Home

Carneades 3.7 User Manual

1. Impairing the freedom of contract would demote the values of innovation lt better alternatives gt Tn the crcumstances Researchers and students increasingly work in more than one Member State The Achieving the poal of making t easier for ers and Harmonizing the copyright exceptions would make it easier for researchers and students to work in Na Further restrictions on the end he introduction of more users of copyrighted materials 4 restrictions on the end users would CIEL should not be enacted create an imbalance between the ing what acts end users C E ER bt do when making use of No The rules about which uses of b The essence of cop copyrighted materials create x tion is the maintenance of a derivate works and thus require a M lance between the rights of the Figure 8 An Argument Map For statement nodes the text of the statement is shown inside the box possibly truncated if the text is too long In argument nodes the circle is filled with a plus sign if the argument is a pro argument or a minus sign for con arguments The edges links between argument nodes and statement nodes show the premises and conclusion of the argument The conclusion of the argument is the statement node pointed to by the edge with the normal arrowhead The other statement nodes linked to the argument without arrowheads are its premises Negative prem
2. 1 revealing implicit premises 2 validating whether the arguments are formally correct by instantiating accepted argumentation schemes 3 asking appropriate critical questions depending on the schemes applied 4 and determining which claims are acceptable taking into consideration the assumptions of users and their collective assessment of the relative weights of conflicting pro and con arguments The first three of these tasks can be accomplished by comparing the argument with its argumentation scheme On the argument page click on the argumenta tion scheme to view a description of the scheme Now you can check whether any of the premises listed in the scheme are missing from the argument The argument is formally valid if all the premises of the scheme are explicitly pro vided by matching premises of the argument and the conclusion of the argument matches the conclusion of the scheme Argumentation schemes define exceptions and assumptions which can be used to ask critical questions The exceptions provide reasons for not applying the argument undercutting it If an exception is true this doesn t mean that the conclusion of the argument is false but only that the argument does not provide a good reason to presume the conclusion to be true The assumptions of the scheme are implicit premises which need to be proven only if they are called into question So if you think an assumption does not hold you should make 21 D Carneade
3. 4 and determining which claims are acceptable taking into consideration the assumptions of the users and their collective assessment of the relative weights of conflicting pro and con arguments The final step of this process determining which claims are acceptable is what we mean by argument evaluation in this section Perhaps a better term for this narrower conception of argument evaluation to avoid confusion would be argument labeling since the result of this kind of argument evaluation is to label the statement and argument nodes of an argument graph as being in out or undecided where in means the statement is acceptable presumably true or the argument is applicable because all of its premises are in and it has not been undercut by an in argument out means the negation of the statement is acceptable equivalently the state ment itself is not acceptable or the argument is not applicable undecided means the statement or argument is neither in nor out Argument graphs are evaluated using the computational model of structured argument presented in Gordon Prakken and Walton 2007 The argument graph is automatically re evaluated updating the labels of the argument and statement nodes whenever the argument graph is modified This includes not only changes to the weights of the statements and arguments but also changes to the assigned proof standards and the addition or deletion of any arguments Addi
4. CUT Performing Al in 81 would have side effects which demote V or some other value 9 14 Argument from Cause to Effect id cause to effect conclusion Event E2 will occur premises minor An event E1 has occurred major Event E1 causes event E2 exceptions CQ1 An event E3 interferred with E1 9 15 Argument from Correlation to Cause id correlation to cause conclusion Event E1 causes event E2 premises major Events E1 and E2 are correlated assumptions CQ1 There exists a theory explaining the correlation between E1 and E2 exceptions CQ2 E3 causes E1 and E2 49 9 16 Argument from Sunk Costs id sunk costs conclusion Action A should be performed premises costs The costs incurred performing A thus far are C waste The sunk costs of C are too high to waste assumptions CQ1 Action A is feasible 9 17 Argument from Appearance id appearance conclusion O is an instance of class C premises minor 0 looks like a C 9 18 Argument from Ignorance id ignorance conclusion S premises major S would be known if it were true minor S is known to be true exceptions CQ1 The truth of S has not been investigated 9 19 Argument from Abduction id abduction conclusion H premises observation observed S 50 explanation Theory Ti explains S hypothesis T1 contains H as a member exceptions CQi T2 is a more coherent explanation than T1 of S 9 20 Ethotic Argument id
5. because they may be defeasible to distinguish them from the inference rules of classical logic which are all non defeasible strict A theory is modelled as a structure having the following fields header metadata e g title description about the theory Descriptions can be in multiple languages and can be arbitrarily long structured texts represented using the Markdown wiki language language The formal language of the theory a dictionary mapping symbols to terms and predicates schemes Strict and defeasible inference rules sections A sequence of sections which in turn consist of a header schemes and sub sections enabling theories to be organized hierarchically similar to the hierarchical structure of books and articles references A sequence of metadata structures for providing bibliographic information about source documents Next we complete this illustration of how to implement argumentation schemes by defining theory1 to be the following theory def theory1 make theory header make metadata title Reconstruction of Liverpool Argumentation Schemes creator Tom Gordon publisher Fraunhofer FOKUS date 2012 description en This is a reconstruction of the version of 41 the Liverpool schemes in Atkison2012a language L schemes S references Atkison2012a make metadata creator Katie Atkinson Adam Wyner and Trevor Bench Capon title Report No D5 2 Report on Prototype
6. ethotic conclusion 2 premises assertion P asserts that S is true trustworthiness P is trustworthy 9 21 Slippery Slope Argument This version of the slippery slope scheme is intended to be used together with the argument from negative consequences schema to derive the conclusion that the action should not be performed Notice that the scheme is represented by two Carneades schemes one for the base case and one for the inductive step 9 21 1 Base Case id slippery slope base case conclusion Performing action A would have negative consequences premises realization Performing A would realize event E horrible costs Event E would have horrible costs 9 21 2 Inductive Step id slippery slope inductive step conclusion Event E1 would have horrible consequences premises causation Event E1 causes E2 consequences Event E2 would haves horrible consequences 51 10 Credits e Conception and Design Tom Gordon Douglas Walton e Programming Pierre Allix Stefan Ballnat Tom Gordon Matthias Grabmair Sebastian Kaiser e Funding and Support Canadian Social Sciences and Humanities Research Council 2007 09 and 2012 2017 with Douglas Walton Standardized Transparent Representations in order to Extend Legal Accessibility Estrella FP6 IST 027655 2006 2008 Quality Platform for Open Source Software Qualipso FP6 IST 034763 2006 2010 Google Summer of Code 2008 Integrated Method for Policy maki
7. An atom formalizing the propositional content of the statement Next using this language we formally define positive and negative versions of a scheme for practical reasoning Schemes may be organized in an hierarchy of sections each section with its own metadata But since there are only four schemes in this example sections are not illustrated here def S make scheme id prasi header make metadata title Practical Reasoning Scheme conclusion should do S1 A premises make premise role circumstances statement circumstances 781 make premise role action statement results in S1 A S2 make premise role goal statement realizes S2 7 G make premise role value statement promotes 7G V make scheme 40 id pras2 header make metadata title Negative Practical Reasoning Scheme conclusion not should do S1 A premises make premise role circumstances statement circumstances 781 make premise role action statement avoids S1 A S2 make premise role goal statement realizes S2 G make premise role value statement demotes G V D Now let s us package the language and schemes together in a theory Concep tually a theory is a set of propositions But since the set may be infinite it is more convenient to represent theories intensionally as a set of axioms and inference rules We call the inference rules argumentation schemes
