Can a debating algorithm shed any light on the nature of debate?

Have philosophical discussions about politics, law, and government.
Featured Article: Definition of Freedom - What Freedom Means to Me
Post Reply
Steve3007
Posts: 5536
Joined: June 15th, 2011, 5:53 pm
Favorite Philosopher: Eratosthenes
Location: UK

Can a debating algorithm shed any light on the nature of debate?

Post by Steve3007 » June 25th, 2018, 6:48 am

Recently, it's been reported, a computer "won" a political debate against a human being:

https://www.cnet.com/news/an-ibm-comput ... mpetition/

Can this kind of experiment help us to analyse what it is that we are doing when we argue the case for a particular political or ethical position? A large part of the process must surely involve defining a goal and then explaining, using empirical evidence and logic, why a particular proposed set of activities are more likely to achieve that goal than some other set of activities. That aspect of debate might be what we would expect a machine to be best at. But anyone who has observed and/or taken part in political debate will know that it is certainly not the whole story. In many cases, it barely features at all. An audience of human beings doesn't just respond to rational evidence-based arguments. Debaters use a huge range of devices including fear, self-interest, tribal loyalty, patriotism, xenophobia, body language, charisma, bluster, showmanship, intimidation, repetition, catchy sound bites, rhymes, alliteration, music, humour, anger, hyperbole, lies, damned lies and statistics to achieve their ends.

Could a computer debater ever learn to use all of these devices? Or would it always simply be a mouthpiece for its computer debater creator?

kordofany
Posts: 44
Joined: April 9th, 2018, 3:57 pm
Favorite Philosopher: Franz Kafka
Contact:

Re: Can a debating algorithm shed any light on the nature of debate?

Post by kordofany » June 25th, 2018, 4:08 pm

It is amazing and very wondrous, but I suspect that it is capable of analyzing non-physical issues, or of associating the subjective with the object as the human brain does. If this happens, we will witness an age when we all become craftsmen and the end of academic work. Because we will only implement the crucial decisions of robots.
Freedom is an Idea in the mind... And reckless behavior is irrational.. I can live free inside the prison.
Image

User avatar
ThomasHobbes
Posts: 1122
Joined: May 5th, 2018, 5:53 pm

Re: Can a debating algorithm shed any light on the nature of debate?

Post by ThomasHobbes » June 25th, 2018, 5:43 pm

Steve3007 wrote:
June 25th, 2018, 6:48 am
Recently, it's been reported, a computer "won" a political debate against a human being:

https://www.cnet.com/news/an-ibm-comput ... mpetition/
"NEWS" items like this are for the hard-of-thinking.
A computer can have no interest in politics.
All important political debates are not 'winnable' by logic.

"A good leader can engage in a debate frankly and thoroughly, knowing that at the end he and the other side must be closer, and thus emerge stronger. You don't have that idea when you are arrogant, superficial, and uninformed." Nelson Mandela

Politics is not cheap point scoring as the stakes are human values.

Steve3007
Posts: 5536
Joined: June 15th, 2011, 5:53 pm
Favorite Philosopher: Eratosthenes
Location: UK

Re: Can a debating algorithm shed any light on the nature of debate?

Post by Steve3007 » June 26th, 2018, 5:16 am

ThomasHobbes wrote:All important political debates are not 'winnable' by logic.
Nothing can be resolved by logic alone except the definition of terminology. But what about a logical/rational argument which appeals to various pieces of empirical evidence? I think political/ethical difference between different people often boil down to disagreements about what is empirically verified as true and what logically follows from what. i.e. they are often not fundamentally irreconcilable differences of worldview, but differences of method.

Take, for example, one of the most basic and venerable divisions in political opinion: Left versus Right; big government versus small government; high taxation versus low taxation; publicly funded services versus private enterprise.

In many cases, the two sides of this debate do not fundamentally disagree as to what is important in life. They do not disagree about the underlying goal, even though they may think they do, because they disagree about relatively superficial goals. They disagree as to what actions are most likely to set in chain the sequence of events most likely to achieve that underlying goal.

User avatar
ThomasHobbes
Posts: 1122
Joined: May 5th, 2018, 5:53 pm

Re: Can a debating algorithm shed any light on the nature of debate?

Post by ThomasHobbes » June 26th, 2018, 2:03 pm

Steve3007 wrote:
June 26th, 2018, 5:16 am
ThomasHobbes wrote:All important political debates are not 'winnable' by logic.
Nothing can be resolved by logic alone except the definition of terminology. But what about a logical/rational argument which appeals to various pieces of empirical evidence? I think political/ethical difference between different people often boil down to disagreements about what is empirically verified as true and what logically follows from what. i.e. they are often not fundamentally irreconcilable differences of worldview, but differences of method.

Take, for example, one of the most basic and venerable divisions in political opinion: Left versus Right; big government versus small government; high taxation versus low taxation; publicly funded services versus private enterprise.

In many cases, the two sides of this debate do not fundamentally disagree as to what is important in life. They do not disagree about the underlying goal,
Yes they do. What do YOU think is that goal?
even though they may think they do, because they disagree about relatively superficial goals. They disagree as to what actions are most likely to set in chain the sequence of events most likely to achieve that underlying goal.
Throw as much empirical statistics at, say, the abortion debate, but where does it get you? All the underlying criteria are emotionally based.
It all depends on what you think is important.
And what is important is not superficial.

User avatar
-1-
Posts: 879
Joined: December 1st, 2016, 2:23 am

Re: Can a debating algorithm shed any light on the nature of debate?

Post by -1- » June 27th, 2018, 9:01 am

Steve3007 wrote:
June 25th, 2018, 6:48 am
Recently, it's been reported, a computer "won" a political debate against a human being:

https://www.cnet.com/news/an-ibm-comput ... mpetition/

Can this kind of experiment help us to analyse what it is that we are doing when we argue the case for a particular political or ethical position? A large part of the process must surely involve defining a goal and then explaining, using empirical evidence and logic, why a particular proposed set of activities are more likely to achieve that goal than some other set of activities. That aspect of debate might be what we would expect a machine to be best at. But anyone who has observed and/or taken part in political debate will know that it is certainly not the whole story. In many cases, it barely features at all. An audience of human beings doesn't just respond to rational evidence-based arguments. Debaters use a huge range of devices including fear, self-interest, tribal loyalty, patriotism, xenophobia, body language, charisma, bluster, showmanship, intimidation, repetition, catchy sound bites, rhymes, alliteration, music, humour, anger, hyperbole, lies, damned lies and statistics to achieve their ends.

Could a computer debater ever learn to use all of these devices? Or would it always simply be a mouthpiece for its computer debater creator?
Winning a debate, or rather, declaring a winner in a debate may yield different winners by different judges. A computer vs man debate may win on the logic ticket, or on the eruditeness ticket, or on the cadence and fluency of speech ticket, but not on a convincing-humans ticket.

A computer will never win an argument on the side of race equality when judged by a panel of KKK members.

A computer will never win an argument on Marxism's side when the judges are multibillionaires.

A computer will never win an argument on the side of capital punishment when judged by people on death row.

A computer will never win an argument on the side of pipeline-construction or on whale hunting when the jury comprises green party members.
This search engine is powered by Hunger, Thirst, and a desperate need to Mate.

Post Reply