Commonsense Reasoning 0809 HC 15: Dialogue systems for argumentation 2 Course synthesis

1
Commonsense Reasoning 08/09 HC 15Dialogue
systems for argumentation (2)Course synthesis

• Henry Prakken
• 14-01-2009

2
Prakken languages, logic, agents

• Lc Any, provided it has a reply structure
  (attacks surrenders)
• Lt any
• Logic argumentation logic
• Arguments trees of conclusive and/or defeasible
  inferences
• Defeat depends on chosen logic
• Semantics grounded
• Assumptions on agents none.

3
Prakken example Lc (with reply structure)
4
Prakken protocols (basic rules)

• Each noninitial move replies to some previous
  move of hearer
• Replying moves must be defined in Lc as a reply
  to their target
• Argument moves must respect underlying
  argumentation logic
• Termination if player to move has no legal moves
• Outcome what is dialogical status of initial
  move at termination?

5
Dialogical status of moves

• Each move in a dialogue is in or out
• A surrender is out,
• An attacker is
• in iff surrendered, else
• in iff all its attacking children are out
• (A p since Q move is surrendered if all its
  premises and it conclusion are conceded)

6
?1 claim (owe 500)
?2 why (owe 500)
?4 argue (owe 500 since contract no payment)
?6 why (contract)
?5 concede (no payment)
?8 argue (contract since notarys document
signed by us)
?11 argue (?notarys document since notarys
seal is forged)
?12 why (notarys seal is forged)
7
Functions of dialogical status

• Can determine winning
• Proponent wins iff at termination the initial
  claim is in opponent wins otherwise
• Can determine turntaking
• Turn shifts if dialogical status of initial move
  has changed
• Immediate response protocols
• Can be used in defining relevance

8
Relevant protocols

• A move must reply to a relevant target
• A target is relevant if changing its status
  changes the status of the initial move
• Turn shifts if dialogical status of initial move
  has changed
• Immediate response protocols

9
P1
O1-
P2-
P4
O2-
O3
P3
10
P1
O1-
P2-
P4
O2
O3
P3-
O4
11
P1
O1-
P2-
P4
O2-
O3
P3
12
P1-
O1
P2-
P4-
O2-
O3
O4
P3
13
?1 claim (owe 500)
?2 why (owe 500)
?4 argue (owe 500 since contract no payment)
?6 why (contract)
?5 concede (no payment)
?8 argue (contract since notarys document
signed by us)
?11 argue (?notarys document since notarys
seal is forged)
What are the relevant targets for ??
?12 why (notarys seal is forged)
14
Example persuasion dialogue
P1 My car is safe. claim P2 Since it has an
airbag. argument P3 why does that not make my
car safe? challenge P4 Yes, that is what the
newspapers say, concession but that does not
prove anything, since newspapers are unreliable
sources of technological information
undercutter P5 OK, I was wrong that my car is
safe. retraction
O1 Why is your car safe? challenge O2 That is
true, concession but your car is still not safe
counterclaim O3 Since the newspapers recently
reported on airbags exploding without cause
rebuttal O4 Still your car is not safe, since
its maximum speed is very high. alternative
rebuttal
15
Prakken example dialogue
P1 claim safe
16
Prakken example dialogue
P1 claim safe
O1 why safe
17
Prakken example dialogue
P1 claim safe
O1 why safe
P2 safe since airbag, airbag ? safe
18
Prakken example dialogue
P1 claim safe
O1 why safe
P2 safe since airbag, airbag ? safe
O2a concede airbag
19
Prakken example dialogue
P1 claim safe
O1 why safe
P2 safe since airbag, airbag ? safe
O2a concede airbag
O2b ?safe since newspaper, newspaper ?
?safe
20
Prakken example dialogue
P1 claim safe
O1 why safe
P2 safe since airbag, airbag ? safe
O2a concede airbag
O2b ?safe since newspaper, newspaper ?
?safe
P3a concede newspaper
21
Prakken example dialogue
P1 claim safe
O1 why safe
P2 safe since airbag, airbag ? safe
O2a concede airbag
O2b ?safe since newspaper, newspaper ?
?safe
P3a concede newspaper
P3b so what since unreliable, unreliable
? so what
22
Prakken example dialogue
P1 claim safe
O1 why safe
P2 safe since airbag, airbag ? safe
O2a concede airbag
O2b ?safe since newspaper, newspaper ?
?safe
O3 ?safe since high speed, high speed ?
?safe
P3a concede newspaper
P3b so what since unreliable, unreliable
? so what
23
Prakken example dialogue
P1 claim safe
O1 why safe
P2 safe since airbag, airbag ? safe
P4 retract safe
O2a concede airbag
O2b ?safe since newspaper, newspaper ?
?safe
O3 ?safe since high speed, high speed ?
?safe
P3a concede newspaper
P3b so what since unreliable, unreliable
? so what
24
Argument graph

so what
unreliable
unreliable ? so what
25
Winning and logic

• A protocol should respect the underlying logic
• We want main claim is in iff it is implied by
  the exchanged information
• (except information that is disputed and not
  defended)
• Ensured in relevant protocols (under certain
  conditions)