8. For most purposes the prepon derance of the evidence standard should suffice See the Editing Argument Graphs Chapter for details value A real number in the range 0 0 1 0 storing the output of the argument evaluation process where 0 0 means the statement is out false or presumably false 1 0 means the statement is in true or presumably true and all other values mean the arguments facts and assumptions are insufficient for drawing any conclusions about the truth or falsity of the statement undecided unknown Real numbers are used to allow experimentation with other models of argument evaluation atom An optional formal representation of the statement in predicate logic This feature is for analysts and need not interest public users main A Boolean value true or false used to indicate whether the statement is one of the main issues modeled by the argument graph Since argument graphs can contain cycles and are not limited to trees we cannot use the roots of trees for this purpose 4 3 Argument Properties id A Uniform Resource Name URN serving as a unique identifier for the argument world wide direction Pro or con 11 strict A Boolean value true or false expressing whether the conclusion of the argument is necessarily true when its premises are true strict arguments or only presumably true Nonstrict arguments are called defeasible arguments scheme The name of the argumentation sche
9. advantage of reducing the number of statements up to 50 resulting in more compact summaries of the arguments 10 4 2 Statement Properties id A Uniform Resource Name URN serving as a unique identifier for the statement world wide text A concise formulation of the statement written by the analyst who reconstructed the arguments from the source documents Paraphrases the various formulations of the statement in the sources Translations of the text in several languages may be included in the model Compare with the description property of the metadata of the statement which can be used to quote some or all of the formulations of the statement in the sources and provide translations in several languages weight A real number in range of 0 0 1 0 representing the degree to which the statement is accepted as true by users audience If the weight is 0 0 the statement is false rejected by the users If the weight is greater than 0 0 but less or equal to 0 25 the statement is assumed to be false If the weight is greater than or equal to 0 75 and less than 1 0 the statement is assumed to be true If the weight is 1 0 the statement is true accepted by the users Otherwise the truth or falsity of the statement is at issue In application scenarios with many users the weights can be collected via polls proof standard The method used to combine pro and con arguments Several proof standards are supported by the system
10. arguments with contradictory conclusions that is arguments pro and con some conclusion come into play Such arguments are called rebuttals The conflict between the rebuttals is resolved by weighing the arguments and applying proof standards See the Proof Standard of this chapter for further information The weight can be entered directly in the argument editor using the slider 6 5 6 Argument Metadata Arguments can be annotated with metadata properties using the Dublin Core metadata elements in the same way as statements See the Statement Metadata section for further instructions You can use the description field of the argument to provide whatever background information you want about the argument including quotations of formulations of the argument in source texts along with hyperlinks to the sources 34 You have the same possibilities for structuring the text of the description using the Markdown wiki language and entering translations of the description as you have for statement descriptions See the Statement Description section of this chapter for further information 6 6 Editing Arguments To edit an existing argument in an argument graph first navigate to the argument page and then click on the Edit Argument button This will open a new tab with an editor for modifying the argument This is the same form used to enter new arguments described in the Entering New Arguments section of this chapter but with t
11. atom optional and its id 17 If metadata had been provided for the statement it would be displayed in a separate tab on the page Descriptions may be provided in multiple natural languages with a tab for selecting the description in each language 5 2 2 Argument Pages Carneades Web App x Waa lt G D localhost 3 rneades projects copyright main argume i c5ac7 cb8 e942 4a6f pd Q D Q s Hi Apps C3 Cloud CJ News CJ Directories C3 Shopping CJ Projects J FOKUS D MARKOS G Politics Cj Carneades Other bookmarks Projects Project Outline Argument Statement Map Outline 9 Argument Edit Argument x Remove Argument Model Metadata Formalization if the following are true Circumstances Not all the material digitised by publishers is scanned with OCR Optical Character Recognition with the purpose of making the resulting content searchable Action Clarifying the law to allow works held in libraries for the purpose of making the resulting content searchable on the Internet would have a transformative effect on research learning and teaching Goal Realizing a transformative effect on research learning and teaching is an important social goal Presumably it is true that Yes The exceptions should be clarified to allow works held in libraries to be scanned for the purpose of making their content searchable on the Internet Properties value Unclear scheme position Figure 6 An Argument Page Argumen
12. be annotated with meta data and documentation in multiple natural languages Carneades includes an inference engine which is able to automatically apply theories to construct arguments Using a high level declarative language for representing theories and generating arguments from these theories makes it easier for domain experts to read and validate the theories Theories may organized in a hierarchical structure of sections and subsections with schemes included at any level Like schemes the theory as a whole and each of its sections can be annotated with its own meta data and natural language description These features facilitate self documenting and isomorphic modeling The source text of the schemes policies or legislation can be included within the model in the same files in such as way as to preserve the hierarchical structure of the sources This makes it easier to maintain the model as the source documents are modified since there can be a one to one correspondance between sections of the source text and sections of the model Theories are represented in Clojure source code files Clojure packages can be used to modularize combine and reuse theories The theory language is an executable knowledge representation language with its own semantics and an inference engine implementing this semantics It is not intended to be used as an interchange format for exporting and importing theories among diverse formalisms with
13. is my first Carneades project de Dies ist mein erstes Carneades Projekt The title and creation date properties should be self explanatory Their values are strings The creation date is not constrained to some specific format for dates The description is a Clojure map to allow multiple descriptions in differ ent languages In the example there are descriptions in English and German Descriptions can formatted using Markdown syntax The theory property specifies which theory to use to construct and reconstruct arguments using the rule based inference engine or interactively using the argument editor Theories are represented using Clojure data structures in Clojure files with the usual clj filename extension In the property map the values of the theory property should be the file name of the Clojure file containing the theory but without the clj file name extension The theory name is re solved relative to the theories directory of the given project or the current project being configured if no project is named In the example the theory property is default walton_schemes This refers to the theory in the de fault theories walton_schemes clj file in the projects directory The databases directory stores all the database files of the project including a database for each argument graph of they project A project may have more than one argument graph For example when using the policy an
14. not delete the conclusion or premises of the argument This can leave some statements in the argument graph being unused in any argument Warning There is no undo function so all editing and delete operations are permanent However you will be asked to confirm all delete operations and have the option of cancelling or saving editing operations 6 1 Reconstructing Arguments To reconstruct a new argument in some source text one which is not already in the argument graph follow the procedure below If instead the source text is another formulation of an argument already in the argument graph you can modify the description of the existing argument to also quote this as an additional source of the argument 1 Click on the New Argument button on the outline page of the argument graph 2 Copy the text of the argument from the source document and paste it into the description field of the Metadata tab of new argument form You can quote the text of the argument using Markdown by preceding each line with a gt symbol 3 Go the Model tab of argument editor and choose an argumentation scheme to apply from the pull down list of schemes The documentation of the selected scheme will be shown and the form will be customized with fields intialized for each of the premises of the scheme You can modify the argument however you want unconstrained by the chosen scheme for example by deleting or adding premises or renamin
