Previous entry : The unbearable inevitability of discretization (1/2)
Next entry : The unbearable inevitability of discretization (2/2)
July 09, 2004Martin's follow-up
Martin Lessard, whose blog I couldn't find when writing my previous post, has committed a follow-up on the said previous post. Here is the english translation of my own summary of his text. I think I have extracted all of his arguments. I could probably have fused some of them, though.
- I was the only one at the RSS meeting defending AI. Anything that only one person defends at a RSS meeting is wrong. Therefore AI is impossible.
- Even though all the other attendants were trying to find a common agreement platform, I was not receptive to what they had to say. An unconvinced person is always wrong. Therefore AI is impossible.
- The ideas that I expressed are in the line of other people's ideas from 150 years ago. I am very naive to ignore the past. Therefore AI is impossible.
- There is no absolute, objective meaning. Will AI be able to interpret, participate and understand a community's moving, fuzzy consensus? No. Therefore AI is impossible.
- Any attempt to formalize language, in the past, has failed because communication and formalization tend to be mutually exclusive. AI cannot work without an extreme lexical AND grammatical formalization of the signs existing in a given context because it can't successfully deal with communication errors. To base an interpretation concept on a definitive signified is to negate the way we (humans) are reasoning. AI's concept interpretation is based on a definitive signified. Someone (maybe Russel or Husserl) said that scientific truths are beliefs. In the sense that they are temporary truths before being invalidated. Therefore AI is impossible.
- A semantic web malformed-query interpreter hyperGoogle that would pass the Turing test is trivial and would only satisfy a novice. Therefore AI is impossible.
- "Unbearable inevitability of discretization", really? Therefore AI is impossible.
- Martin wishes not to paralyze research. He only wants to prevent researchers to repeat their predecessors' mistakes. I should be grateful. Therefore AI is impossible.
Am I convinced? You bet! :) Am I wrong to not consider this as an attempt for reconcilliation? I don't think 1, 2, 3, 7, 8 need any comment from me and I won't reply to 4, 5, 6 today because I'm not done with the second part of my post. Notice that the title was The unbearable inevitability of discretization (1/2)
. That second part was meant to be my personal answers to the questions that end the first part.
Despite what it may look like, I'm trying to get to peacefully agree on some sort of common ground. So, please be nice. Okay? BTW, Martin, I can't trackback your post.
Posted by Vincent-Olivier Arsenault at July 9, 2004 08:27 AMSeeing my post translated that way, I may understand "you get pretty...err...intensely aggressive".
I like to specify 2 things that might have been overinterpreted:
1. By adding the repetitive "Therefore AI is impossible" for each point, which I never wrote "as is", you're adding, if not a confrontation smell, at least a serious denial of communication of my part. Which wasn't my intention. But my post can't prevent you to think so if effectively you were repeating this sentence after every of my arguments.
2. By adding your 2 first points (to be the only one to defends your point at the meeting and "an unconvinced person is always wrong") your post entice readers to think I use such silly arguments to promote my points. Please read again and take it out of the discussion as I think you want a higher debate.
That said, as a matter of fact, I have to agree you are feeling very strong about this topic. If you want me to be a proponent of "Therefore AI is impossible" side, we have before to agree on what AI exactly means. We might not have overlaping conceptions of it. I am looking forward to understand better your project.
PS: I corrected my backlink feature on my blog.
Martin Lessard
zeroseconde.blogspot.com