26
Prakkens relevant protocols characteristics

• Protocol
• Multiple-move
• Multiple-reply
• Not deterministic in Lc
• Immediate-response
• Dialogues
• Can be long (stepwise construction of arguments,
  alternative replies)
• Both sides can develop arguments
• Logic
• Used for single agent construct/attack arguments
• Used for outcome players jointly build
  dialectical graph
• Commitments
• Not used (could be used in protocol)

27
Filibustering

• Many two-party protocols allow obstructive
  behaviour
• P claim p
• O why p?
• P p since q
• O why q?
• P q since r
• O why r?
• ...

28
Possible sanctions

• Social sanctions
• I dont talk to you any more
• Shift of burden of proof by third party
• ...
• P q since r
• O why r?
• referee O, you must defend not-r!

29
Contents of Synthesis

• Theoretical issues
• Defeasibility and probability
• Intuitions and validating logics
• Dynamics
• Final project options

30
Commonsense reasoning and probabilities

• Principles of defeasible reasoning cannot be
  founded on probability theory (PT)
• PT produces mere probabilities
• Not all defeasibility is probabilistic
• Frame axioms, norms, goals, ...
• But NMLs ignore statistical dependencies between
  variables

31
Defeasibility and probability a paradox

• The lottery paradox
• A lottery with 1 million tickets and 1 prize.
• The probability that some ticket wins is 1
• The probability that a given ticket Li wins is
  0.000001.
• Is the conclusion that a given ticket will not
  win justified?

32
The lottery paradox in default logic

• Problem for no I is ?Li justified.
• Solution HP this is not a reasoning but a
  decision problem

33
Validating logics with intuitions (1) the
general case

• The problem check that a logic adequately
  formalises a reasoning practice
• A method formalise examples, and check whether
  the logic satisfies ones intuitions
• But whose intuitions of logicians, of ordinary
  language users,
• And what if intuitions conflict

34
Validating logics with intuitions (2) the
defeasible case

• A special problem how to distinguish
  counterexamples from abnormal situations?
• Hypothesis many counterexamples are based on
  new information that invalidates a defeasible
  inference.

35
An examplefloating conclusions

• People tend to live in the same cities as their
  spouses
• People tend to live in the city where they work
• Carole works in A, her spouse works in B, so they
  live in A or B.

• Hortys counterexample Carole works in A, her
  spouse John works in B, so they live in U
  (between A and B)
• But isnt this an exceptional situation?

36
Which semantics is the right one?

• Alternative semantics may each have their use in
  certain context
• E.g. epistemic vs. practical reasoning
• Dynamic aspects of reasoning makes this problem
  less urgent

37
Floating conclusionsstill invalid? (Horty)

• Witness John says the suspect shot the victim to
  death
• If a witness says P then usually P is the case
• So, the suspect shot the victim to death
• So, the suspect killed the victim
• Witness Bob says the suspect stabbed the victim
  to death
• If a witness says P then usually P is the case
• So, the suspect stabbed the victim to death
• So, the suspect killed the victim

38
Floating conclusionsDont ignore dynamics

• Any judge would ask further questions
• Did you hear anything?
• Where did you stand?
• How dark was it?
• The laws way of dealing with dynamics
• Procedures for fair and effective dispute
  resolution

39
A simpler (imaginary) example

• American civil law evidence has to prove claim
  on the balance of probabilities
• (Imaginary) statistic 51 of American husbands
  commits adultery within 10 years.
• Mary has been married to John for 10 years can
  she sue John for divorce?

40
Topics for final Msc project

• Apply a logic to a certain problem
• Medical, legal,
• Implement
• Argument games, argument structuring, dialogue
  protocols, dialogical agents
• Theoretical research
• Defeasible vs. probabilistic reasoning
• Combining different modes of reasoning
• Dynamics
• Intelligent agents and argumentation
• Dialogue protocols and strategies/attitudes
• Learn optimal strategies
• Argumentation in negotiation or deliberation