Next: Finding Information in Text
Up: Historical roots of Data-Intensive
Previous: Key applications
You may not feel in a position to answer these questions. However,
I doubt if anyone is actually able to answer them definitively, so
you should attempt them anyway.
- 1.
- Is bird song a language? How about whale songs? Why? Can you
think of ways of supporting your claims using corpus analysis?
- 2.
- How would you go about determining authorship of a collection
of disputed text?
- 3.
- What is the difference between a language and an artificial
code? Why exactly does Weaver's idea of treating Russian as
a funny encoding of English seem so strange?
- 4.
- A regular seeming signal arrives from a distant star. How would
you try to determine whether this signal is a sample from a language
spoken by some unknown intelligent life form?
- 5.
- People don't seem to assign probabilities to sequences of words,
so why should machines?
- 6.
- A possible objection to the
statistical approach is:
``Statistical models can't be right because
they assign a score even to obvious drivel. This makes them
worse than non-probabilistic grammars, which
reject such trash.''
Is this objection reasonable?
- 7.
- What use could linguistics make of a million hours of
annotated
video of everyday conversation?
Next: Finding Information in Text
Up: Historical roots of Data-Intensive
Previous: Key applications
Chris Brew
8/7/1998