15. s should do make predicate symbol should do arity 2 forms en make form positive In cirumstances As we should do 7s negative In circumstances 4s we should not do fs question In circumstances s should we do 4s results in make predicate symbol results in arity 3 forms en make form positive In circumstances s doing s would result in circumstances s negative In circumstances s doing s would not result in circumstances 4s question In circumstances 4s would doing s result in circumstances s avoids make predicate symbol avoids arity 3 38 forms en make form positive In circumstances s not doing 4s would avoid circumstances fs negative In circumstances s not doing fs would not avoid circumstances 7s question In circumstances 4s would not doing ds avoid circumstances 4s realises make predicate symbol realises arity 2 forms en make form positive Circumstances s would realize goal 7s negative Circumstances s would not realize goal s question Would circumstances s realize goal s promotes make predicate symbol promotes arity 2 forms en make form positive Achieving s would promote the value fs negative Achieving s would not promote the value fs question Would achieving s promote the value 4s demotes make predicate symbol demotes arity 2 forms
16. social goal Arguments against this statement Argument from Ignorance o Ifthe rules regulating the scanning of works held in libraries were unclear this would be known 9 Itis not known that the rules regulating the scanning of works in libraries are unclear Arguments having this statement gs a premise Figure 5 A Statement Page he proof standards available are dialectical validity preponderance of the evidence clear and convincing evidence and beyond reasonable doubt See the section on Evaluating Arguments for further details about proof standards he next section displays the content of the statement in natural language This formulation of the statement is written by the analyst or analysts who reconstructed the arguments to build the argument graph or generated from a template in a theory when arguments are constructed automatically using the Carneades inference engine Next arguments pro and con the statement are displayed as well as arguments which use this statement or its negation as a premise The premises of these arguments are also displayed This makes it possible to navigate to nearby arguments and statements in the argument graph by simply clicking on the links in these lists The bottom of the statement page displays the properties of the statement its value proof standard default value preponderance of the evidence weight whether or not it is a main issue its formal representation as a logical
17. 0 The new statement form is divided into two sections named Model and Metadata On smaller screens these two sections are shown in separate tabs labelled accordingly The model consists of the text of the statement in one or more languages along with some additional properties described in detail below The metadata consists of generic properties not specific to statements such as a title and description Be careful not to confuse the description of the statement with the text of the statement The text is a formulation of the statement whereas as the description provides further information about the statement 6 2 1 Main Issue An argument graph should have at least one main issue These are the issues which are central to the debate All the other issues are subsidiary issues only important to the extent that they are relevant for resolving one of the main issues 6 2 2 Statement Text The text property of a statement node of an argument graph is for expressing the statement concisely in natural language You should always provide such a text As for descriptions translations of the text in several languages may be included in the model and the statment may be structured using Markdown This is the text that will appear in statement boxes in argument maps and in hypertext views of the argument graph Whereas descriptions are optional and part of the metadata about the statement this text is the content rep
18. 04 Argument from Abduction 22e Ethotic Argument ar A dd een Slippery Slope Argument 20 9 21 1 Base Case 2 2 2 ee 9 21 2 Inductive Step 02 22 0000 10 Credits 11 User Manual License References 44 44 45 45 46 46 46 4T 47 47 48 48 48 48 49 49 50 50 50 50 51 51 51 51 52 52 53 1 Introduction The Carneades argumentation system is series of open source research prototypes freely available for downloading at http carneades github com The version of Carneades described in this user manual Carneades 3 7 provides web based collaborative software tools for e re constructing arguments using a rulebase of argumentation schemes e visualizing arguments in diagrams argument maps e critically evaluating arguments using a formal model of structured argu ment based on argumentation schemes and with support for variable proof standards Gordon Prakken and Walton 2007 The next chapter Getting Started explains how to download install and run the Carneades system locally on your personal computer The projects chapter explains how to create and access projects where each project can contain any number of argument graphs The argument graph chapter provides an overview of the underlying data model used by all the tools The browsing visualization and evaluation chapter explains how to read and view argument graphs The chapters on editing argument graphs
19. 1 of Structured Consultation Tool publisher IMPACT date 2012 8 System Administration 8 1 Building and Installing from Source Code 1 Use a Git client to clone a copy of the Carneades source code from the Carneades GitHub repository 2 Follow the instructions in the following README file carneades src CarneadesWeb README md 8 2 System Configuration To modify the configuration edit the config carneades clj file in the instal lation directory The configure files are Clojure source files with the properties represented as a Clojure map The projects property provides the full path name of the directory used to store Carneades projects The default directory is the projects directory of the installation directory Example projects usr local carneades projects Additional properties can be configured for each project as described in the next section 42 8 3 Project Structure and Configuration Carneades projects are stored in the directory stated in the carneades clj configuration file Each project is a directory with the following structure properties clj databases theories documents The properties c1j file defines the attributes of the project The properties are represented as a Clojure map in the file Here are the contents of an example properties clj file title My First Carneades Project creation date January 10 2015 theory default walton schemes description en This
20. Carneades 3 7 User Manual Tom Gordon Contents Introduction Getting Started 2 1 Downloading Carneades less 2 2 T enses 2 3 5 aaa Rec uen pe Aud bu ot 2 4 Binary Installation i2 o R9 ORE 2 3 1 Using the Web Application Locally The Projects Page Argument Graphs 4L Data Model xxx ete ae PE Xe wa 4 2 Statement Properties 2000000000 e 4 3 Argument Properties en 4 4 Premise Properties saaana aaa a a 4 5 Metadata andes ho SE 20 8 ar ake e D nn O E E AA a Ye de vr tad Browsing Visualizing and Evaluating Arguments b The OutlneJP ge 2 2ec046 4 amp 4RARV6GR RS N NN 5 2 Using Hypertext to Browse an Argument Graph 5 2 1 Statement Pages llle 5 2 2 Argument Pages lees 5 8 Visualizing Argument Graphs in Argument Maps 5 4 Evaluating Arguments a es ao ct c ea A 11 11 12 12 6 Editing Argument Graphs 6 1 Reconstructing Arguments 6 2 Entering New Statements 6 2 1 6 2 2 6 2 3 6 2 4 6 2 5 6 2 6 6 3 Editing Statements 6 4 Deleting Statements 6 5 Entering New Arguments 6 5 1 6 5 2 6 5 3 6 5 4 6 5 5 6 5 6 6 6 Editing Arguments 6 7 Deleting Arguments 6 8 Evaluating Argument Graphs Main Issue 544 Leon e eR GH XE eee rey ee cs Statement Text s Atom Formalizing Statements using Predicate Logic ProofStandard amp 4 c s gk kso ne s Statement Weight State
21. D are classified as instances of class G 9 9 Defeasible Modus Ponens id defeasible modus ponens conclusion B premises major If A is true then presumably B is also true minor A 4T 9 10 Argument from an Established Rule id established rule conclusion C premises major Rule R has conclusion C minor Rule R is applicable assumptions CQ1 Rule R is valid 9 11 Argument from Positive Consequences id positive consequences conclusion Action A should be performed premises major Performing action A would have positive consequences 9 12 Argument from Negative Consequences id negative consequences conclusion Action A should not be performed premises major Performing action A would have negative consequences 9 13 Argument from Practical Reasoning id practical reasoning conclusion A1 should be performed premises circumstances S1 is currently the case action Performing A1 in 81 would bring about 82 goal G would be realized in 82 value Achieving G would promote V 48 assumptions CQ1 V is indeed a legitimate value CQ2 G is a worthy goal CQ3 Action A1 is possible exceptions CQ4 There exists an action that when performed in S1 would bring about 82 more effectively than A1 CQ5 There exists an action that when performed in 81 would realize G more effectively than A1 CQ6 There exists an action that when performed in S1 would promote V more effectively than A1
22. ajor Source E is an expert in subject domain D domain Domain D contains the statement S minor E asserts that S is true exceptions CQ1 E is untrustworthy CQ2 S is inconsistent with the testimony of other witnesses assumptions CQ3 S is based on evidence 9 5 Argument from Analogy id analogy conclusion 2 premises major Case C1 is similar to the current case case 8 is true in case C1 minor E asserts that A is true exceptions CQ1 There are relevant differences between case C1 and the current case CQ2 S is false in case C1 which is more on point than case C2 9 6 Argument from Precedent id precedent conclusion 2 premises major Case C1 is similar to the current case ratio Rule R is the ratio decidendi of case C1 46 conclusion Rule R has conclusion S exceptions CQ1 There are relevant differences between case C1 and the current case CQ2 Rule R is inapplicable in this case 9 7 Argument from Verbal Classification id definition to verbal classification Strict true conclusion 0 is an instance of class G premises individual 0 satisfies definition D classification Objects which satisfy definition D are classified as instances of class G 9 8 Argument from Definition to Verbal Classification id definition to verbal classification Strict true conclusion O0 is an instance of class G premises individual 0 satisfies definition D classification Objects which satisfy definition
