pickett (1988)?

By using the Dispersion plot function of WordSmith 3, and a version of the BEC that was divided into macro-generic categories, it was possible to gain a clear picture of the macro-generic distribution of the chosen lexis across the BEC. All the words under analysis were multi-generic, i.e. they occurred across a number of genres. However, the words displayed marked difference in range of occurrence, from the word takeover that was found to occur in only 5 macro-genres, to business that appeared in all 28 macro-genres in the BEC. A further finding was that the words could be placed in one of four quadrants marked along two sliding scales: written-spoken and doing business-talking about business. This gave some credence to Pickett’s (1988) assertion that the language used to talk about business and the language used to do business would be essentially different. Pickett was more referring to phrases and linguistic patterns and routines that would be used in doing business, but not necessarily to talk about it. This analysis has gone a step further than Pickett did, by showing that the same can be said about individual words and they, too, can be placed on a doing business-talking about business scale. Moreover, it provides a framework for placement of business lexis that is only partially dependent on specific genre for positioning – it is related to, but at the same time independent of genre. It is an over-arching structure under which words can be placed that is above the level of individual words, clusters and genre. This phenomenon is even clearer at the level of multi-word clusters.

5. What kind of clusters can be found in Business English and do business-specific clusters exist?

Clusters found were both totally business-specific and unlikely to be used out of it, and also others that, whilst being independent of business, were, of course, usable within it. At a general level it was found that the longer clusters were totally genre-specific (e.g. If you have any difficulty receiving this transmission), whilst the smaller clusters were more genre-independent and multi-functional. This was an expected finding, and has been discussed in the literature before (Henry 1996, Williams 1998). Additionally, the fact that shorter clusters were often part of longer clusters was clearly in evidence (Sinclair 1991). An example from the BEC is given below with the phrase one of the:

one of the

one of the biggest

one of the biggest businesses in

one of the biggest businesses in South Korea ….

one of the

one of the first

one of the first things

one of the first things that

one of the first things that they ….

However, perhaps the most interesting finding was that, like the individual words discussed above, the clusters could be placed into clear quadrants of usage on the written-spoken/doing-about axes. This suggests that even the very short clusters of two and three words can be placed into specific positions and are not as genre-independent as was previously thought.

6. How do words associate with each other in Business English?

The way words associate with each other in the Business English environment is not random. Analysis of the key key-words and their associates has shown clear patterning in the way that words associate with each other in the manner described by Scott (1997, 1999). Some words were found to have very broad associative patterns, for example the word business, whilst others formed very limited associations, for example stock.

7. Business English: an overview

Business English lexis is a structured whole. Little appears in the lexis of Business English by chance, and the impression gained from this study is of a lexis forming an integrated world of its own by an infrastructure that grows ever larger, until its outer limits meet those of general English. This is shown in the diagram (Fig. 53) below. The outside edge is marked here with a broken line to show that the boundary of Business English is porous and not completely self-contained. Thus, all the above must be seen in terms of strong tendencies rather than in absolute terms. Business English lexis does demarcate itself from the lexis of the everyday world, but the crossover point is fuzzy and open to interpretation.

The gathering of a Business English corpus of this nature presents not just an opportunity for analysis, it presents an opportunity to see what lexis is used in Business English both in terms of frequency and keyness. Thus, in one sense, the words themselves are as important as any analysis carried out on them. The words represent lexis that is core to Business English and accordingly should be core to any student of Business English. This moves the discussion on to more pedagogical areas, where the implications for materials design and the classroom of these findings can be discussed. Before that, however, the next section will look at how Business English published materials have thus far viewed business lexis, and will assess what assumptions have been made as to its nature.

Word

Cluster

Collocation

Semantic Prosody

Colligation

Associates

Genre

Supra-generic structure

Fig. 53 Business English lexis: ever-expanding segments that form part of the whole

9.4 Business English published materials

The second major hypothesis of this thesis stated that the lexis used in published Business English teaching materials is significantly different from the lexis found in real-world business life. The main research question asked whether significant lexical differences could be found between the language used in published Business English materials and the language used in business. It is the aim of this section of the chapter to answer this question.[195] The PMC was subjected to a similar but more limited analysis as the BEC – key words, semantic prosody, colligation and clusters – but with two important differences. Firstly, two sets of key words were computed and analysed from the PMC, and, secondly, only 5, rather than 50 words, were subjected to detailed analysis. The process of analysis is now explained in more detail.

