[Reading] Tell Me a Story, Chapter 1: Knowledge is Stories

I'm so excited about this book. Thanks Henry for recommending and lending me this. Here are some of the sentences I found very inspiring in Chapter 1:

1. "A good teacher is not one who explains things correctly but one who couches explanations in a memorable (i.e., interesting) format."
2. "Knowledge, then, is experiences and stories, and intelligence is the apt use of experience and the creation and telling of stories."
3. "A good theory of mind must include theories about how the stories of others are decoded to find in indices to enable their retrieval and storage, as well as theories of how and why our own stories appear in our minds in response."
4. "performance in conversation is an excellent measure of intelligence"
5. "When someone tells you a story in response to one you have told that captures an important generalization between the two, you believe that you have been 'really understood'. "
6. "Storytelling and understanding are functionally the same thing"
7. "We are always looking for the closest possible matches. We seek to retrieve stories that seem to us to be identical"
8. "Is being very intelligence just having a great many stories to tell? Does it mean being better at finding relevant stories to tell? Is it adapting superficially irrelevant stories into relevant ones, i.e., finding a story in one domain and applying it by analogy to another? Maybe it means combining stories and making generalizations from them -- or, perhaps intelligence is embodied in the intitial procecss of collecting stories to tell in the first place."

in which that last two are particularly interesting to me.

I like the 7th one because it talks about exactly what I'm trying to do in the storytelling system - "looking for the closest possible matches; retrieving stories that seem to be identical". It's interesting that the interaction between the user and my system actually constitutes collaboratively the thinking process of someone, since what Schank's saying right here is that the story retrieving process refers to thinking. The thinking of this someone, in fact, is really something I interpret as the construction process of the narrative. During the succession of popping up stories to each other, the user and the system are really working together and to think about what's going on next, which, as a whole, eventually would turn out to be a storyline.

In theory, this process may be used as a learning enviornment of "how to think" for the AI behind the screen, because it is allowed to see how the human participant responds, how he/she chooses the recommended results, and so on. It will learn by observing which similar abstraction pairs his/her choices are according to, and maybe even learn to select different kinds of abstraction categories under different context.

And I like the 8th one because it raises a good question about what I've long taken as an undoubtful assumption - "Does it mean being better at finding relevant stories to tell?" I think right here what Schank wants to address is that the point should be the strategy of designing the mechanism of understanding - e.g. "combining stories and making generalizations from them", as opposed to one of the consequences that it leads to, which is a search result of relevant stories - "being better at finding relevant stories to tell".