Etymology is the study of the history of individual words. The origins and developments of Medieval Latin words are extremely interesting in their own right, but for the DMLBS etymology is of particular importance because, although it is not generally used to order the senses and subsenses of a word (as in some dictionaries), the history of a word is used to determine the headword or ‘lemma’ form that is used for an entry and to decide which examples belong together within a single entry. This is necessary because of the complex variation found in Medieval Latin spelling, meaning that the ‘same’ word can be found in an assortment of different spellings.

Many modern languages, such as English and French, have highly standardized spelling systems, but Medieval Latin was rather freer in its spelling. Classical Latin, which did have a well-standardized spelling system, retained its prestige in the Middle Ages, and so words with origins in Classical and Late Latin and their derivatives tended, as might be expected, to show some degree of adherence to Classical norms in spelling. However, especially for words borrowed from other contemporary languages (which themselves generally did not yet have standard spelling systems), we find a considerable degree of variation because naturally they do not have a standard Classical form and may even contain sounds that Classical Latin simply did not have. (This variability in spelling often extends to words with Classical origins too, depending on the writer's level of proficiency in the Classical spelling system or desire to adhere to it.)

In view of this the DMLBS selects as the lemma form for each entry the standard Classical or Late Latin spelling if one existed. For words derived from these by regular processes of word-formation (e.g. the creation of adverbs from third-declension adjectives by the addition of the suffix -iter to the stem) again the standard form expected as a result of the derivation is adopted.

For words borrowed from contemporary languages, we identify the form that best reflects the word and its history within the language from which it has been borrowed. This can sometimes be difficult to establish, particularly when there are a number of equally possible sources: we consult the best available reference works on the contemporary vernacular languages (Old and Middle English, Anglo-Norman French, etc.), but often we find clear evidence of related words in use in more than one of these languages but the relationship between the words (and which is the source) is not so clear. Particularly problematic are examples where the apparently related vernacular words are first attested only some time later than our sources, as might be expected when Latin was more often the language of written record than the vernacular. Sometimes we find the vernacular word is itself a descendant of a word borrowed into that language from earlier Latin, and when this vernacular word is borrowed back into Latin we end up with a so-called ‘doublet’. In these and similar cases we may decide to keep the evidence together in a single entry (quoting both headword forms) or separate the forms into two entries, depending on the complexity of the meanings.