ICPW2007.Hoffman

17
Logical Argument Mapping (LAM): A cognitive-change-based method for building common ground Michael H.G. Hoffmann October 27, 2007 [email protected]

Transcript of ICPW2007.Hoffman

Page 1: ICPW2007.Hoffman

Logical Argument Mapping (LAM):A cognitive-change-based method for buildingcommon ground

Michael H.G. Hoffmann

October 27, [email protected]

Page 2: ICPW2007.Hoffman

Outline

• Argument visualization: State of the art• Some definitions• Specific differences of Logical Argument Mapping (LAM)• The goal: Building common ground

through cognitive change in four areas• Cognitive change and Peirce’s concepts of

diagrammatic reasoning and pragmatism• The compulsory power of diagrams• Requirements for cognitive-change-based argument visualization tools• The normative standard of LAM: Three rules• The LAM procedure• The essential ideas behind LAM• Analysis of an exemplary argument• Conventions for constructing LAM arguments• Conclusion• References

[email protected]

Page 3: ICPW2007.Hoffman

Argument visualization: State of the art in three areas

Argumentation in a broader sense (focus on clarifying issues,sensemaking, problem solving, collaborative learning):

• Belvedere: Dan Suthers• Compendium, ClaiMapper: Simon Buckingham Shum• Dialog mapping: Conklin, 2006Argumentation in a narrow sense:• Toulmin, 2003 <1958>• Wigmore Diagrams (1931): Rowe & Reed, 2006• Carneades: Gordon, Prakken, & Walton, 2007• Rationale: van Gelder, 2007• Araucaria: Reed & Rowe, 2004Systems to translate various argumentation styles• Argument Interchange Format (AIF): Chesnevar et. al., 2006• World Wide Argument Web (WWAW): Rahwan, Zablith, &

Reed, 2007

Page 4: ICPW2007.Hoffman

Some definitions

• Argument:An instantiation of an argument scheme. The generalform of an argument scheme is always that of relatingat least one reason to a claim (various lists of argumentschemes can be found in Walton, 1995; Pollock, 1995; Katzav & Reed,2004; hist. overview: Garrson, 2001).

• Logical Argument:An argument whose argument scheme is a valid ruleof inference (modus ponens, complete induction, etc.)

• Argumentation:A set of arguments and statements that support,object to, or evaluate elements of those arguments

Page 5: ICPW2007.Hoffman

Specific differences of Logical Argument Mapping (LAM)

• Uses primarily logically valid argument schemes• Main function: To induce cognitive change• Central idea: Logical inference forms establish a

normative standard for arguments. In her attempts tomeet this standard, the user is challenged to enter akind of dialectical process that leads her back andforth between improving her own understanding of theissue in question and the way she represents it

• That means: LAM is more an interventional than adescriptive tool

• Following Thomas Aquinas: Before you attack anargument, make it as strong as possible

• The focus is on representing subjective (andintersubjective) perspectives, not on an objectivistreconstruction of some “truth.” Since everybodyframes a problem or conflict differently, the“authorship” of an argument is important

Page 6: ICPW2007.Hoffman

The goal: Building common ground through cognitivechange in four areas

1. Facilitated conflict negotiations2. Deliberative decision making

In (1.) and (2.) LAM can be used to deepen mutualunderstanding and to stimulate cognitive change in caseswhere mutual understanding is a central problem

3. Analysis of texts and narrativesLAM can help the analyst to find common ground betweenher interpretation and the material’s rationality

4. Intercultural communicationThere is some hope that through an interculturaldevelopment of LAM argument schemes a sort of universalargument language can be formed

Page 7: ICPW2007.Hoffman

Cognitive change and Peirce’s concepts ofdiagrammatic reasoning and pragmatism

• Diagrammatic reasoning: by externalizing ourreasoning in diagrams, we create “something(non-ego) that stands up against ourconsciousness. … reasoning unfolds when weinhibit the active side of our consciousness andallow things to act on us” (Hull, 1994)

• “Diagrams” are those “icons” that areconstructed by means of a certain“representational system” (Peirce, CP 4.418)

• E.g. an axiomatic system: A system of axiomsdoes not only define the representational meansthat are available in a field, but it determines alsothe necessary outcome of any operation orexperimentation we perform within such asystem.

Page 8: ICPW2007.Hoffman

The compulsory power of diagrams

Kant’s construction toprove that the sum of thetriangle’s inner anglesequals 180° degrees

It is the ontology (elementsand relations) and the rules ofthe chosen system ofrepresentation that determineswhich experiments withdiagrams are possible, andtheir necessary outcome.