1. Two sets of key words: When using the key word statistic, the role of the reference corpus is central to the results gained,as what is determined as key is only key in comparison to the point of reference used. In the case of the BEC, Business English key words were computed using a corpus of general English (the BNC), as a statistical reference point. For the PMC, the key word analysis was done twice, using two separate points of reference, thus creating two sets of key words that were essentially different in content.

Analysis 1: The first set of key words computed – resulting in Analysis 1 – were computed using the general English of the BNC as a reference point. The resulting key words were those which occurred significantly more frequently in published Business English materials than in general English. Analysis of these key words showed, therefore, how published materials lexically demarcate Business English from general English. This meant that the PMC was analysed in reference to the same corpus as the BEC – i.e. the BNC. Thus, as both the PMC and the BEC key words were formed by reference to the same corpus, it was possible to compare the key words gained from the PMC to the key words computed from the BEC ->BNC analysis, discussed in the first part of this chapter. This showed the different ways in which the PMC and the BEC were lexically demarcated from the same source – general English.

[158] The term sub-business words is used here to describe words that appear in the background of business, but are not pure business words themselves. They often take on different meanings or usage when appearing in the Business English environment as opposed to their use in general English.

[159] This was not without difficulties as will be seen later on, but did provide an approach that would enable both easier analysis and pedagogical application.

[160] Berber Sardinha (1999) discussed how key words can be used in text analysis and suggested two main ways of obtaining a representative sample of key words – either by using the majority of key words gained (50%+1), or by statistically determining a significant sub-set (1999:4-6). For work on the BEC it was decided to take all the key words into the analysis, in order to provide as broad a picture of Business English lexis as possible.

[161] This is not to say that the list provided by the key word analysis is exhaustive of all people concerned with business, but that of the people that it found, the vast majority are clearly related to business.

[162] Formal division into tangible and abstract was not without problems as some words, e.g. relationship, could fit to one or the other category. The observations noted, therefore, are based on an intuitive impression of the words gained.

[163] This category overlaps to some extent with the activities category, but has here been treated as a separate entity in itself. There are several ‘activities’ that can also be taken as ‘events’, e.g. delivery and installation, for example. The designations into the two categories rested on intuition and are therefore potentially flawed, but the categories themselves are valid: session, merger, valuation, for example, are clearly events, and distribution and competition are clearly activities. It is possible, however, that some of the words assigned to one or the other could be interchanged (delivery and installation, for example).

[164] Other finance-related words were also placed into other categories, for example, bankruptcy into the events section. It was admittedly difficult at times to know where to best place some of the words.

[165] And all of these can also be used as nouns.

[166] It must be acknowledged that ‘mobile’ is not an obviously positive adjective, collocating as it does mostly as in mobile phone, but arguably it is not negative in any way and may be seen in some circumstances as positive. Mobile is also on its way to becoming a noun, as in my mobile (phone).

[167] It is obvious that some of the key words, as with favorite, were key owing to their US spelling.

[168] Analysis of the word ‘partner’, as seen in Appendix 6 (Key word 1.8) reveals that it is the business sense of this word that is by far the most frequent usage in the corpus, as opposed to its usual most common meaning of ‘life or sexual’ partner.

[169] See discussion Chapter 3 on Business English and genre.

[170] The word God has been included here, not because God is considered a person per se, but (without going into a prolonged philosophical debate) a person-like figure.

[171] FT is probably included as being key here because of its high use in the BNC.

[172] It is interesting that the word ‘truth’ occurs far less in business life than in non-business life.

[173] US use of the non-standard gonna and wanna account for the presence of these verbs. Gonna is used 111 times out of 112 total occurrences, and wanna 18 times out of 18 total occurrences in transcripts of US television programmes. BEC/BNC differences in transcription account for their negative key value here.

[174] The same reservations regarding word class distribution apply to this category as were expressed when dealing with the positive key words, i.e. words in noun and verb form can differ semantically and are, therefore, more difficult to categorise semantically.

