Exception when parsing ' ' with tokenize=False

This throws an exception with pattern 2.6:

from pattern.en import parse
result = parse(' ', tokenize=False)

The last frame of the exception:

/pattern/text/__init__.pyc in _v(self, token, previous, next)
    310         v = {}
    311         f(v,  "b", "b")         # Bias.
--> 312         f(v,  "h", token[0])    # Capitalization.
    313         f(v,  "w", token[-6:] if token not in self.known or token in self.unknown else "")
    314         f(v,  "x", token[-3:])  # Word suffix.

IndexError: string index out of range

Reproduce the exception with pattern.en.parse(' ', tokenize=False), then inspect pattern/text/init.py around the _v method and the call path from parse. Determine the expected handling for a space-only input and add a regression test; done means the example no longer raises IndexError.