[Continuing coverage of the 2009 Singularity Summit in New York City.]
Gary Marcus, an NYU psychologist, is underway with the first of the afternoon talks, “The Fallibility and Improvability of the Human Mind.” (Abstract and bio.)

Marcus starts off discussing natural language, and the common argument that it is highly imperfect. He says that many linguists, such as Noam Chomsky, have argued that natural language is in fact very close to an ideal system — that is, one that a set of super-engineers would design if they had built it from scratch. He says that he doesn’t want to make the argument that necessarily “things that look like human bugs are actually features,” but he says it’s something we should keep in mind.
Next Marcus moves on to discuss aspects of “the human system” that are problematic (he mentions the spinal column), and asking about limitations of the human mind. He starts with basic things, like how easily we forget simple things like where we left our keys. Computers, he note, store all memory in specific locations; our brains don’t. Minds are also susceptible to irrationality, like “framing effects” (where the same issue can yield very different opinions depending on how it’s described).
Now he’s launching into the selfsame systematic descriptions of the human mind that he seemed to be warning us against earlier: he says this last problem is due to something like “garbage in, garbage out” — that is, we remember best the last thing we have heard, and that’s why we are susceptible to framing. Marcus doesn’t seem to have considered his own advice of whether this apparent “bug” might actually be a “feature.”
Marcus concludes by saying that the current state of human biology is an accident, and there is room for improvement, “if we dare.” At least as far as this talk goes, that claim is short on evidence and long on opinion.
During the questions, an audience member asks about the case of those rare individuals who seem to remember everything. Marcus says this is more a disorder where people obsess about recording and documenting their lives, and that enhances their memory. He wrote a Wired article earlier this year about Jill Price, the most famous recent non-forgetter. Marcus says that while the media has focused on the sad parts of Price’s story — she cannot forget bad things — he knows of another case where the unforgetter is a DJ who seems quite happy. (More data points, please!)


  1. Is the fact that evolution has to stick to whatever currently exists a bug or a feature? Is nearly half of the American population dead by age 55 in the 1950's a bug or a feature?

    Those two alone could generate thousands of design problems. You really need examples?

  2. Matthew, if you're talking about my "more data points, please" comment, I was referring specifically to Marcus's claim about enhancing memory so that people don't forget. I was pointing out, rightly, that one data point showing it's bad and one data point showing it's good is nowhere near sufficient to conclude that it is either entirely bad or good.

    If you're talking about the idea that everything that appears to be a "bug" in human nature (or human evolutionary design, if you wish) is in fact a "feature" — I made no such claim in this post.

Comments are closed.