[175] These comments by Williams were echoed by Watson (1997), discussed in Chapter 3, in his study of managerial discourse inside a Midlands company. The very same quote was used: the new owners of the company ‘spoke a very different language’ compared to the older, original workers in the company. Watson concluded, to reiterate the point made in Chapter 3, that ‘management is talk. And work organisations are constituted through the dialogues and discursive processes which that talk creates and expresses’(Watson 1997:226).

[176] Stubbs (1996) did analyse some words that can be regarded as business-related, e.g. labour, employment, worker, career and job. See Stubbs (1996) Chapter 7.

[177] A statistical analysis of collocates is included in Appendix 10 in Vol. II using the MI statistical procedure. This has not been included the full analysis here owing to the skewed nature of MI results (see e.g. McEnery & Wilson (1996:71-72) for more discussion on this).

[178] Here we can see the link between semantic prosody and colligation: meaning and grammatical form match and the two concepts cannot easily be separated. As Hargreaves noted in Chapter 4 when discussing collocation and colligation ‘there is no sharp distinction between the two’ (Hargreaves 2000:213). This will be discussed in more detail later in this chapter.

[179] As mentioned previously, high and low are also have a nominal function, but are here used mostly as adjectives (high 96.11% adjectival use, low 92.27% adjectival use).

[180] The same finding was made for low: words referring to amount/measurement tended to stress the extreme:

Colligationally however, high and low differed and this will be discussed in the next section of this chapter.

[181] Mc Carthy & O’Dell presented the phrase up to the ears in work – Hoey showed that this form is colligationally untypical, and that the phrase occurs almost exclusively with a possessive pronoun, as in up to his ears in debt.

[182] Autasys worked well, having a high recognition rate, but had problems especially in recognising verbs. For example, it assigned all patterns of ‘to + word’ into the verb category whether the word after ‘to’ was a verb or not. The tagged version of the BEC can be found on the CD ROM accompanying this thesis. It uses the LOB tag-set and can be viewed using WordSmith.

[183] See Willis (1990:53-56) for examples of how this was done in the original COBUILD project.

[184] COBUILD uses the term ‘meaning’, though in many cases it was felt that the definition was just a different ‘sense’ of a word. Thus, the word ‘sense’ has also been used.

[185] The words chosen for this group are not as homogenous as Hoey’s and so could anyway be expected to differ more colligationally. Moreover, the word tendency must once again be stressed – here we are talking about typical patterning; it is not to say that less common patterns are not possible or important.

[186] Two points must be made here. Firstly, although there is little doubt that Business English does display fewer meanings than general English, limitations in the corpus may account for some of the lack of word meanings. Secondly, some of the meanings given by COBUILD were tied to word classes and not taken into analysis here, i.e. only head-word analysis was carried out. Thus, if, for example, a word meaning utilised a passive form, it was excluded from the analysis.

[187] Stenström (1990) reported on lexical items peculiar to spoken discourse, using the LLC and LOB corpora for comparison. Stenström was interested in spoken discourse items where simple frequency of a lexical item was not enough, but it needed to have a discourse function. She found that these ‘D-items’ could be single words like well, oh, all right, or longer clusters such as as you know (Stenström 1990:139).

[188] This cluster was also the most frequent 6-word cluster found by Williams (1998:42) in her 117,000 word corpus of negotiating simulations.

[189] There was a small number of negative key clusters that were multi-functional, i.e. appearing less in Business English than in general, but the general trend was that that they were equally used.

[190] These are in fact 2-word collocations, but the term ‘cluster’ has been kept in order to maintain uniformity in this section of the chapter.

[191] Key key-words are words that are key in a number of texts. WordSmith ranks the words in order of how many files the word was key in, with the highest first.

[192] In other macro-genre the relative frequencies of the words were reversed. For example in Agreements, customer occurred 98 times and company only 59 (total sample size 29,602 tokens). Thus, a reversal of keyness is seen: where one is more common the other is less, and vice versa.

[193] This is a very interesting area. For example, whilst the word product does not associate with company, the plural products does (10 instances). This shows: a) that the words possibly occur in different semantic environments; and b) that words and their lemmas must be treated separately in linguistic analysis.

[194] There will be more discussion on sub-technical language the section on pedagogic implications later in this chapter.

[195] It should be stated at the outset, however, that the analysis of the published materials plays a lesser role in this thesis than that of analysis of real-life Business English. This was a necessary compromise due to considerations of time and space, and the possibility of more extended work in this area could be one theme to be taken up by later researchers.