Oh, yet another exam time post! This one will be short since I actually am kind of worried I’m going to do badly on this one and thus actually have to study.
I’m going over my linguistics notes from early in the school year. We had a question about “the lack of invariance problem”. I myself have a problem with this problem and it’s blatant use of double negatives in order to sound more complicated than it really is. It’s the variance problem, plain and simple. Different people say words differently. That’s all it is.
Anyway, the question was what problems does the variance problem cause for speech recognition software? I’d answered the question on my own and then had written down what others had said in seminar.
Those in seminar had said things like “computers can’t understand sarcasm — they don’t understand emotion! A sentence means something different if there’s emphasis on it or not and computers just don’t get it!”.
My response was that computers simply can’t store all the possible variations of words.
Common sense, people.
It was the same with most other seminar questions. “Why teach chimps sign language?” Oh, well, it helps us to understand how humans acquire language blah blah blah blah. Or maybe because no one will give their kid to science so we can learn how they acquire language. I’m considering doing that, actually. Well, not quite that, but my kids are going to be total science experiments one day.
Anyway, just to brag about the final thing I’ll be able to brag about in linguistics, I got 95.5% on my annotated bibliography, which was, as always, the highest mark in the class. Sigh.