The dataset is in three parts each being in the form of a table converted to a text file. There is a table of Verbs (split across two disks with the separate tables labelled Verbs_1.txt, Verbs_2.txt), a table of nominalizations from those verbs, including an indication of where a verb fails to license a nominalization (Nouns.txt) and a table which associates the primary key (Identifier) of each Noun entry with the primary key of the corresponding Verb entry (Links.txt). The latter table, when incorporated into an appropriate relational database, allows a very large range of sophisticated queries to be run on the Nouns and Verbs tables concurrently.This dataset is the result of two projects (ESRC Project numbers R000234783 and R000236115). The database was originally designed by Andrew Bredenkamp and Louisa Sadler. It was redesigned by John Gregory. Responsibility for further modifications and for the inputting of the data lies with Marina Zaretskaya and Andrew Spencer.
Standard dictionaries, native speaker judgements