In this Issue
Letters to the Editor
Write to Us
Spread the Word
Back Issues

How do you go about writing
a bilingual dictionary?


Language interference
The new Finglish:
from sex shops to big business

Book review
Faux Pas: A No-Nonsense
Guide to Words and Phrases
from Other Languages

British and American
Getting around in Great Britain

New word of the month
(Don’t) Read all about it!
New words in the media

MED Web Watch
Language Log

Your questions answered

How do you go about writing a bilingual dictionary?
by Sinda Lopez

• Introduction
• Dictionaries past and present
• Five basic steps to writing a bilingual dictionary
      1 Compilation of a framework
      2 Translation of the framework entries
      3 Bilingual edit
      4 Consistency checks
      5 Final edit and proofreading
• Conclusion


If I had a pound for every time I’ve been asked that question, I could probably stop writing them. I don’t think I’ve ever given anyone who has asked a straight answer though. It’s a process that not that long ago took about a decade, so how do you go about describing that at a dinner party, on a bus journey or indeed in a short article?


Dictionaries past and present

The very first bilingual dictionary I worked on, which passed through several publishers’ hands before ending up on the market, was the work of over 100 people and took around 13 years to put together. In those days, it was a painstaking manual process and computer use was limited to the IT people and keyboarders capturing the lexicographers’ work.

The whole business of dictionary writing – especially the timescale – has changed enormously since those days, as indeed have the dictionaries themselves, which in part answers the second most common question you’re asked when you reveal you’re a lexicographer: Why do we need any more dictionaries? My reply to that is usually simpler: Would they expect schools to still be using textbooks from the 1950s? A dictionary is a tool that reflects language and its use and must therefore keep abreast of developments and the times as well as increasingly meet the specific needs of particular user groups. Dictionaries need to be pitched to maximize their benefits. A dictionary aimed at young learners of English, for example, will differ significantly from one intended for university students or professionals. There are general dictionaries, business dictionaries, legal dictionaries; an endless list of types and permutations. Content aside, technological developments – most dictionaries these days are simultaneously published in print and electronic format – as well as the look and feel of products we choose to buy all contribute to the need for publishers to regularly update their products as well as occasionally aim to come up with ground-breaking new angles or distinguishing features that will further help and interest each new generation of dictionary user.

The essential process of compiling a bilingual dictionary remains basically the same although technological advances have improved and dramatically shortened the process. Whereas once you could have got away with just working on two or three major projects for your entire lexicographic career, these days a project can last from as little as six months to a couple of years, depending on how much data a publisher already holds in its databanks and how sophisticated their data-processing systems are to get at a targeted starting point for a new product.

One thing to remember about bilingual dictionaries that cannot be underestimated in the planning, is that a bilingual dictionary has two sides, each involving the source and target language with the basic process having to be worked through for each side by native speakers of both languages.


Five basic steps to writing a bilingual dictionary

1 Compilation of a framework
Two teams of lexicographers each produce a framework in their own language in which they categorize data into meanings and give examples of use as well as any helpful background information and comments. There are several steps to this initial process:

Gather linguistic evidence
The gathering of material needs to be in line with the publisher’s specification which nowadays usually has a pretty well-defined target market in mind. For example, the range of words selected as well as the detail of information to be covered and the examples chosen to illustrate the language in use, all need to be relevant to the age, level and purpose of the user that the publisher intends to sell the product to.

Whereas in my early days, this process involved a couple of afternoons a week leafing through publications ranging from The Economist to Hello, armed with a highlighter pen, a set of index cards for recording finds, and an all-important cup of tea and a biscuit, these days the use of a corpus, with millions of words captured and numerous search and query tools at the lexicographer’s disposal, takes out a lot of the effort (and fun!) as well as offering a more scientific and representative map of the language to start out from.

Painstakingly analyze and categorize it
Once the list of words to be covered is selected, lexicographers trawl through the evidence for those words to identify the most common structures and patterns that they usually appear in. They then start categorizing the different meanings and uses into separate senses and select natural examples and contexts to illustrate those meanings. In bilingual dictionary compilation, you have to go a step further and add several contexts that you know will not be published in the finished product but that aim to give the translator a fuller picture of the word and how it is used so that they can test out the most appropriate general translation that can be offered as an all-purpose equivalent.

Construct a framework
All this slowly builds up into a framework for an entry. With bilingual compilation, the framework is only half the story as you cannot decide exactly what will go into the final entry until the other language is in place and a lot of teasing and tweaking is required along the way.

Now that framework compilation is computerized, there are complex tagging structures to categorize the data down to the last detail. Everything from the headword to the punctuation is controlled by different tags and this makes the output and manipulation of data a lot easier and more cost-effective than in the old days of long-hand compilation and keying.


2 Translation of the framework entries
With the frameworks in place, translators can begin to translate the entries, sometimes working through several screens of data to arrive at a publishable two-line entry. Dictionary translation is a fine art as, no matter how many times you have previously worked through the same A-Z text, the perfect translation can still elude you as language is not an exact science and different languages relate to each other better on some levels than others. Concrete things do not tend to present too many problems but it’s the conceptual or cultural aspects of a language that can have translators immersing themselves in the evidence, looking up existing dictionaries, following a whole series of related entries and clues all promising something closer and better, sometimes to find that they’ve just been on a wild-goose chase.


3 Bilingual edit
With the bilingual text in place, the editing can begin. Screens of information have to be boiled down to tight but comprehensive entries that capture the key elements of the word without taking up any unnecessary space. This part is a delicate balancing act full of compromise.


4 Consistency checks
Much of the data in a dictionary is compiled systematically following product-specific rules and guidelines that could themselves fill another tome. Computers and tagging structures which allow extractions of the different types of information in entries have made checking things like groups of related entries, coverage of specific types of information as well as statistics on length and elements within entries a lot simpler and more efficient.


5 Final edit and proofreading
When every part of the process is officially finished, you need to go through the work one last time when the focus turns to assessing the quality of the finished product and approving it for publication. Once that is done, it can finally be proofread to eliminate any typos and remaining errors before passing on the finished text for printing.



That’s more or less it in a proverbial nutshell, as far as the A-Z text is concerned at least. Any special features throughout or supplementary pages are each a mini-project in their own right.

So, whenever I get asked the same old question again, I can just whip out a copy of this article and that should make anyone sorry they ever asked! On the few occasions I’ve mistaken someone’s passing curiosity as genuine interest and engaged with their opening question, another question invariably follows: But doesn’t it get boring? Well, yes, sometimes it does as no doubt aspects of every job do. The point is that nothing much is that interesting until you’re immersed in it and you have to struggle to make sense of it and get the job done. A bilingual dictionary is like an enormous jigsaw puzzle where you’ve got to keep trying the different pieces – although unlike a jigsaw puzzle, there are many possible combinations and pictures those pieces could make. That struggle, together with different language combinations, product specifications and the chosen angle, keep the process alive and bring a new dimension to the work each time.