For a proper understanding of lexicographic methodology, it is important to distinguish sharply between the lexical database and the dictionary:

• The lexical database is a database in the technical sense, i.e. a computer file whose basic components are, logically, records with a certain internal structure.
• The dictionary is a text which importantly comprises a list of entries. Technically, it is a text file which may or may not be printed on paper.

The relationship between these two files is complex. Part of it is a mapping in the sense that the dictionary file is a (partial) export from the database.¹ Part of it is irregular, i.e. brought about by human discretionary intervention.

It is important to understand that the entire lexicographic research deals with the lexical database. During the entire compilation period, the dictionary does not exist. It is produced as the very last step by exporting part of the database in a suitable format. That step is discussed in another section.

Structure of the lexical database

In a first approximation, the structure of a lexicon may be implemented in a database as follows:

• The main sections of the framing structure of a dictionary are generally not apt to be elaborated in a database, with the one exception of the entry list. The other sections are essentially text files or sections of a text file.
• The entry list is implemented as a lexical database.
• Each lexical entry, constituted by a lemma, is a record of the database.
• Depending on the type of DBMS chosen, the database is just a set of records, each of which essentially contains the entire information pertaining to that entry, or the database is a set of tables, only one of which contains lemmas with certain kinds of information pertaining directly to them, while the other tables are linked to the primary one.
• The categories of lexical information (lemma, word class, inflection class, meaning etc.) are fields of a record.

The sections on lexical relations and on relations between the lexicon and the rest of the language system show that much if not most of the information associated with a lexical entry is by its very nature relational. The kind of database appropriate for representing a lexicon is therefore, in principle, a relational database. However, for practical reasons concerning features of currently (2010) available software, one has to make a choice between a relational database and a free-field-structure database. These are discussed in the section on databases.

¹ The opposite, i.e. the creation of a lexical database by importing a structured textfile, also occurs.