For Peirce, this is thefoundation of his pragmatism:It is a “practical consideration”that “if one exerts certainkinds of volition, one willundergo in return certaincompulsory perceptions. …certain lines of conduct willentail certain kinds ofinevitable experiences”(CP 5.9).

Page 9: ICPW2007.Hoffman

Requirements for cognitive-change-based argumentvisualization tools

1. Since a diagram is the more “compelling” the strongerthe rules of the representational system, and the betterwe understand and realize these rules, we need, first, astandard of argumentation that is as strong aspossible and, second, the readiness of people topursue the goal of meeting this standard as strictly aspossible.

2. Whatever is relevant for the possibility of cognitivechange, or what might have an impact on theacceptability of an argument, must be visible

3. To reduce cognitive load, only what is relevant shouldbe visible

4. To allow the integration into the World Wide ArgumentWeb (WWAW) proposed by Rahwan, Zablith, & Reed(2007), each element of an argumentation should betagged using the ontology of the ArgumentInterchange Format (AIF)

Page 10: ICPW2007.Hoffman

The normative standard of LAM: Three rules

1. Structure your map according to anargument scheme whose logicalvalidity is evident and generallyaccepted

2. make sure that all your premises(reasons and warrants) are true, andprovide further arguments for theirtruth if they are not evident

3. make sure that all your premises areconsistent with each other

Page 11: ICPW2007.Hoffman

The LAM procedure

1. Identify the logical argument scheme thatrepresents best what you try to map as anargument

2. Transform what you identified as an argument intoa logical argument by adding what is missing, andby reformulating the elements of the argument in away that its validity in accordance with the schemebecomes evident

3. Consider possible objections against both thereason and the warrant. (At this point, thecompelling character of LAM as a representationalsystem plays out. Since we are challenged toexplicate everything that is needed to get alogically valid argument, we can see exactly wherethe argument can be weakened)

4. Decide whether to develop new arguments againstthe objections, or to reformulate it in a way that itcan be defended against the objections, or to giveup the whole argument

Page 12: ICPW2007.Hoffman

The essential ideas behind LAM

• The normative standard of the three rules challenges the LAMuser to explicate everything that is necessary to get a logicalargument map, and to refine her or his map as long as it takes tomeet this standard

• This means1. that all those implicit background assumptions that determine how we frame

an issue—and that are mostly responsible for problems of mutualunderstanding—become visible and an object of reflection

2. that all the parts of an argument—not only what someone explicitlymentions—are on the table and can be questioned so that a process ofbuilding common ground will be motivated

• Visualizing what hinders most in building common ground isessential for cognitive change

• From an epistemological point of view, the truth of premises inarguments is either evident or has to be justified in an ongoingprocess of argumentation. Thus, Logical Argument Mapping leadseither to assumptions that can be accepted as socially shared, orto a certain modesty regarding truth claims

• Whatever the outcome might be, it is a process that we engage inwhen mapping the logical structure of an argument.

Page 13: ICPW2007.Hoffman

Analysis of two exemplary arguments

The map1.

2. Analysis of anargument aboutthe importanceof jihad(877 KB)

Page 14: ICPW2007.Hoffman

Conventions for constructing LAM arguments

Layout• The structure of a LAM map is determined by Western reading habits that direct

our attention from the top left corner of a page to the right and downwards• Since the understanding of an argument is facilitated when we know the central

claim from the very beginning, this claim is located on top of the map in the leftcorner

• Starting from there, we work to the right and downwards to reconstruct thereasons and warrants in an ongoing process of argumentation

Ontology: statements and relations• Statements are presented in two different text box forms: rounded rectangles and

ovals. Based on their importance for cognitive change, the warrants arehighlighted by using oval text boxes; everything else is presented in roundedrectangles

• The ground color specifies a coherent position, all statements in this color mustbe consistent; objections and other considerations are presented in differentcolors

• Relations are represented by arrows. Each arrow must be specified by1. Its function: “therefore” for arguments; “opposes,” “refutes,” “rejects,” “questions,”

“supports,” etc. for other functions2. By naming the chosen logical argument scheme (S-R: rule of inference scheme) or

a conflict scheme (S-C)3. By naming the person/group/institution that claims this relation (AU: author)

Page 15: ICPW2007.Hoffman

Conclusion

• The purpose of Logical Argument Mapping (LAM) isto facilitate processes of building common groundin three areas: Conflict negotiations Deliberative decision making Analysis of texts and narratives Intercultural communication

