Description of Spanish and English Assistant Programmes 4.

Discussion with business students in companies who need to write emails in another language has also revealed that many or even the majority of emails are not complete entities consisting of an opening, introduction, body and close but single sentences or short phrases which make up an on-going dialogue:

Company A: We're sorry, but we haven't got any of article number 123/789 left. You could try again next month.

Company B: Can you offer an alternative?

Company A: The only alternative we can offer is 126/456. Are you interested?

Company B: What size is model 126/456?

etc. etc.

Using the same tools in Fig. 12 the user can locate single question/answer type sentences for this style of email communication. There is also a section for on-going email exchanges which attempt to list questions and answers in groups so the user can access several alternative phrases together and select the most suitable. (Fig. 16.)

Fig. 16

Top

2.7 Assistant Trainer: for more advanced students.

Finally, both Assistant Programmes offer a Trainer version of Email Compiler. The Trainer is designed to allow users to write their emails making choices from phrases and expressions directly in the L2 rather than searching for language in L1 and then producing the translation. I am making the assumption that such a task will prove beneficial to new language acquisition but presumably for students who already have an intermediate or post-intermediate level in the foreign language (more is explained on these points in the section on Teaching Potential. As can be appreciated in Fig. 17, the student is now obliged to choose suitable English phrases from the drop-down lists. Choices made can be checked against a translation in Spanish but the user must pay particular attention to the English version before making a choice otherwise finding a suitable phrase or expression could be very time-consuming.

Fig.17

Other pages of the Trainer which require the user to use English is the Word Search page which must be searched in English to locate the desired phrases. Only when the drop-down list is located via the Word Search page can the user check his or her choice against a Spanish translation. See Fig. 18.

Fig. 18

Top

2.8 Building the Corpora for the Assistant Programmes.

There are two types of corpus in the Assistant Programmes. In an earlier version of these programmes the Email Compiler section was absent and the objectives of each Assistant Programme was different from those of the current version. Before the popularity of email writing there existed a greater use of business letter-writing via hard-copy (paper) and traditional mailing means. Most letters of this genre would be written in what we could term a “formal” style. A definition of formal can be found in the work Email English, a modern-day text book for teaching English email-writing to foreign students:

“This is the style of an old-fashioned letter. Ideas are presented politely and carefully, and there is much use of fixed expressions and long words. The language is impersonal. Grammar and punctuation are important. This style is not common in emails, but you can find it if the subject matter is serious (for example a complaint).” (Emmerson, p.8, 2004).

The original corpus in the Assistant Programmes was practically entirely formal English and Spanish. The time, effort and cost of writing and sending letters by traditional mail (often referred to as snail mail) meant that informal business letters were very rare; letters only had a practical commercial purpose and traditional formal letter-writing expression persisted - even into foreign language teaching.

This formal language corpus was taken from authentic business letters from both Spanish and English sources; mostly businesses who donated them. The personal details: addresses, names, dates and figures etc. were changed to maintain author and company anonymity. Another source of reference for set expressions and style analysis was taken from the works by W. Foulsham (English language) and R.M. Mata (Spanish language). These letters were then inserted into the original pre-email Assistant Programmes in two ways:

In the format of complete letters. An L2 translation was made of each copying the style of similar L1 originals. These bilingual versions were then placed one under the other on a page with a subject index so the user could locate them. The reader of this dissertation can locate the letter index at the link from the main menu in Email Compiler.

In the format of expressions and phrases extracted from the business letters and then rearranged with the foreign language translation. The bilingual pairs were then grouped according to some category with other bilingual pairs in a similar format adopted by a standard bilingual phrase book. For this reason, this corpus can be found in a suitably named page: (Libro de Frases). An extract from the category Complaints can be seen below in Fig. 19.

Fig. 19

Unlike paper-book phrase books, the entire corpus is machine searchable; a user can insert a word or phrase in the on-page search engine and instantly pinpoint the desired phrase or expression.

A second phrase book was also devised in a similar way to the first but including all phrases and expressions connected with the subject of writing cover letters to accompany a curriculum vitae for the specific pupose of seeking a post in a Spanish or English-speaking company. This page can be viewed at (CV frases).

2.8.1 Objectives of the original corpus.

This first, pre-email corpus of bilingual sentences, expressions and words required the user to be well-versed in the foreign language. The user had to locate a suitable sentence and then make necessary adaptations to the text which would include both lexical and grammatical modifications. Therefore, non-L2 language users or even low-level L2 users were unable to compile letters in the foreign language. This corpus was similar to a Translator Memory (TM), discussed in another part of this dissertation, or to the bilingual data bases found in Translator Workstations.

