How is the sentiment scoring derived?

[geobetts]

How have the scores been assigned in https://github.com/sloria/TextBlob/blob/dev/textblob/en/en-sentiment.xml?

For example, how have polarity, subjectively, intensity or confidence been derived in:
<word form="afloat" cornetto_synset_id="n_a-533320" wordnet_id="a-00076921" pos="JJ" sense="borne on the water" polarity="0.0" subjectivity="0.1" intensity="1.0" confidence="0.8" />

Has someone labelled these manually? Or have they been learnt (via a machine learning model)?