23. alysis tool an 43 argument graph is created for each case The database files are in the format used by the H2 database engine for the Java Virtual Machine The theories directory contains the Clojure source files of the theories of the project Finally the documents directory can be used to store copies of source documents documentary evidence or any other project files 9 Builtin Argumentation Schemes This chapter presents the argumentation schemes included with the distribution of the Carneades system The system is pre configured to use these schemes but you can configure the system to use other schemes or modify these schemes to meet your requirements The argumentation schemes are shown here in pseudocode for readabililty See the Modeling Argumentation Schemes chapter for a description of the syntax used to formally define the schemes The schemes can be viewed online using Carneades by clicking on the Theory link near the top of an argument outline page if the project has not been configured to use some other schemes Most of the schemes here are derived from the book Argumentation Schemes Walton Reed and Macagno 2008 The schemes for arguments from credible source and practical reasoning are based on Wyner Atkinson and Bench Capon 2012 and Atkinson and Bench Capon 2007 respectively The schemes from these sources been modified to fit the Carneades computational model of argument For exam
24. and modeling argumentation schemes are somewhat more demanding and some basic prior knowledge of argumentation theory Walton 2006 is recommended The system administration chapter describes how to build and install the system from its source files as well as how to configure the system The final chapter shows how some well known argumentation schemes Walton Reed and Macagno 2008 can been represented using the system 2 Getting Started This chapter explains how to download install and run the Carneades system locally on your personal computer For information about how to build the system from source code and advanced configuration options see the System Administration chapter 2 1 Downloading Carneades The Carneades Argumentation System is open source software available at https github com carneades You can download binaries and source code of Carneades releases from the releases page To use the version of Carneades described in this manual download the release named Carneades 3 7 2 2 License The source code of the Carneades system is licensed using the Mozilla Public License MPL version 2 0 An English version of the license is distributed with the software in the licenses directory The MPL license is certified by the Open Source Initiative OSI The MPL is a simple copyleft license The MPL s file level copyleft is designed to encourage contributors to share modifications they make to your co
25. b browser Depending on your operating system and how you started the server the Carneades web application can be shut down by either quitting the Carneades application or if you started the server from a command line using a termi nal application by ending this process typically by typing control c in the terminal 3 The Projects Page The projects page is the first page shown in your Web browser when you start Carneades It lists example projects and the projects you have created and provides a way to create new projects To access a project click on its name in the list of projects To create a new project click on the New Project button complete the form and then click on the Save button The theory field allows you to specify which argumentation schemes to use to construct and reconstruct arguments It is initialized to use the model of Walton s argumentation schemes distributed with the system from the default project default walton schemes Projects are stored in files on your file system in the directory specified in the carneades clj file in your home directory See the System Administration chapter for further configuration details You can change the theory to use other argumentation schemes including ones you have written yourself See the Argumentation Schemes chapter for information on how to write theories with your own argumentation schemes Save or move the theory to the theories dir
26. de while still allowing them to combine your code with code under other licenses open or proprietary with minimal restrictions In particular if you write an application which links to Carneades as a library you are free to use any license you wish for your own code See http www mozilla org MPL 2 0 FAQ html for more information 2 3 Binary Installation Prerequisites e Version 7 or better of a Java Runtime Environment Installation Procedure 1 Download the carneades 3 7 zip file from Carneades 3 7 release on the releases page 2 Unzip the carneades 3 7 zip Zip archive file using some Zip tool This will create a directory folder with the following hierarchical structure e carneades 3 7 carneades 3 7 jar config carneades clj doc manual pdf timestamp txt projects README txt You can move this directory to some other location on your file system at any time This creates a standard installation with the default configuration 2 3 1 Using the Web Application Locally To start the Carneades web application server double click on the carneades 3 7 jar file in your file system browser for example the Finder on Mac OS X or the Windows Explorer on Windows PCs To start the server to from a command line for local use type java jar carneades 3 7 jar Either way after the server starts it will open up the projects page of the Carneades web application in your default we
27. ections for the purpose of preserving and distributing online the cultural heritage of Europe Aktionsb ndnis p 6 A 912 How should the cross border aspects of the orphan works issue be tackled to ensure EU wide recognition of the solutions adopted in localhost 3000 projects copyright main statements urn963Auuid963A69272f7c 27da 4371 8119 e383c2c88814 Figure 4 The Outline of an Argument Graph Page 16 5 2 1 Statement Pages Carneades Web App x Way G B locathost 3000 carneades projects copyright main stateme d 27da 4371 8119 ex zv D Q iii Apps Cloud News Directories Shopping Projects FOKUS MARKOS Politics Cj Carneades Other bookmarks Projects Project Outline Argument Statement Map Outline Statement x Remove Statement Edit Statement New Argument Formalization Yes The exceptions should be clarified to allow works held in libraries to be scanned for the purpose of making their content searchable on the Internet Arguments for this statement position Notallthe material digitised by publishers is scanned with OCR Optical Character Recognition with the purpose of making the resulting content searchable 9 Clarifying the law to allow works held in libraries for the purpose of making the resulting content searchable on the Internet would have a transformative effect on research learning and teaching o Realizing a transformative effect on research learning and teaching is an important
28. ectory of your project in the projects directory specified in your carneades clj configuration file Suppose you have named this theory my_schemes clj and stored it in the theories directory of your project To have this theory used by your project instead of default walton_schemes edit the properties clj file in the project directory D Carneades Web App x D localhost 3 av 00 9 z pps Cloud CJ News CJ Directories CJ Shopping G Projects C3 FOKUS C3 MARKOS D Politics Cj Carneades 3 Other bookmarks Projects Projects New project Default Copyright in the Knowledge Economy Policy Compass MARKOS Figure 1 Example Projects Page and then change the theory to my_schemes Notice that you need not name the project since the theory is in this project and you also should not add the lj Clojure file extension Theories are implemented using a rule language embedded in the Clojure programming language This must be done locally on the computer running the Carneades server There is currently no Web user interface for editing the properties of projects or managing the theories of projects 4 Argument Graphs Argument graphs model relationships among arguments and statements claims propositions The arguments represented in an argument graph can be new invented arguments constructed from knowledge evidence and facts or re constructed from arguments in source documents such a
29. emise has exactly one statement node A statement may be the conclusion or premise of more than one argument A statement may be both a conclusion and a premise resulting in complex argument graphs representing chains or trees of reasoning Argument graphs may contain cycles A simple cycle would result if a statement is a premise and conlusion of the same argument There are methods for resolving these cycles when evaluating argument graphs A statement in an argument graph represents a propositional variable whose value may be true in false out or unknown undecided To continue with our example the sentences Johnny rode his skateboard in the park and Johnny did not ride his skateboard in the park would be represented by a single statement in an argument graph Conclusions and premises of argument can negated using con arguments and negative premises respectively T here are two kinds of arguments pro and con An argument is pro if its conclusion claims the statement is true and con if it claims the statement is false Similarly there are two kinds of premises positive and negative A positive premise holds if its statement is in Conversely a negative premise holds only if its statement is out Prior models of argument graphs did not distinguish pro and con arguments or positive and negative premises Rather in these prior approaches all argument nodes are pro and all premises are positive Our approach has the
