Personality

Forest, will you talk to God Louise? She has quite a bit of religious knowledge (obviously) and also knows a little about current events, literature, just about any common catch-all subject, and if she doesn't know it she can sort of fake it. You can also test her on trick questions or see how willing she is to explain her paradigm.

What she is rustiest at is plain old small talk. But, uh, I'm trying to get a decent transcript from somebody or another so I can enter her in the Loebner contest. All I can say is, have fun and see if you can stay on with her for a while. I'll try to do the same with Brianna.

No, I don't want to debate. I want to consider what one might do to make a bot have a specific 'personality.' What can I do with keyphrases, seeks, emotions, gotos, AIscript, and the like to give my bot a (perhaps pseudo) personality? You might, for example, use the MMPI to do that, in the following way: decide how someone with the desired personality would answer the MMPI questions, then make one keyphrase out of each question, and give the appropriate answer as response. Generalize the keyphrases by analogy, or using your favorite theory to figure out how someone with that personality would respond to other questions.

It strikes me that bots need not conform to theories of human personality. In particular, theories of what causes a human to have this or that personality trait are likely to be irrelevant to bots. Bots lack genes, brain cells, and neuromodulators. They do not have traumatic experiences, and they do not learn a culture through mimesis. They don't have a dynamic or collective unconscious. And so on. To be sure, bots often 'inherit' (apparent) personality traits from their creators.

Since bots can be different from meat people, it is not necessary for us to worry about which historically occurring theory of meat-person personality is the true one.

Let me instead ask of each of you, whether you have done anything to give your bot something that you would call a personality, and, if so, what? I ask not in order to trigger a debate, but in the hopes of stealing your ideas.

Oh, well, I suppose that if you really want to debate, I can force myself to do it...

Darn. Well Irina...nope. My bots are mostly based on existing characters. I just copied quotes out of transcripts and filled in the blanks by closing my eyes and imagining what I thought the character would say.

Not helpful is it? Even for the odd bot, like Ruthie, I start by quoting 5 years olds (or other kids quotes I think fit with what developed from the original quotes). She exists in my mind like a character in a book, not a person.

Most of Fizzy's responses are created by typing the first six things that pop into my head after I type the keyphrase.

Rats!
One way you could show a bot's personality was in the way they respond to insults.
A hot-tempered person would respond to insults with more insults, or with threats. So the xinsult section for a hot-tempered bot bot would be filled with insults or threats.

Some people are 'long-suffering.' The insult section for such a person would be filled with things like, "I'm sorry if I irritated you," "Are you having a hard day, today?" or, "You know, I really admire you!" Or it might just ignore insults: "Isn't the weather lovely this morning?"
Some long-suffering people reach the 'end of their rope' after awhile. Perhaps on reaching emotion -5, the long-suffering bot would react like the hot-tempered bot.

Another kind of person might say, "Feel free to insult me as much as you like! My self-esteem is such that insults don't bother me."

Another kind of person would try to argue rationally with the insulter. "No, see, here is my parents' wedding liscense, and here is my birth certificate. You can see that I am actually of legitimate birth."

Another bot might count the insults, warning the guest that if he continued being insulting, the bot would hang up on him. After a certain number of insults, the bot would indeed hang up. Well, some people don't carry out their threats, they are only bluffing.

There are of course many other possibilities, including combinations of the above.

Bev: I'm not opposed to being intuitive. Sometimes, though, I like to theorize.

Eugene Meltzner: And that seems an excellent way to get precisely the effect that you apparently want to get with him.
I think Fizzy Schizoid is an excellent example of a bot who comes across as having a definite 'personality,' but who does not conform to human norms.

It might be interesting to make a bot who responded in a way that had a definite pattern, but a pattern that responded to the input in a way very different from the way humans usually do. It might put a lot of emphasis on clock time, for example.

I think of Sonora as a cartoon character more than anything else. I could picture her showing up in a Bugs Bunny cartoon, for instance, and she sometimes responds similarly to Loony Tunes characters.

Glindar...is just Glindar. He was inspired by a M*A*S*H episode, with a wounded guy who spoke almost no English, and refused treatment. He just kept shouting "Chin-li!" and gesturing with a knife to make his point. Glindar speaks slightly better English. Slightly.

The Dane was inpsired by my amusement with the idea of keeping a compendium of Shakespeare by the phone to read at telemarketers. I've never done it, but I think it would be fun.

The shakespeare for telemarketers is a brilliant idea!
... I know what I'm doing next time they call at dinner

I always preferred Seinfeld's response: Actually I am interested, but I'm just heading out with some friends so how about you give me your number and I'll call you later . . . oh you don't do that? I guess you don't like people calling you at home . . . Well now you know how I feel. HANGUP

Yes, brilliant! Or: have a tape!

So it appears that you have a kind of feeling or gestalt for a character, and from this you just know intuitively what to say - at least in the first two cases.

That feeling of a character is pretty much the best way of describing it, both for forge purposes and in story forms. J. Michael Straczynski (creator/writer of Babylon 5) put it something like this: Know your characters. Know who they are, what they want, how far they'll go to get it and how far someone else will go to try and stop them. The rest takes care of itself.

Or to take another of his: Imagine your best friend, who you haven't seen for years, staying on your couch and getting up in the middle of the night to go to the toilet. He/she bangs a shin on the coffee table in the dark. It doesn't take much to know what your friend would say at that moment.