This original corpus has still been maintained in the current version of the Assistant Programmes but its relevance and practical application has been reduced to:

writing business letters (an activity less common than 10 years ago)
very formal emails where formal language could be appropriate eg. in complaint emails.
a combination of a “hybrid” neutral style so as to avoid an email which sounds to informal.

2.8.2 The Email Compiler Corpus.

Two reasons motivated me to update the original corpus.

Business letter-writing was becoming less popular as was the formal writing-style used in such letters. Correspondence via email in business was taking over as the most popular medium for in-company, inter-company and company-client communication. “...emails are by far the most common method of written communication...” (Emmerson, p.6). A survey carried out by Global Integration on the daily routines of businesses stated that: “The average manager in our survey receives 57 emails per day...” (Hall, 2006). An excessive amout surely! But it does emphasise the point. The speed and relatively little effort involved in email writing even meant that communication by phone was yielding to the power of email correspondence. The linguistic consequence of this revolution meant former letter-writing styles were being abandoned by native and non-native speakers of English and Spanish for a style more suitable to the email medium. Neither should we assume this style is just informal language where nobody minds if you make mistakes. Emmerson points out that business emails “take awareness and practice to write in a style that fits the context.” (Admittedly, I have studied this phenomenon in greater depth regarding the English language, where more published material is available. My knowledge of this occurrence in Spanish has been obtained only by personal perceptions in Spanish companies - not from publications on the subject. As a result Spanish email writing styles are more difficult to define.)

I wanted to help my students in the acquisition process of the L2 in question. The original programme was limited to a bilingual data base for reference purposes. This might have been useful for translators but for language learners the programme proved uninspiring and difficult to use and thus rated poorly as a didactic tool. (I also realized that without a large team of co-operators, I could not produce sufficiently extensive corpora to satisfy professional translator needs either.) I devised a hypothesis based on teaching methodologies such as Task Based Learning (TBL) which involved the prerequisite of a dynamic programme where the learner could be actively and cognitively involved in a translation process from L1 to L2. However, learner participation must range from a necessary zero cognitive involvement. That is to say, a user should be able to produce a translation of an email without the obligation for study. Only in this way would employee/learners use the Assistant Programmes as practical tools in the workplace. Language acquisition would occur, I reasoned, when the user had time to turn to the language help tools provided, such as the Trainer version, help pages, grammar sections etc..

2.8.3 Construction and compilation of the second corpus.

As mentioned above, I have identified three types of international business email and written correspondence:

In-company emails refers to correspondence between branches of the same company. Naturally, from a language viewpoint I refer to only when those branches are outside the country so that a foreign language is required for communication. From perceptions gleaned at Brenntag SA, Seville, in-company emails are usually between counterparts and discuss technical subjects common to the particular speciality of the sender and receiver. An example of in-company email correspondence would be where counterparts are attempting to solve problems in a common Intranet system.

Inter-company emails are those emails dealing with business operations such as purchases and sales - another company, then, is often the client of another. This category was defined from perceptions at Bordas SA, Seville. As the companies are likely to be related in product or service type there may be common jargon used in international communication, which may be difficult to understand for somebody outside that sphere of business. An example may be the sale of spare parts between the manufacturer and the assembling factory. In the case of Bordas SA, it is the import of base chemicals for the production of other chemical compounds used in foodstuffs.

Company-client communication is also based around the activities of purchases and sales. By client, here, I refer to private non-corporate entities. Jargon common to both will be used less in many cases during international communication unless the field is technical. There may also be a need here for letter-writing contact especially where no contractual bonds link the serving company and the client so confirmation of transactions must be via hard copy. An example of this type of correspondence (email or letter) would be the sale, before and after-sales service of products via a catalogue, brochure either in paper form or online viewed at a website.

An important point must be made at this stage: company correspondence is not readily available for the researcher. There exists the obvious policy of corporate secrecy which means obtaining large quantities of email samples is complicated. My own collection of donated letters I obtained from the few companies where I worked as a language consultant/teacher and therefore a trust relationship had already been established. The companies relied on my discretion and the rendering anonymous of all the letters handed over to me.

