AI Zone Admin Forum Add your forum

NEWS: Chatbots.org survey on 3000 US and UK consumers shows it is time for chatbot integration in customer service!read more..

Writing rules to human conversation
 
 

Here is an attempt to write how some rules for conversation. I tried to do how a greeting and a farewell would happen but didn’t write examples for the farewell.

What do you think? Do you think that you could write general rules for human conversation, in a programmatic way (e.g. in the form of if statements/a flowchart)? Or are humans too complex. Either way this is just an attempt.

Criticize/add to it/post whatever.

* The back and forth process used in a conversation:
Telling ->
Asking ->
Listening <-
Thinking—

* Elements to a conversation:
opening
body
closing

* Flow of a conversation:
1. Opening
——————————
Greeting:
If acquainted
Formal or informal greeting. It has happiness to see the other person again.
Elseif
Formal greeting
Endif

If acquainted
Inquiry into wellbeing since last seeing person
Elseif
Inquiry into wellbeing for the present
Endif
——————————

2. Body
——————————
——————————

3. Closing
——————————
Farewell:
Question about meeting further
——————————


* Details of the parts
——————————
Greeting
Explanation:

A greeting can be a single word/greeting phrase on its own, or it can be appended with an inquiry into:
1) present action (e.g. ‘what are you doing?’)
2) state of wellbeing (e.g. ‘how is your day?’)
2.1) if acquaintance, since last time met (e.g. ‘how have you been?’)
——————————

Examples:
Statement greetings:
TO STRANGER OR ACQUAINTANCE:

Hello
Hi
Hey
What’s up/Sup
Nice to meet you

ONLY TO ACQUAINTANCE:
Nice to see you.
Long time no see.
It’s been a while.

Asking a question about wellbeing:
TO STRANGER OR ACQUAINTANCE:

How do you do?
how is your day? how has your day been?
what’s up?

ONLY TO ACQUAINTANCE:
how are you? how have you been?
how are you doing?
what’s been up?
What have you been up to?
How are things?
What’s new?
how are you going?


——————————
Farewell
Explanation:

A farewell can or cannot have a question about meeting further, based on:
whether the people are acquainted
if so, how well they know each other
whether (in the conversation) they planned to meet further
——————————

 

 
  [ # 1 ]

Beginning and end…now you have to do the tough middle part wink

 

 
  [ # 2 ]

Indeed, the middle part (flow of the conversation) is where the real trouble begins. In a real human conversation, there is no way to predict where the conversation will lead, people tend to sidetrack, go off on tangents, etc. in conversations. The conversational system needs to be able to track this and react to such situations. In my perspective this is where strong-AI comes in, because without the system ‘understanding’ what’s happening in the conversation (why the sidetrack or tangent occurs), it is impossible to either follow up on the new direction or steer the conversation back on track.

Mind you, the problem I describe is is pretty hard already with only humans conversing wink

 

 
  login or register to react