The Role of Explanatory Value in Natural Language Processing
It addresses a foundational issue for the NLP community by advocating a shift in research focus toward explanatory value, which could influence institutional policies.
The paper argues that explanation of linguistic behavior should be a primary goal in NLP, distinct from model explainability, and compares recent models of human language production to illustrate this point.
A key aim of science is explanation, yet the idea of explaining language phenomena has taken a backseat in mainstream Natural Language Processing (NLP) and many other areas of Artificial Intelligence. I argue that explanation of linguistic behaviour should be a main goal of NLP, and that this is not the same as making NLP models explainable. To illustrate these ideas, some recent models of human language production are compared with each other. I conclude by asking what it would mean for NLP research and institutional policies if our community took explanatory value seriously, while heeding some possible pitfalls.