30. en make form positive Achieving goal s would demote the value 4s negative Achieving goal s would not demote the value 4s question Would achieving goal s demote the H value s An argumentation scheme is represented as a structure with six fields 39 id header conclusion premises exceptions and OF Quom ee RS ES assumptions The id is a term in the language used to reify argumentation schemes and repre sent statements about argumentation schemes in domain models The header enables metadata about the scheme to be represented e g title description Descriptions can be represented in multiple natural languages The conclusion is a formula schema which may contain schema variables Schema variables are represented by symbols beginning with a question mark e g Ag 7A and G and range over both terms and propositions Thus the conclusion of an argumentation scheme can be a schema variable This feature is needed for representing schemes such as arguments from expert witness testimony whose conclusion may be any proposition whatsoever The premises exceptions and assumptions fields of schemes are vectors of premise structures where each premise has the following properties role A string naming the role of the premise in the argumentation scheme e g major minor circumstances goal positive Boolean False if the premise is negated Default true statement
31. ering new arguments is shown when you click on the New Argument button in the menu bar of the argument graph page or a statement page When used on a statement page the form for entering the new argument will be initialized with the statement as the conclusion of the argument See the section on Reconstructing Arguments for tips about how to proceed when interpreting and modeling arguments in source texts 6 5 1 Using Argumentation Schemes Arguments can be entered into the system without using argumentation schemes Their use is entirely optional That said argumentation schemes can be helpful for interpreting source texts when trying to reconstruct arguments They serve as templates which can guide you in your task of understanding the source text For example if the argument looks like it might be an argument from expert opinion using the expert opinion scheme will help you to remember the conventional form of this kind of argument Can the text reasonably be understood as including all the premises of the expert witness scheme Were any missing premises left implicit by the author or would this be reading too much between the lines Does the argument perhaps better fit some other scheme See the section on Reconstructing Arguments for tips about how to proceed when interpreting and modeling arguments in source texts You can view a list of the available argumentation schemes using the pull down menu Type in a string such as expe
32. for the statement to be deemed acceptable presumably true The proof standard is used by the computational model to argument to compute the acceptability of the statement Several proof standards are available Dialectical Validity DV This standard is the only one that does not make use of arugment weights It is satisfied if at least one pro argument is in and no con argument is n Preponderance of Evidence PE This standard is met if at least one pro argument is in that weighs more than any in con argument Clear and Convincing Evidence CE This standard is satisfied if the pre ponderance of evidence standard is met and in additional the difference between the strongest in pro argument and the strongest in con argument is above a certain threshold Beyond Reasonable Doubt BRD This standard is met if the clear and convincing evidence standard is meet and in addition the weight of the weakest in con argument is below a certain threshold The default proof standard is preponderance of the evidence and for most applications this proof standard should be sufficient Note that the preponderance 29 of evidence standard is met whenever the dialectical validity standard is met If arguments are not weighed the dialectical validity and preponderance of evidence standards will give the same results The preponderance of evidence clear and convincing evidence and beyond reasonable doubt standards are ordered by the amount of proof req
33. g premise roles The schemes are there to help you not constrain you See the Argumentation Schemes chapter for documentation of the initial set of schemes provided with Carneades These schemes may be modified or extended or replaced entirely as described in the Argumentation Schemes chapter 4 Enter the conclusion of the argument by choosing a statement already in the graph with the same meaning as the conclusion of the argument in the source text If the needed statement is not listed create one first following the instructions for entering new statements below 5 Similarly add the premises of the argument by choosing existing statements in the argument graph from the pull down lists below the role of each premise in the form Negative premises can be entered by changing the Positive property of the premise to False If the premise is implicit in 26 the source text you can note this by changing the Implicit property of the premise to True 6 At the bottom of the form click the Save Argument or Cancel button Warning Any changes you make using the form will not be saved to the argument graph until you execute the Save Argument command by clicking on this button 6 2 Entering New Statements The form for entering new statements is shown when you click on the New Statement button on the outline page of the argument graph page Carneades Web App x WESS lt Q D localhost 3 arneades inde
34. he fields of the form filled in with the current values of the properties of the argument After making changes click on the Save button at the bottom of the form to have the changes stored in the argument graph database or the Cancel button to abort the editing process and retain the prior values of the properties of the statement Warning The changes cannot be undone after saving them 6 7 Deleting Arguments To delete an argument from an argument graph first navigate to the argument page and then click of the Remove Argument button You will be asked to confirm or cancel the deletion Deleting an the argument does not delete the statements used in its conclusion or premises Any arguments undercutting this argument will also be deleted since the con clusion of such an undercutter refers to this argument specifically and can serve no function once the argument has been deleted Warning The deletion cannot be undone after it has been confirmed 6 8 Evaluating Argument Graphs The term argument evaluation has two different but related meanings The broader meaning described in the Evaluating Arguments section of this manual concerns the process of critically assessing the validity of arguments by 1 revealing implicit premises 2 validating whether the arguments are formally correct by instantiating accepted argumentation schemes 3 asking appropriate critical questions depending on the schemes applied 35
35. ine a2 ODARE Apps Cloud C3 News G Directories C3 Shopping G Projects J FOKUS C3 MARKOS Gj Politics Cj Carneades Other bookmarks Outline This outline provides an overview of the main arguments Click on an item in the outline for more information including further arguments Q9 Should the law be clarified with respect to whether the scanning of works held in libraries for the purpose of making their content searchable on the Internet goes beyond the scope of current exceptions to copyright EJ pro Yes The exceptions should be clarified to allow works held in libraries to be scanned for the purpose of making their content searchable on the Internet Not all the material digitised by publishers is scanned with OCR Optical Character Recognition with the purpose of making the resulting content searchable Clarifying the law to allow works held in libraries for the purpose of making the resulting content searchable on the Internet would have a transformative effect on research learning and teaching Realizing a transformative effect on research learning and teaching is an important social goal If the rules regulating the scanning of works held in libraries were unclear this would be known a It is not known that the rules regulating the scanning of works in libraries are unclear pro a Question 9 Is irrelevant pro Public libraries have the right indeed the responsibility to publish digital copies of their coll
36. ises are displayed with a circular dot arrowhead on the statement side of the edge The statement and argument nodes in argument maps contain hyperlinks Click ing on a statement or argument node displays the details of the node in a statement or argument page respectively 20 In argument maps argument nodes whose conclusion is another argument node rather than a statement node visualize undercutting arguments These are arguments which question the applicability of another argument This is the only case where two nodes of the same type are directly connected in the map Argument maps are represented using structured vector graphics SVG not bitmaps You can zoom the map in or out to any scale without loss of resolution How this zooming is done depends on your device and web browser When argument graphs have been evaluated the status of the argument and statement nodes is visualized in argument maps using both color and icons Nodes which are in are filled with a green background and contain a checked box Nodes which are out are shown with a red background and contain a crossed box a box filled with an X Nodes which are neither in nor out are filled with white background color and contain an empty checkbox The colors are redundant to accommodate black and white printing and color blind users 5 4 Evaluating Arguments By argument evaluation we mean the process of critically assessing arguments by
