Ann Irvine asked me to help with her project distinguishing scientific bird names with common bird names. Ann is trying to classify different bird names based on whether the common name is a direct translation of the scientific name. A direct translation relies on mostly Latin and Greek roots, with usually two different roots combined to form the scientific name. Common bird names always have three words associated with them, but the first two words are based on a translation and the last word is associated with the birds genus. She suggested I look at this webpage to learn more about the science behind bird names.
She wanted me to rank bird names based on this scale she provided me with:
1 = seems like a translation pair (e.g. if you see ‘flavi’ translated as yellow OR golden OR blond or… semantically similar, mark as 1.)
2 = half-way a translation pair (e.g. leucogaster=white chinned instead of white bellied, leuco/white is translation but gaster/chinned is not, so mark 2. OR if you see, e.g., ‘gularis’ translated as ‘blue throated’, mark as 2)
3 = have nothing to do with one another (e.g. minor/black-faced or ellioti/wine-throated)
Ann plans to use the rankings I give to test her code that attempts to rank bird names translation consistency.
In order to rank the words, I used a variety of sources to find Latin and Greek translations of possible roots in the scientific bird name. Ann told me to use a website about Etymology. I used Wiktionary to look up the roots and any of their possible meanings. I was also given a text file that contained the computer generated ranking list for the different bird names. I could search in the already ranked list for specific roots from the bird name I was ranking. If the file showed a higher number for the scientific root and the following word, I could assume the scientific root was associated with the English word. For example in the testing file:
flavivertex yellow crowned
This means the scientific name is flavivertex and the common name is yellow crowned. The word flavivertex can be broken up into flavi and vertex. Flavi is a root word for yellow and vertex is a root word for top or crowned. I would then mark a 1 next to the name pair. I can see that flavi is a word for yellow by searching through the training data. Just a section of the output when searching for flavi:
flavipes yellow legged 2
flavicans yellow breasted 1
flaviventris yellow bellied 4
flavigularis yellow throated 3
flaviventer yellow bellied 1
The higher the number following the name pair, the more likely it is a direct translation. From this output, it can be concluded that yellow is a translation of the root word flavi.
I ranked about 300 different name pairs and found that I ranked 77 as direct translations, 114 as half translations, and 147 as non-translation pairs.