Every politician knows: if you start being cartooned then this is a good sign.
So this time there's one about Computational Linguists. You don't know
what CLs are? Welcome.
This one is from
BoingBoing originally going back to the artwork done by
Randall Munroe.
As you might know I am a Computational Linguist myself
and so -- though I don't like taking a defensive stance -- I feel inclined
to drop some words.
So what's the message? Well, if you visit the original
cartoon you will notice that I somewhat cropped it, i.e. the ``Fuck Computational Linguists''. Sigh, yeah. Not, that Computational Linguists don't fuck. Maybe he just missed
a comma somewhere like in ``Fuck, Computational Linguists'' (thanks for the advice) or ``Fuck computational, Linguists'' (how bizarre) ...
The other more substantial part of the cartoon says that
the field of Computational Linguists was ``ill-defined''.
In fact, Computational Linguists know quite well that given a room of
n CLs there are
n+m (m>0) opinions. So the field is
highly contradictory not only among scientists but also everybody him/herself.
Is this a weakness? Definitely no. If you meet someone who is strong enough
to revise his angle given he followed one school for ages with quite some
evangelism, then you are a lucky one because those scientists are rare.
Most follow their track for ages not accepting what goes on to the left
and to the right, taking all of their youngers with them. Hard facts and plain numbers are ignored just because it does not fit into the picture. If it happens
that people come up with contradictory models and these get discussed
and compared then this is good. I don't have to elaborate on that, I think.
``Ill-defined'', this also subsumes something like ``fragmented''. As said
the field of CLs
is fragmented in the way that there is a huge variety
of models and methods at a first glance, which is good. What is bad -- and
that's fragmentation too -- is the fact that Computational Linguists find
themselves in the more broader area of Cognitive Science where we have
all sorts of overlapping research going on that is much more connected to
each other than we would like to modularize research. Neuroscience, Psycholinguistics, Philology, Anglistics, Theoretical Linguistics, and so on.
And most of them stay in their camp, albeit they share the same object
of interest: your blabla machine between your shoulders. That's not supprising
because the material is very hard to get within one span of life, progress
made very slowly. Some may even say there is a crisis of disorientation while
overcoming Chomsky.
So maybe Computational Linguists are all loony because they contradict
one another and themselves all the time but how about you: you talk talk
talk all day. Did you wonder anytime why you say what you say, why you
(don't) understand what someone tries to tell you, why you talk differently
than you write, how this speech and language thingy can happen to work out
at all? Probably not. For you, language is a natural given. You emit it. And electricity comes from of outlets in the wall.