37. me applied if any Optional amp Examples argument from credible source argument from practical reasoning weight A real number in range of 0 0 1 0 representing the relative weight of the argument compared to other arguments pro and con the conclusion of the argument The weight is assessed by users If there are multiple users the weights can be collected using polls value A real number in in the range 0 0 1 0 used to record the output of the argument evaluation process where 0 0 means the argument is out not acceptable 1 0 means the argument is in acceptable and all other values mean the arguments in the graph taken together are insufficient for determining the acceptability of this argument undecided unknown Real numbers are used to allow experimentation with other models of argument evaluation 4 4 Premise Properties polarity Positive or negative role The role of the premise in the argumentation scheme applied Examples minor major implicit A Boolean value true or false Can be used to note that the premise was not explicit in the source documents from which the argument node was reconstructed 4 5 Metadata The argument graph as as whole as well as each of its statements and arguments can be annotated with metadata using the Dublin Core There are 15 elements in the Dublin Core Each element may contain zero or more values Here is a list of the Dublin Co
38. ment Metadata Using Argumentation Schemes Entering and Deleting Premises ss Strict or Defeasible Arguments Argument Direction len Argument Weight lle Argument Metadata e 7 Argumentation Schemes 7 1 Modeling Argumentation Schemes 8 System Administration 8 1 Building and Installing from Source Code 8 2 System Configuration 8 3 Project Structure and Configuration 24 26 27 28 28 29 29 30 30 30 31 31 31 33 33 34 34 34 35 35 35 36 37 9 Builtin Argumentation Schemes 9 1 9 2 9 3 9 4 9 5 9 6 9 7 9 8 9 9 9 10 9 11 9 12 9 13 9 14 9 15 9 16 9 17 9 18 9 19 9 20 9 21 Argument from Position to Know les Argument from Credible Source llle Argument from Witness Testimony Argument from Expert Opinion llle Argument from Analogy a Argument from Precedent e Argument from Verbal Classification less Argument from Definition to Verbal Classification Defeasible Modus Ponens llle Argument from an Established Rule sss Argument from Positive Consequences les Argument from Negative Consequences Argument from Practical Reasoning Argument from Cause to Effect 20 2000 Argument from Correlation to Cause 2 0 Argument from Sunk Costs e Argument from Appearance Argument from Ignorance 0000
39. ng using Argument modelling and Computer assisted Text analysis IMPACT FP7 IST 247228 2010 2012 Simple Procedures for Cross Border Services SPOCS CIP 238935 2009 2012 The MARKet for Open Source An Intelligent Virtual Open Source Marketplace MARKOS FP7 ICT 317743 2012 2015 Policy Compass FP7 ICT 612133 2013 2016 Enhanced Government Learning EAGLE FP7 ICT 619347 2014 2017 11 User Manual License IEO This user manual is licensed under a Creative Commons Attribution 4 0 International License 52 References Atkinson Katie and Trevor J M Bench Capon 2007 Practical Reasoning as Presumptive Argumentation Using Action Based Alternating Transition Systems Artificial Intelligence 171 10 15 855 74 Gordon Thomas F Henry Prakken and Douglas Walton 2007 The Carneades Model of Argument and Burden of Proof Artificial Intelligence 171 10 11 875 96 Walton Douglas 2006 Fundamentals of Critical Argumentation Cambridge UK Cambridge University Press Walton Douglas Chris Reed and Fabrizio Macagno 2008 Argumentation Schemes Cambridge University Press Wyner Adam Katie Atkinson and Trevor Bench Capon 2012 A Functional Perspective on Argumentation Schemes In Proceedings of the 9th International Workshop on Argumentation in Multi Agent Systems ArgMAS 2012 edited by Peter McBurney Simon Parsons and Iyad Rahwan 203 22 53
40. nother scheme any statements you have selected for the conclusion and premises of the argument will remain but you will need to check whether they are still appropriate and add modifiy or delete premises as necessary Again the argumentation schemes are meant to be helpful not get in your way They do not constrain you to enter arguments matching the schemes You can however check whether arguments correctly instantiate the schemes they have been assigned See the Validating Arguments section for further information 6 5 2 Entering and Deleting Premises In the premise of the argument editor you can choose an existing statement in the argument graph from the pull down list You can filter the statements in the list by typing some text in the text field at the top of the list You can insert additional premises by clicking on the Add Premise button Premises can be deleted by clicking on the Remove this premise button immediately below each premise 6 5 3 Strict or Defeasible Arguments An argument may be strict or defeasible not strict An argument is strict if and only if its conclusion must be true with no exceptions if its premises are true whether or not its premises are in fact true For example the following arguments are strict e The moon is made of green cheese since the moon is made of green cheese and unicorns have horns e The figure is a triangle since it has three sides The first example argument i
41. o the graph by clicking on the New Statement or New Argument buttons In both cases you will be presented with a form to enter the required information The form will be displayed in a new tab so that you can toggle back and forth between different views of the argument graph will editing Warning A known bug in the user interface causes your work to be lost if you change to another tab without first saving your work See the Entering New Statements or Entering New Arguments sections for further information To add a new argument pro or con some existing statement go to the statement page and click on the New Argument button The conclusion of the new argument will be set to the existing statement Then complete the rest of the form as described in the Entering New Arguments section 24 D Carneades Web App x G D localhost 3000 carne projects copyright main outline E OQ D Q H Apps Cloud News 3j Directories Shopping CJ Projects FOKUS MARKOS Politics Cj Carneades Other bookmarks Projects Project Outline Map Theory Reconstruction of Comments on the EU Green Paper Copyright in the Knowledge Economy Edit Metadata amp New Statement New Argument Share The purpose of the Green Paper on Copyright in the Knowledge Economy is to foster a debate on how knowledge for research science and education can best be disseminated in the online environment 6GreenPaper p 3 The Green Paper ha
42. onal model of argument is used to compute the values Gordon Prakken and Walton 2007 If you do not agree with the result of the evaluation there are at least three reasons why you may be right and the system s evaluation wrong 1 Not all relevant arguments have been included in the model or put forward in the debate If you are aware of some missing argument consider adding it to the argument graph 2 The weights assigned to the statements and arguments may not reflect your opinion If these weights are averages of weights assigned collectively by a group of users you may disagree with the collective view Minority views can be correct 3 The formal model of argument we are we are using to compute acceptability may be incorrect Of course specialist knowledge is required to assess the correctness of the model If you like most people do not have this knowledge then we recommend a skeptical but respectful attitude If you agree with the results of the model then the model gives you a reason to have more confidence in your opinion If you do not agree with the results of the model then you may want to take pause to reconsider your views even if in the end you do not change your mind 6 Editing Argument Graphs To edit an argument graph first go to its outline as described in the Chapter entitled Browsing Visualizing and Evaluating Arguments An example argument graph page is shown in the figure below You can add nodes t
43. opping Projects CJ FOKUS MARKOS Politics D Carneades Other bookmarks Argument from Practical Reasoning Source Atkinson K and Bench Capon T J M Practical reasoning as presumptive argumentation using action based alternating transition systems Artificial Intelligence 171 10 15 2007 855 874 If the following are true S1 is the case in the current circumstances Performing A in S1 would bring about S2 Gwould be realized in S2 Achieving the goal G would promote the value V Presumably it is true that Action A should be performed Unless There exists an action that would bring about S2 more effectively than A There exists an action that would realize the goal G more effectively than A There exists an action that would promote the value V more effectively than A Performing A in S1 would have side effects which demote V or some other value practical reasoning 0 5 Argument from Cause to Effect If the following are true An event E1 has occurred Event E1 causes event E2 Presumably it is true that An event E2 will occur Unless An event E3 interfered with event E1 cause to effect 0 5 Figure 10 A Scheme for Argument from Practical Reasoning 23 an issue of it by adding arguments pro or con the assumed statement in the argument graph Any time an argument graph is modified it is automatically reevaluated to update the values of its arguments and statements A computati
