Lecture | LUCL Colloquium - Series '24/'25
Engineering in language research
- Date
- Thursday 12 December 2024
- Time
- Series
- LUCL Colloquium - Series '24/'25
- Location
-
P.J. Veth
Nonnensteeg 1-3
2311 VJ Leiden - Room
- 0.06
Abstract
Language research is dominated by the Popperian approach. Bold hypotheses are presented and falsified. Most research is about rejecting proposals presented by others. An alternative approach is one in which verification and optimization play a crucial role. Before attempting to reject hypotheses, one must have a good idea of the environment in which one is working and the tools at one's disposal. This engineering approach is primarily aimed at improving existing techniques. Better prediction is more important than deeper understanding. I argue that both approaches have their strengths and weaknesses and that language research is better off embracing both rather than glorifying one at the expense of the other.