Contemporary Folk Music


How did we tag our songs?

First, since we were looking at songs, it was important to maintain a specific structure throughout our corpus. Our schema started out with just the structure tags and elements. We had chosen a total of ten songs for each language so our XML documents contained the songs that were grouped according to what language they were.

As we looked at each individual song within our XML documents, tags such as meta and lyrics were the first important structure tags to group different kinds of information. We knew we needed to look at descriptive information about each of the songs as well as the lyrics themselves.

Within the meta element we had elements such as country, language, title, artist, year, dialect, and most importantly theme. Since not all of our songs within languages were from the same country, country could be a useful element for later analysis. Language was a more important element for our research. However, dialect within a language was not as important an element as we originally thought since it seems to be prevalent only in Italian folk songs. We chose to include the song title and artist name in order to differentiate both within our song collection and between the original version and any remakes that have been produced of the song. We chose year as another element to note because we were looking at a specific timeframe, from the 1960s to the present.

Within each individual song we had a tag for lyrics that came after our metadata header. Within the lyrics tag we had stanza, line, and noun elements. Our stanza element had a type attribute to distinguish verses, chorus, and bridges because we thought it may help our analysis. However, we did not end up using the distinction for analysis within the time frame of this semester. Line elements had no attributes and were included simply to maintain lyrical structure that most songs have when they are written down. Our noun tags were our most important tags.

Since our research question has to do with themes, we decided that a lot of themes can be determined based on word choice. More specifically, nouns tend to show or at least hint at the overall theme of a song. Therefore, we decided to look at all of the nouns, pronouns included. Within our noun elements we had the following attributes: typeN, cat, ref, foreign, subtheme. The typeN attribute was used to differentiate between pronouns, proper nouns and common nouns. The cat attribute denoted whether the noun was a person, place, thing or idea. The ref attribute was mainly used for pronouns that referred to a previous noun. The foreign attribute was used to determine whether or not the word was from another language or not since borrowing occurs in many languages.

The most imporant attribute for our analysis was the subtheme attribute. Unlike our theme element that was chosen amongst specific predetermined themes, we allowed for our subthemes to be more specific and determined by the context of the noun. The noun itself, regardless of the context of the whole song, has its own subthematic meaning. Not all of the nouns in our songs had a subtheme attribute, but there were a substantial number that were marked up in our corpus.

Creative Commons License