Dictionary building procedure

The Dictionary Express method

Our cost and time saving method for creating dictionaries consists of 4 steps:

corpus building
automatic generation of the dictionary content
post-editing
publishing or deployment

Steps 2 and 3 are closely interconnected and typically carried out simultaneously. While some dictionary data are being generated, other data are already being postedited. The postedited data may affect the generation of the data for the next entry component.

Access

The access to the Dictionary Express method is not through a tool or a user interface. The method uses the existing tools and interfaces of the Sketch Engine corpus query system and the Lexonomy dictionary writing system.

The added value consists in the workflows and editor and data management developed by Lexical Computing.

NEXT: Corpus

Traditional vs modern

There is a major difference between the traditional methods of dictionary building and the Dictionary Express method.

entries from A to Z

The traditional approach of dictionary editing followed the alphabet and involved whole-entry editing. The work started from entries starting with the letter A and after the first entry was written, the work moved to the next entry. This required the lexicographer, or dictionary author, to possess deep linguistic expertise to be able to edit all the components of the dictionary entry.

entry component by component

Dictionary Express builds the dictionary by components. The same type of component, e.g. example sentences, is generated for all dictionary entries at once. The component is post edited in all the dictionary entries, and then the next component can be generated for all the dictionary entries. This is followed by postediting again. The process continues until all components of all the dictionary entries are generated and post edited.

Finding an expert editorial team for a new project is often difficult or even impossible because professionals with a lexicographic expertise are very scarce. In the case of less resourced languages, it may be completely impossible to find such experts. In addition, highly-trained experts can affect the costs.

With this approach, editors only need to be trained in one narrow and clearly specified skill. This makes it easy to train the editors in a short period of time. For example, for the headword list editing task, the editors only need to be trained in distinguishing words which should and should not be headwords. Working with non-expert editors means lower costs.

NEXT: Corpus

The Dictionary Express method

Access

Traditional vs modern

entries from A to Z

entry component by component

We are

Our corpus tool

Learn from us

Our term extractor