python print str.decode（＆＃34; utf-8＆＃34;）UnicodeEncodeError

Question

我想将python字符串（utf-8）转换为unicode。

word = "3——5" # —— is u'\u2013', not a english symbol
print type(word).__name__ # output is str
print repr(word)          # output is '3\xe2\x80\x935'
print word.decode("utf-8", errors='ignore')

我收到了这个错误

UnicodeEncodeError: 'ascii' codec can't encode character u'\u2013' in position 1: ordinal not in range(128)

但是当我改变时

word.decode("utf-8", errors='ignore')

到

word.decode(errors='ignore')

错误消失。

为什么呢？ word是一个utf-8字符串，为什么我可以指定utf-8来解码？