Hacker News Reader

Comment by shmageggy | Hacker News Reader

I'm continually astonished at how much sway Universal Grammar still holds. This entire field has completely ignored the fact that AI and machine learning exist. Chomsky was and is a genius, but Max Planck was right when he said science advances one funeral at a time.

View on HN

I definitely agree that universal grammar has been dying a long slow death, but despite press coverage, I suspect universalists are the minority nowadays. Either way I hardly think linguistics has ignored AI/machine learning. In general there's an anticipation of eventual synergy between the domains, but ML has yet to magically solve the "poverty of stimulus" issues with pattern recognition that originally prompted hypotheses like UG...

Ah sorry, by "entire field" I meant the subset of linguists who are doing this kind of research. And I agree that AI/ML hasn't solved every problem raised by the poverty of the stimulus, but more importantly I think it is clear at this point that we should be framing these phenomena in terms of learning and inductive biases rather than solely in terms of UG.

The "poverty of the stimulus" is no more than a dubious hypothesis.

Chomsky's main influence came not so much from any particular theory as from the fact that he offered a coherent research program to a generation of linguistics grad students.

The resulting hangover is long and painful.

Universal grammar is a theory about how the human mind processes language. AI and ML don't bear on it at all.

Reply on news.ycombinator.com