String edit distance (also called Levenshtein distance) is used in Gabmap for measuring linguistic distances based on phonetic transcriptions. String edit distance determines the distance between two different pronunciations of a lexical item by finding the smallest cost for changing one pronunciation into the other. Changing one pronunciation into the other is done by inserting, deleting or substituting characters. This process results in an alignment of the two pronunciations.
You can click on alignments under Measuring technique in the project view. This will bring you to a view where you can choose any of the lexical items and a site in order to see the alignments.
|
What is the linguistic distance between the pronunciations of the word ‘rose’ in the counties Lebanon and Monroe? |
When using string edit distance to measure the linguistic distances in the data set, the distance between two dialects is first calculated separately for all lexical items. After that, the aggregate distance between the two dialects is calculated as the average of all the item distances.
(Alignments are created in Gabmap only for string data. For other types of data than transcriptions other distance measures are used which will not result in alignments.)