4 ms·

The flaw in the attitude of Sherlock Holmes is that he does not consider the, for the lack of a better term, "compressibility" of knowledge. Then again, Shannon

The flaw in the attitude of Sherlock Holmes is that he does not consider the, for the lack of a better term, "compressibility" of knowledge. Then again, Shannon wouldn't do his thing for another half century, so it would not be fair towards Sir Doyle to expect him to have taken this into consideration. The brain actually does a lot of lossy compression tricks to remember things together, plus I doubt that knowledge about the Solar system takes up any significant amount of space in the brain, since physics works with models. Very "compressible" information, really. But if we look at how much lore people remember about modern pop-culture trivia, like comics or TV series or games, it's a different story. The amount of mental effort and "mind space" that fans dedicate to remembering, say, World of Warcraft lore is mindblowing to me at times.

foxyv 8y ago
I think a significant portion of my brain (Similar in size to the knowledge from my Bachelor's degree) was dedicated to remembering every quest and map up to WotLK. It's surprising how much I still remember from 7 years ago.
- vanderZwan 8y ago
  Funny enough, that might also reveal something about how human brains store information: locations on a map sounds suspiciously compatible with the method of loci (aka "memory palace", thanks to a more recent incarnation of Sherlock Holmes) [0] https://en.wikipedia.org/wiki/Method_of_loci https://en.wikipedia.org/wiki/Method_of_loci
  - Retra 8y ago
    Music is also very easy to remember.
    - vanderZwan 8y ago
      IIRC there is recent research confirming it is at least partially stored in the same are responsible for muscle memory, which supposedly explains quite a few things.
      - ZeikJT 8y ago
        If that is true, the obvious question is why is it stored there?
        vanderZwan 8y ago
        Well, there are strong ties to movement in music. There is a reason that you don't dance to paintings, literature or poetry. So one option could be that music arises partly from our motor functions to begin with.
        vram22 8y ago
        >There is a reason that you don't dance to paintings, literature or poetry. Ha ha, well put. Although poetry is sort of like music, a bit, at least rhyming poems are. Just looked it up, this article is interesting: https://en.wikipedia.org/wiki/Rhyme https://en.wikipedia.org/wiki/Rhyme
Retra 8y ago
I think there's a related problem that would have been understandable at the time: if you don't understand how the world works, you are potentially missing out on patterned information which allows one to form more accurate models of the world across domains. The assumption that's made seems to be that different problem domains are all independent, but that's not completely true. That's probably the main reason that mathematics and language are so useful.
- vanderZwan 8y ago
  > The assumption that's made seems to be that different problem domains are all independent, but that's not completely true. That's probably the main reason that mathematics and language are so useful. To be fair, the effectiveness of multi-disciplinary research seems to only have been acknowledged in recent decades, even though by and large the biggest breakthroughs have almost always been thanks to people looking outside of the silos of their field. The sciences are really "tribal", and it used to be much worse. For example, when I started studying in the early 2000s I remember the ML bachelor (then: AI) having recently split off from CompSci and being super-territorial. I also heard stories of CompSci having split off from mathematics a few decades before that, and during my Interaction Design (IxD) master I learned that it separated from HCI in the nineties (short version: HCI was idolising the methods of the quantitative sciences too much for its own design-oriented good, IxD decided to bring in methods and insights from the qualitative fields). I wonder if this was better or worse in Doyle's time.

foxyv 8y ago

I think a significant portion of my brain (Similar in size to the knowledge from my Bachelor's degree) was dedicated to remembering every quest and map up to WotLK. It's surprising how much I still remember from 7 years ago.

vanderZwan 8y ago
Funny enough, that might also reveal something about how human brains store information: locations on a map sounds suspiciously compatible with the method of loci (aka "memory palace", thanks to a more recent incarnation of Sherlock Holmes) [0] https://en.wikipedia.org/wiki/Method_of_loci https://en.wikipedia.org/wiki/Method_of_loci
- Retra 8y ago
  Music is also very easy to remember.
  - vanderZwan 8y ago
    IIRC there is recent research confirming it is at least partially stored in the same are responsible for muscle memory, which supposedly explains quite a few things.
    - ZeikJT 8y ago
      If that is true, the obvious question is why is it stored there?
      - vanderZwan 8y ago
        Well, there are strong ties to movement in music. There is a reason that you don't dance to paintings, literature or poetry. So one option could be that music arises partly from our motor functions to begin with.
        vram22 8y ago
        >There is a reason that you don't dance to paintings, literature or poetry. Ha ha, well put. Although poetry is sort of like music, a bit, at least rhyming poems are. Just looked it up, this article is interesting: https://en.wikipedia.org/wiki/Rhyme https://en.wikipedia.org/wiki/Rhyme

Retra 8y ago

I think there's a related problem that would have been understandable at the time: if you don't understand how the world works, you are potentially missing out on patterned information which allows one to form more accurate models of the world across domains. The assumption that's made seems to be that different problem domains are all independent, but that's not completely true. That's probably the main reason that mathematics and language are so useful.

vanderZwan 8y ago
> The assumption that's made seems to be that different problem domains are all independent, but that's not completely true. That's probably the main reason that mathematics and language are so useful. To be fair, the effectiveness of multi-disciplinary research seems to only have been acknowledged in recent decades, even though by and large the biggest breakthroughs have almost always been thanks to people looking outside of the silos of their field. The sciences are really "tribal", and it used to be much worse. For example, when I started studying in the early 2000s I remember the ML bachelor (then: AI) having recently split off from CompSci and being super-territorial. I also heard stories of CompSci having split off from mathematics a few decades before that, and during my Interaction Design (IxD) master I learned that it separated from HCI in the nineties (short version: HCI was idolising the methods of the quantitative sciences too much for its own design-oriented good, IxD decided to bring in methods and insights from the qualitative fields). I wonder if this was better or worse in Doyle's time.