Language Log

One of Matt Wuerker's 2017 political cartoons:

Yesterday BBC's Radio 4 program "Today", the cultural counterpart of NPR's "Morning Edition", invited into the studio a robot from the University of Sheffield, the Mishalbot, which had been trained to conduct interviews by exposure to the on-air speech of co-presenter Mishal Husain. They let it talk for three minutes with the real Mishal. (video clip here, at least for UK readers; may not be available in the US). Once again I was appalled at the credulity of journalists when confronted with AI. Despite all the evidence that the robot was just parroting Mishalesque phrases, Ms Husain continued with the absurd charade, pretending politely that her robotic alter ego was really conversing. Afterward there was half-serious on-air discussion of the possibility that some day the jobs of the Today program presenters and interviewers might be taken over by robots.

The main thing differentiating the Sheffield robot from Joseph Weizenbaum's ELIZA program of 1966 (apart from a babyish plastic face and movable fingers and eyes, which didn't work well on radio) was that the Mishalbot is voice-driven (with ELIZA you had to type on a terminal). So the main technological development has been in speech recognition engineering. On interaction, the Mishalbot seemed to me to be at sub-ELIZA level. "What do you mean? Can you give an example?" it said repeatedly, at various inappropriate points.

Today's SMBC:

From Soonish: Ten Emerging Technologies That'll Improve and/or Ruin Everything, by Kelly Weinersmith and Zack Weinersmith:

In celebration of Geoff Pullum's 700th LLOG post, "World domination and threats to the public", we'll be meeting for a quiet (virtual) drink this evening. But meanwhile I'll quietly suggest that Geoff has been too hasty in joining Randall Munroe at xkcd in assigning to the field of Linguistics a "low likelihood of being a crucial tool for a supervillain, and low probability of anything breaking out of the research environment and threatening the general population".

In fact LLOG posts have described at least two fictional counter-examples  over the years, and I expect that commenters will be able to suggest some others.

Linguistics is in the most desirable quadrant according to today's xkcd: low likelihood of being a crucial tool for a supervillain, and low probability of anything breaking out of the research environment and threatening the general population.

But I'm not at all sure that everything is positioned correctly. Molasses storage should be further to the right (never forget the Great Boston Molasses Flood of 1919); dentistry should be moved up (remember Marathon Man); robotics in its current state is too highly ranked on both axes; and entomology, right now (October 18, 2017), in addition to being slightly too low, is spelled wrong. Lots to quibble about, I'd say. But not the standing of linguistics as a safe thing to work on.

Today's Frazz:

Today's SMBC:

Today's SMBC:

Today's xkcd:

Mouseover title: "It's hard to train deep learning algorithms when most of the positive feedback they get is sarcastic."

Pearls before Swine for 7/23/2017:

A couple of Generation Z language consultants confirm the accuracy of the translations. Or as one of them put it, "Haha right".

Today's SMBC:

For a more complex but pointed analogy, see "The Pirahã and us", 10/6/2007.

Today's xkcd:

Mouseover title: "If you draw a diagonal line from lower left to upper right, that's the ICP 'Miracles' axis."

[The ICP reference, if you don't get it…]