Bearing in mind the above point, I shall now continue to analyse the decisions I made regarding corpus choice. The above definitions of international business email and written correspondence types directly influence the type of corpora to be developed. In-company emails are usually technical and the extent of subjects that counterparts may discuss via email communication very large. Furthermore, the subjects would vary depending on the commercial activity of the company in question. Developing corpora for this type of correspondence would require a study of a large number of companies and the emails they sent. For these reasons, I therefore discarded this option.

For the above reasons, Inter-company and company-client email proved a more reasonable option for a corpus. However, this does not mean that access to such documents is easier but that due to the large number of monolingual and bilingual publications available, and often written by authors with experience in companies, corpus generation should prove more straightfoward. The main books I referred to were The Complete Letter Writer (1998) and Correspondencia Comercial y Privada (2000). As publications devoted entirely to formal letter-writing, the language used did now not interest me. Instead, I examined the subject matter that the letters were written about. Email styles may have revolutionized correspondence expression in the business world but the subjects remain the same: all activities revolve around the capitalist necessity for purchase and sales. The only additions I needed to make were the online/Internet contexts: website services, software product downloads etc. not covered by these publications.

2.8.4 The basic email subject list.

I first identified 16 basic written business correspondence subjects or topics (not in order of importance or popularity):

1, Business invitation.	2, Replying to a business invitation
3, Enquiring about a price.	4, Sending price quotes.
5, Offering goods or services.	6, Ordering goods or services.
7, Not satisfied and returns.	8, Complaints and late payments.
9, Sending things late, early or on time.	10, Replying late to correspondence.
11, Replying on receipt of correspondence.	12, Saying no to enquiries.
13, Asking for a reply to a previous email.	14, Confirming, postponing or cancelling.
15, Enquiring about training and courses.	16, Business trip travel booking.

(These subjects in English Assistant Business can be viewed at
. )

With reference to letters from the publications mentioned and those business letters from the first corpus, I brainstormed each subject for variables. This process was fairly straightforward. For example, on the subject of Replying on receipt of correspondence the brainstorming process would be similar to the following:

Line one

Occurrences: We acknowledge receipt of your email.
Reactions to events: We were happy to receive your email.
When the events happened: We were happy to receive your email yesterday.
Variables of the subject and verb: I was happy to receive your email yesterday.
Variables in the verb: I was happy to read your email yesterday.
Variables in the complement: I was happy to read your report yesterday.

Within the context of each topic a large number of possible scenarios is therefore devised for the first line of the email. The variables can be placed in convenient drop-down lists with colour bands added to make the reading of long lists easier on the eye. The same process is then carried out for a second line of the same email with as many lines as deemed necessary to cover the possible issues which could arise for each topic.

The opening lines (introducción) and closing lines (despedida) were actually compiled after all the 16 basic email subjects had been written. It was then easier to decide which openings and closes to choose by studying each subject and brainstorming possible sentences for each scenario. Where possible and appropriate, each sentence in the openings and closing lines was written in a less formal and more formal register:

We apologize for any inconvenience we may have caused...(more formal)

We're really sorry if we've put you out in any way...(more informal)

2.8.5 The questions and answers corpus.

The next consideration was to tackle what the sender of an email would do on receipt of a reply. Some topics in the original list of 16 are, in fact, replies, eg. Replying to a business invitation etc. However, replies are not always offered for all topics and then there is the case of replies to replies! It was not sufficient to provide the user with tools to write one email only per topic; email correspondence, more than letter correspondence, is comparable to a dialogue where several interactions between sender and receiver take place. It was therefore essential to devise lists of questions and answers for each of the 16 basic email subjects so that correspondence could continue and be interactive. Sub-categories of the 16 basic email subjects had to be devised to make it easier for the user to locate a specific topic. (Examples of these “dialogue emails” can be seen in the section in English Assistant Business.) This required more brainstorming around each subject to draw up a list of possible question and answer interventions in their various sub-categories. In Email Compiler, these lists are accessible either by subject (see and or via the Word Search tool.) An eg. from the English Assistant Business question list can be seen in Fig. 20.

Fig. 20

2.8.6 Other corpora included.

Promotional language glossary: (Phrasebook format) for writing marketing literature, spam-type emails, mail shot material etc. The glossaries have been taken from genuine material stuffed into my letterbox over the last few years! It covers many subjects from cameras to cots, soft drinks to software, mobile phones to mobile homes, etc!

Insurance terms glossary: Included in the Phrasebook. Contributions courtesy of business students at Willis S.A., Seville and for which I am very grateful.

3. State of the Art

3.1 Machine Translation systems and how they work.

Top