44. ple generic critical questions which undermine premises undercut the argument or rebut its conclusion have been omitted since these critical questions apply to all defeasible arguments in Carneades All the argumentation schemes presented here are defeasible unless they have been explicitly declared to be strict 9 1 Argument from Position to Know id position to know conclusion S premises major W is in a position to know about things in a certain subject domain D minor W asserts that S is true 44 domain S is in domain D exceptions CQ1 W is dishonest 9 2 Argument from Credible Source id credible source conclusion S premises Source W is a credible source about domain D assertion W asserts S domain S is in domain D exceptions CQ1 W is biased CQ2 W is dishonest CQ3 Other credible sources disagree with 2 9 3 Argument from Witness Testimony id witness testimony conclusion S premises position to know W is in a position to know about things in a certain subject domain D in domain Domain D contains the statement 2 truth telling Witness W believes S to be true minor W asserts that 2 is true assumptions CQ1 S is internally consistent exceptions CQ2 S is inconsistent with the facts CQ3 S is inconsistent with the testimony of other witnesses CQ4 W is biased CQ5 S is implausible 45 9 4 Argument from Expert Opinion id expert opinion conclusion 2 premises m
45. re elements Title Creator Subject Description Publisher Contributor Date Type Format WON eo P cS I PE 12 10 Identifier 11 Source 12 Language 13 Relation 14 Coverage 15 Rights See Dublin Core for a detailed description and usage guidelines for each element The Dublin Core is intended to be useful for describing a wide range of resources on the World Wide Web In addition Carneades allows each metadata record to be assigned an optional key a string which can be used as a label to refer to the metadata record such as BenchCapon 2008 similar to the way citation keys are used in bibliographic databases such as BibTeX At most one key should be provided Carneades provides special support for providing description elements of the Dublin Core in multiple languages English German French and for formatting these descriptions using the Markdown language This feature can be used to include quotations from and links to source documents in the descriptions of both statements and arguments 5 Browsing Visualizing and Evaluating Argu ments This chapter of the Carneades user manual explains how to e Access the outline page to view the title description issues and outline of the top level arguments of the argument graph e Use hypertext in web pages to browse an argument graph e Visualize argument graphs in diagrams called argument maps and using these maps to navigate to more de
46. rence engine which can be used to find arguments in this way Both methods can be used together Carneades is an interactive argumentation assistant not a fully automatic problem solver Arguments which have been found automatically can be manually edited by users 4 1 Data Model Metadata statement Argument Premise Figure 2 Entity Relationship Diagram The entity relationship diagram above shows the elements of argument graphs and their connections The figure does not visualize a particular argument graph but rather relationships between the elements of argument graphs in general The two main elements of argument graphs are statements and arguments Statements represent propositions claims and assertions Arguments represent simple inferences from one or more premises to a single conclusion Again there should be only one statement or argument in the graph for each statement and argument in the source documents no matter how many different ways the statement or argument has been expressed in source documents Some or all formulations of the statement or argument can be quoted or referenced in the metadata of the statement or argument node See the discussion of metadata below for further information As can be seen in the entity relationship diagram arguments are linked to statements in two ways in argument graphs Each argument has exactly one conclusion which is a statement and zero or more premises where each pr
47. resenting the statement itself Translations of the text can be entered in several natural languages En English De German Fr French It Italian NI Dutch Sp Spanish 28 6 2 3 Atom Formalizing Statements using Predicate Logic Optionally statements can be formalized in predicate logic by providing a value for the atom property in the form This is an advanced feature that may be needed only for more specialized application scenarios Atoms are formalized in Carneades using the prefix notation of s expressions symbolic expressions as in the Lisp family of programming languages For example Socrates is a man could be formalized as man Socrates Here man is a unary predicate symbol The fact that Socrates died in 399 B C could be represented using a binary predicate for example as died Socrates 399 The language for Atoms is quite expressive e The arity of atoms can be 0 or more as in predicate logic not restricted to unary and binary predicates as in description logic RDF and OWL Examples good between O 1 2 e Compound terms may be used For example 2 1 1 The is the atom language is not restricted to Datalog e Atoms can be higher order Example believes Gerald and made of Moon green cheese really exists Yeti Notice that compound propositions can be reified as terms as in this example 6 2 4 Proof Standard The proof standard of a statement determines how much proof is required
48. rguments submitted but rather to model a sufficient number of representative arguments and policies for the purpose of illustrating features of the IMPACT argument toolbox The Corpus Selection Working Group of the IMPACT project has chosen 4 of the 25 questions raised in the Green Paper as well as 12 of the 323 comments submitted representing a wide range of stakeholders to be used for the research and development purposes of the project The four questions covered by this model are listed below Click on a question for further information Main Issues Figure 3 An Outline Page 14 5 2 The title of the argument graph This title usually includes the topic of the discussion or debate A menu bar of commands A description of the topic of the discussion modeled in the argument graph The description can be arbitrarily long and include multiple sections paragraphs images hyperlinks lists and other content A list of the main issues of the discussion Each item in the list is linked to a page providing detailed information about the statement in the argument graph at issue The outline which presents the top five levels of the arguments in the argument graph The first level of the outline lists the main issues again The second level lists the arguments pro and con each issue The third level lists the premises of each of these arguments The fourth level lists the argument pro and con each premise Finally the fifth le
49. rt to display the schemes containing 31 D Carneades Web App x Q B localhost 3000 carneades index htmlit projects copyrigh ew av 0D z Apps Cloud News G Directories Shopping Projects CJ FOKUS C3 MARKOS Politics Cj Carneades Other bookmarks Projects Outline Statement Argument New Argument Map Outline New argument Model Metadata Scheme Argument from Witness Testimony z Source Douglas Walton Henry Prakken Chris Reed Argumentation Schemes and Generalisations in Reasoning about Evidence Proceedings of the 9th International Conference on Artificial Intelligence and Law Edinburgh 2003 New York ACM Press 2003 pp 35 Douglas Walton Witness Testimony Evidence unpublished book manuscript to appear Premises Positive True False Role position to know Statement X Remove this premise Figure 13 New Argument Form 32 the string in the title Then select the desired scheme from the list This will display a description of the scheme perhaps with example arguments along with references to publications about the scheme You can go back and select another scheme at any time Whenever a scheme is selected the form will customized to include premises and exceptions fields for the chosen scheme The roles of the premises and exceptions will be modified to match the selected scheme The premise roles can be edited You are not constrained by the scheme If you change your mind and select a
50. s Web App x 4 Q D localhost 3000 carneades re uid 4cdf9 4 a3a G 60 09 S Apps Cloud C3 News G Directories C3 Shopping G Projects J FOKUS C3 MARKOS Ci Politics Cj Carneades Other bookmarks 9 Argument Edit Argument x Remove Argument Model Metadata Formalization If the following are true Circumstances The 2001 EU Copyright Directive regards copyrights and exceptions but not remedies for violations of copyrights Action Using a stand alone instrument to regulate orphaned works instead of amending the 2001 Copyright Directive would cause the separate topic of remedies for copyright violations to be regulated by a separate instrument Goal Separate legal topics should be regulated by separate instruments Values Promoted Regulating separate legal topics with separate instruments promotes the value of legal clarity Presumably it is true that While we do not support such an approach to the extent that an orphan works standard is adopted throughout the EU we recommend that a Community statutory instrument dealing with the problem of orphan works should be a stand alone instrument SIIA p 9 Properties Unclear Argument from Practical Reasoning false 0 5 urn uuid 4cdf9c48 2104 4773 a3af d6bGeB3ec35d Figure 9 An Argument from Practical Reasoning 22 D Carneades Web App x Q D localthost 3000 carneades proje opyright theories default p z Apps Cloud C3 News G Directories Sh
