1 / 40

How to change somebody’s mind? Modeling a Process of Convincing.

How to change somebody’s mind? Modeling a Process of Convincing. Katarzyna Budzyńska Faculty of Christian Philosophy Cardinal Stefan Wyszyński University in Warsaw & Magdalena Kacprzak Faculty of Computer Science Bialystok University of Technology.

lacey
Download Presentation

How to change somebody’s mind? Modeling a Process of Convincing.

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. How to change somebody’s mind? Modeling a Process of Convincing. Katarzyna Budzyńska Faculty of Christian Philosophy Cardinal Stefan Wyszyński University in Warsaw & Magdalena Kacprzak Faculty of Computer Science Bialystok University of Technology

  2. The Project (Budz-Kacp): The Formal Theory of Persuasion motivation Philosophy Reflection on Persuasion Process application Comp-sci The Formal Theory of Persuasion modeling Simulation of Persuasion Process application Logic modeling Formal Description of Persuasion Process

  3. The Project (Budz-Kacp): The Formal Theory of Persuasion motivation Philosophy Reflection on Persuasion Process application Comp-sci The Formal Theory of Persuasion modeling Simulation of Persuasion Process application Logic modeling Formal Description of Persuasion Process

  4. Outline of the Presentation • The Nature of Persuasion • The Aspects of Persuasion • Subjectivism • Gradation of Beliefs • Dynamism • The Logic for Persuasion Theory • The Logic of Graded Modalities • The Algorythmic Logic • Conclusions

  5. 1. The Nature of Persuasion

  6. How we change one another's minds • Persuasion (lat. persuasio) is a way to induce somebody tobelieve in our rights or to do something. • Persuasion is one of the methods of negotiation which allows to reach an agreement.

  7. Specific techniques used topersuade people Techniques for changing minds (verbal versus non-verbal arguments) • By appeal to reason: • logical arguments, scientific methods, proofs • By appeal to emotion: • body language, tradition, faith, deception, praise • Aids to persuasion: • bribery, blackmail, seduction, brainwashing, torture

  8. How we change one another's minds Thus, we assume that: • a persuasion is an action which is initiated by the conflict of opinion& aimed to change beliefs • aproponentperforms a sequence of actions to meet his goal, • an opponent comes into a conflict with a proponent • an audience is a target of persuasion • the goal of a persuasion may be satisfied in a different degree

  9. An Example • I would like to propose you the insurance policy. • I am not interested - it's beyond my pocket. Maybe it is very profitable, but I can't afford such a luxury. • But we offer the lowest premium. Moreover, our products are matched to the individual needs of our clients. I am sure we will find something that best meet your needs like for example low monthly payments. • It sounds very interesting. Tell me more - what are the details of your proposal?

  10. 2. The Aspects of Persuasion:Subjectivism

  11. Philosophical Intuitions • Persuasion refers to BELIEFS not knowledge • „conflict of opinion” not „conflict of knowledge” • knowledge – unquestionable beliefs – controversial, subjective • episteme () - true knowledge basing on rational, fundamental principles doxa () - common belief originating in sensual experiences and thus being uncertain information that may turn out to be false

  12. Beliefs: the standard approach Let M= (S,RB1,...,RBn,v) be a model where • S is a set of states, • RBiSSis an accessibility relation defined for agent i (i=1,...,n), • v : S{0,1}PV is a valuation function.

  13. Beliefs: standard approach In a standard doxastic logic it is possible to express three types of belief attitudes: • Bi • an agent i is absolutely sure that , • Mi • an agent i allows  to be true (MBi()), • Ni • an agent i is neutral with respect to logicalvalue of (NBiBi()).

  14. Beliefs: standard approach • M,s |= Bi iff for every state s’ which is i-accessible from s, M,s’ |=   RBi s  RBi RBi 

  15. 2. The Aspects of Persuasion:Gradation of Beliefs

  16. after after before before I did not believe the thesis I did not believe the thesis I do believe the thesis stronger, but not absolutely arguments arguments I do believe the thesis Philosophical Intuitions • Persuasion changes the GRADES of beliefs • „Black-and-white” types of convincing • Persuasions that increase the degree of certainty in not a fully range

  17. Beliefs: Hoek-Meyer approach • M,s |= Md iff there are more than d accessible states verifying   s   M,s |= M1

  18. Beliefs: Hoek-Meyer approach • Bd - at most d accessible states refute  (Bd  Md ) • M!d - exactly d accessible states satisfy  (M!0B0, M!dMd-1Md, if d >0)  s   B1& M!2

  19. Graded beliefs:some extensions • M,s|= Mid iff there are more than d i-accessiblestates verifying  • New deriverable modality: • Mid1,d2  Mi!d1  Mi!d2 • there are d1 i-accessible states verifying  & d2 i-accessible states verifying   • agent i beliefs  with the degree d1/(d1+d2) • observe that d1/(d1+d2)  [0,1]

  20. The Example w.r.t. Gradation w1 s1 too exp high v(s1, T) = 0 • Thesis T: I should buy a life insurance • M,s |=Mi1,2T • The housewife beliefs T with the degree 1/3 s2 exp high v(s2, T) = 1 s s3 ? exp low v(s3, T) = 0 s4 cheap low v(s4, T) = 1 s5 cheap high v(s5, T) = 1 The housewife beliefs about insurance (before persuasion)