• Its main objective is to motivate cognitive change• If cognitive change is the goal, then more important

than finding the truth with regard to an issue is topromote self-reflexivity: revealing implicitassumptions and motivating both insight into one’sown limitations and an ongoing process ofreframing

Page 16: ICPW2007.Hoffman

LAM: Older examples on the web

• Searching for common ground onHamas (March 31, 2007; 279 KB)http://cmapspublic2.ihmc.us/servlet/SBReadResourceServlet?rid=1175354427380_673614899_4820&partName=htmltext

• Hume on causality(March 12, 2007; 2.0 MB!)http://cmapspublic2.ihmc.us/servlet/SBReadResourceServlet?rid=1174485108126_1315080200_6415&partName=htmltext

• Regulating kidney supply(Feb 27, 2007; 618 KB)http://cmapspublic2.ihmc.us/servlet/SBReadResourceServlet?rid=1172634181185_1938633584_8077&partName=htmltext

• Middle East conflict. An Argumentationon the sovereignty over al-Haram al-Sharif/Temple Mount in Jerusalem(May 30, 2006; 763 KB)http://cmapspublic2.ihmc.us/servlet/SBReadResourceServlet?rid=1174484935398_1054902877_6189&partName=htmltext

Page 17: ICPW2007.Hoffman

References

Chesnevar, C., McGinnis, J., Modgil, S., Rahwan, I., Reed, C., Simari, G., et al. (2006). Towards an argument interchange format. Knowledge EngineeringReview, 21(4), 293-316.

Conklin, J. (2006). Dialogue Mapping: Building Shared Understanding of Wicked Problems. Chichester, England; Hoboken, NJ: John Wiley & Sons.Garrson, B. (2001). Argument Schemes. In F. H. v. Eemeren (Ed.), Critical concepts in argumentation theory (pp. 81-100). Amsterdam: Amsterdam University

Press.Gordon, T. F., Prakken, H., & Walton, D. (2007). The Carneades model of argument and burden of proof. Artificial Intelligence, 171(10-15), 875-896.Hoffmann, M. H. G. (2004). How to Get It. Diagrammatic Reasoning as a Tool of Knowledge Development and its Pragmatic Dimension. Foundations of Science,

9(3), 285-305.—— (2005). Logical argument mapping: A method for overcoming cognitive problems of conflict management. International Journal of Conflict Management,

16(4), 305–335.—— (in press). Cognitive conditions of diagrammatic reasoning. Semiotica (special issue on "Peircean diagrammatical logic," ed. by J. Queiroz and F.

Stjernfelt).Hull, K. (1994). Why Hanker After Logic? Mathematical Imagination, Creativity and Perception in Peirce's Systematic Philosophy. Transactions of the Charles S.

Peirce Society, 30, 271–295.Katzav, J., & Reed, C. A. (2004). On Argumentation Schemes and the Natural Classification of Arguments. Argumentation, 18(2), 239 - 259.Kirschner, P. A., Shum, S. J. B., & Carr, C. S. (Eds.). (2003). Visualizing Argumentation: Software Tools for Collaborative and Educational Sense-making.

London: Springer.Peirce. (CP). Collected Papers of Charles Sanders Peirce. Cambridge, Mass.: Harvard UP.Pollock, J. L. (1995). Cognitive carpentry. A blueprint for how to build a person. Cambridge, Mass. : MIT Press.Rahwan, I., Zablith, F., & Reed, C. (2007). Laying the foundations for a World Wide Argument Web. Artificial Intelligence, 171(10-15), 897-921.Reed, C. A., & Rowe, G. W. A. (2004). Araucaria: Software for Argument Analysis, Diagramming and Representation. International Journal of AI Tools, 14(3-4),

961-980.Rowe, G. W. A., & Reed, C. A. (2006). Translating Wigmore Diagrams [Electronic Version]. Retrieved Oct. 18, 2007, from

http://babbage.computing.dundee.ac.uk/chris/publications/2006/comma2006-wig.pdfToulmin, S. E. (2003 <1958>). The Layout of Arguments. In The uses of argument (Updated ed., pp. 87-134). Cambridge, U.K.; New York: Cambridge University

Press.van Gelder, T. J. (2007). Rationale: Making People Smarter Through Argument Mapping [Electronic Version]. Law, Probability and Risk, submitted, from

http://www.austhink.com/pdf/vangelder_submitted.pdfWalton, D. (1995). Argumentation Schemes for Presumptive Reasoning: Lawrence Erlbaum.Wigmore, J. H. (1931). The Principles of Judicial Proof (2nd ed.): Little, Brown & Co.