  21. The example – some modification w1 - modified s1 too exp high v(s1, T) = 0 • M,s |=Mi1,2T • The housewife beliefs T with the degree 1/3 • M,s |=Mj3,1T • The businesswoman beliefs T with the degree 3/4 s2 exp high v(s2, T) = 1 s s3 ? exp low v(s3, T) = 0 s4 cheap low v(s4, T) = 1 s5 cheap high v(s5, T) = 1 Ri – accessibility relation of the housewife Rj – accessibilityrelation of the businesswoman

  22. 2. The Aspects of Persuasion:Dynamism

  23. Philosophical Intuitions • The persuader uses arguments as tools in order to CHANGE other’s beliefs • Argument – action aimed at changing beliefs • It requires performing some activity: • the persuader expresses verbal argument or • she executes nonverbal argument

  24. The example w.r.t. Actions w1: before w2: after s1 s1 too exp high too exp high v(s1, T) = 0 v(s1, T) = 0 s2 s2 exp high exp high v(s2, T) = 1 v(s2, T) = 1 s s’ s3 s3 argument ? ? exp low exp low v(s3, T) = 0 v(s3, T) = 0 s4 s4 cheap low cheap low v(s4, T) = 1 v(s4, T) = 1 s5 s5 cheap high cheap high v(s5, T) = 1 v(s5, T) = 1 M,s |= Mi1,2T M,s’ |= Mi3,1T She beliefs T with the degree 3/4 She beliefs T with the degree 1/3

  25. Arguments - interpretation Let M= (S,Ra1,..., Rak,v) be a model where • S is a set of states, • RatS{1,...,n}Sis an interpretation of action t, (t=1,...,k), • v : S{0,1}PV is a valuation function.

  26. a1 a2 ak ...... s s’ Arguments - interpretation • Let • 0 be a set of atomic actions, • P=a1;a2;a3;...;ak where a1,...,ak0be a program. • We say that s’ is reachable from s by agent i via program P iff there exists a sequence of states s0,...,sk such that s0=s, sk =s’ and for every t=1,...,k (st-1,i,st)Rat. • When M,s’|=Mj1,0T then we say that M,s |=(i:P) (Mj1,0T) (s,i,s’)  Rat(P)

  27. 3. The Logic for Persuasion Theory

  28. What a logic do we need to reason about persuasion? • Beliefs rather than knowledge (doxastic logic versus epistemic logic) • Graded beliefs (two-valued logic, multi-valued logic, fuzzy logic, probabilistic logic etc.) • Change of beliefs (dynamic, algorithmic logic, logic of belief revision)

  29. What a logic do we need to reason about persuasion?

  30. Characteristics of our Persuasion Theory • broad sense of notion of persuasion • gradation of beliefs-attitudes • the specific aspect of research on the persuasion process

  31. What a logic do we need to reason about persuasion? • Multimodal logic of actions & graded beliefs (AGn) • Inspired by: • The Logic of Graded Modalities (W. van der Hoek & J.J. Meyer in: Modalities for Reasoning about Knowledge and Quantities. Elinkwijk, Utrecht, 1992) • The Algorythmic Logic (G. Mirkowska & A. Salwicki. Algorithmic Logic. Polish Scientific Publishers, Warszawa, 1987)

  32. Syntax of AGn •  ::= p |  |  | Mid |(i:P) where iAgt,P, dN

  33. Model for AGn Let M=(S,RB1,..., RBn,Ra1, ... , Ram,v) with • a non empty set of states S, • doxastic relations RB1, ... , RBn, RBi S  S for i=1,...,n, • argument relations Ra1, ... , Ram, Raj S  Agt  S for j=1,...,m, • a valuation function, v: S{0,1}PV

  34. Semantics of AGn • M,s |= Mid1 iff the number of states reachable via relation RBi which satisfy  is MORE THANd1 • M,s |= Bid1 iff the number of states reachable via relation RBi which satisfy  is AT MOSTd1 • M,s |= Mid1,d2 iff • the number of states reachable via relation RBi which satisfy  is d1and • the number of states reachable via relation RBi which satisfy  is d2

  35. Semantics of AGn • M,s |= (i:P) iff there exists s’S which is reachable by agent i via program P and satisfies  ( there exists s’S such that (s,i,s’)R(P) and M,s’|= )

  36. Axioms for graded beliefs • All propositional tautologies • Bi0() (Bid Bid) • Bid Bid+1 • Bi0()  [(M!id1 M!id2)  M!id1+d2()] • Bid Bi0Bid • Bi0 (true)

  37. Axioms for actions • (i:P1;P2)(i:P1)((i:P2)) • (i:P)() (i:P) (i:P) • (i:P)(true) • (i:Id) 

  38. Rules of inference • If  and  then  (modus ponens) • If  then Bi0 (generalisation) • If  then (i:P) (i:P)

  39. Conclusions • A formalism for • simulations of human discussions • an automatic process of convincing • Wemay use it to research • whether the different scenarios of persuasion (various arrangement of arguments)will arrive at the same result, • how they will change the audience's outlook on life, • whethermost of a given group become convinced to the thesis, • how entering of a new proponentwill affect the course of discussion etc.

  40. Thank you for your attention!

More Related