51. s court opinions or newspaper commentaries Argument reconstruction is a challenging task requiring the interpretation of natural language texts Consider the following simple legal argument Johnny violated the law by riding his skateboard in the park This same argument can be expressed in many different ways in natural langauge Here are some examples 1 Because Johnny rode his skateboard in the park he violated the law 2 Vehicles are prohibited in the park Someone who does something which is prohibited violates the law Johnny rode his skateboard in the park A skateboard is a vehicle Therefore Johnny violated the law 3 Johnny hat gegen das Gesetz verstofen weil er sein Skateboard im Park benutzt hat The first example just reordered the premise and the conclusion of the argument putting the premise first The second example reveals some implicit premises of the original formulation of the argument The third example is a German translation of the original formulation of the argument All four of these texts including the original formulation express the same argument but in different ways In a large scale debate for example about European policy issues the same argument might be expressed in thousands of different ways in many different languages The European Union has 23 official languages One of the purposes of argument graphs is to provide a way to summarize the arguments put forward in complex debates with exac
52. s of course cyclic not to mention fanciful But it is strict nonetheless Arguments are defeasible if they are not strict With defeasible arguments the conclusion is only presumably true when the premises are true The argument 33 gives us a reason to accept the conclusion but there may be exceptions or other reasons leading to the opposite conclusion Here are a couple of famous examples from the computational models of argument literature e The object is red since it looks red e Tweety flies sinces Tweety is a bird The object may look red because for example if is being illuminated by a red light And Tweety the bird may be a penguin or have a broken wing and so on The conclusion of a defeasible argument need not be probably true in some statistical sense It is not necessary to have good empirical data allowing us to draw conclusions about what is probably the case in order to make a defeasible argument The argument only needs to give us some good reason to believe the conclusion at least until we have heard arguments giving us reasons to the contrary Select whether the argument is strict in the form using the radio buttons By default arguments are defeasible 6 5 4 Argument Direction Arguments can be pro or con their conclusion An argument con a statement is semantically equivalent to an argument pro the negation of the statement 6 5 5 Argument Weight Argument weights only become important when two or more
53. s two parts The first deals with general issues and the second deals with specific issues related to the exceptions and limitations which are most relevant for the dissemination of knowledge and whether these exceptions should evolve in the era of digital dissemination GreenPaper p 3 Here we present a reconstruction of some of the policies and arguments put forward in the comments submitted In reponse to the Green Paper Our aim is not to comprehensively model all the policies and arguments submitted but rather to model a sufficient number of representative arguments and policies for the purpose of illustrating features of the IMPACT argument toolbox The Corpus Selection Working Group of the IMPACT project has chosen 4 of the 25 questions raised in the Green Paper as well as 12 of the 323 comments submitted representing a wide range of stakeholders to be used for the research and development purposes of the project The four questions covered by this model are listed below Click on a question for further information Main Issues Figure 11 An Argument Graph Page 25 To edit or delete existing statements or arguments first go to the page of the statement or argument and then click on the Edit or Remove button Editing statements and arguments is done using the same forms used to create new statements and arguments Deleting a statement will also delete the arguments pro or con the statement Deleting an argument does
54. sing the Markdown wiki language The form includes a Markdown editor to make this easier for you As for the statement text translations of descriptions can be entered in several natural languages 6 3 Editing Statements To edit an existing statement in an argument graph first navigate to the statement page and click Edit Statement button This will reveal a form for modifying the statement This is the same form used to enter new statements described in the Entering New Statements section of this chapter but with the fields of the form filled in with the current values of the properties of the 30 statement After making changes click on the Save button at the bottom of the form to have the changes stored in the argument graph database or the Cancel button to abort the editing process and retain the prior values of the properties of the statement Warning The changes cannot be undone after saving them 6 4 Deleting Statements To delete a statement from an argument graph first navigate to the statement page and click of the Remove Statement button You will be asked to confirm or cancel the deletion Warning The deletion cannot be undone after it has been confirmed Deleting a statement also deletes all arguments pro or con this statement i e with this statement as the conclusion of the argument The statements for the premises of these arguments are not deleted 6 5 Entering New Arguments The form for ent
55. t pages are quite similar to statement pages At the top the page the premises and conclusion of the argument are shown If available the role of each premise in the argumentation scheme applied is shown e g major or minor The bottom of the page displays the properties of the argument its id the argumentation scheme applied if any whether it is a strict or defeasible argument its weight and value The argumentation scheme contains a hyperlink Click on the link to view a description of the scheme If metadata had been provided for the argument it would be displayed next Descriptions can include quotations of one or more source texts expressing the argument along with hyperlinks to the sources on the Web 18 D Carneades Web App x lt Q B localhost 3000 carneades projects copyright main arguments urn uuid c5ac7cb8 e942 4a6f bb18 f6 3 OQ D Q HE pps Cloud News Gj Directories CJ Shopping CJ Projects C3 FOKUS MARKOS Politics Cj Carneades Other bookmarks Map Outline 9 Argument Edit Argument x Remove Argument Model Metadata Id Source LIBER Description The Association of European Research Libraries LIBER in response to Question 9 of the Green Paper argues that yes the law should the law be clarified with respect to whether the scanning of works held in libraries for the purpose of making their content searchable on the Internet goes beyond the scope of current exceptions to copyright as follo
56. tailed views of statements and arguments e Evaluate arguments to reveal missing premises check the form of arguments ask critical questions and assess the acceptability of statements 5 1 The Outline Page An outline page consists of the following parts lhttp en wikipedia org wiki BibTeX 13 D Carneades Web App x C 5 localhost carneades projects copyright main outline av ODAB E Apps Cloud C3 News G Directories C3 Shopping G Projects J FOKUS C3 MARKOS Gj Politics Cj Carneades Other bookmarks Projects Project Outline Map Theory Reconstruction of Comments on the EU Green Paper Copyright in the Knowledge Economy Edit Metadata New Statement New Argument Share The purpose of the Green Paper on Copyright in the Knowledge Economy is to foster a debate on how knowledge for research science and education can best be disseminated in the online environment 6GreenPaper p 3 The Green Paper has two parts The first deals with general issues and the second deals with specific issues related to the exceptions and limitations which are most relevant for the dissemination of knowledge and whether these exceptions should evolve in the era of digital dissemination GreenPaper p 3 Here we present a reconstruction of some of the policies and arguments put forward in the comments submitted In reponse to the Green Paper Our aim is not to comprehensively model all the policies and a
57. tional arguments not only deletions can cause the labels to change since acceptability is a nonmonotonic inference relation 7 Argumentation Schemes Argumentation schemes are represented using a high level scheme rule language embedded in the Clojure programming language Argumentation schemes generalize the notion of an inference rule to cover defeasible as well as strict reasoning patterns We use the term scheme instead of rule to emphasize that the rules are usually defeasible The scheme language is expressive enough to represent axiomatizations ie axioms and inferences rules of theories in many domains including laws regulations and policies in addition to argumentation schemes per se Computational models of theories have been called many things in computer science including knowledge bases and deep conceptual or semantic models We prefer the term theory to knowledge base because knowledge may suggest consensus or truth while it is clearer that theories may be controversial or contested 36 The scheme language is similar to logic programming languages such as Prolog Any Prolog rule clause can be represented in Carneades in a straight forward way The rule language has some additional features for representing argumen tation schemes such as scheme variables ranging over atoms and a means to represent premise roles e g major minor Moreover schemes in this language can
58. tly one node in the graph for each argument no matter how often or in how many ways it has been expressed A single argument graph is used to represent all the arguments put forward in debate from all participants The nodes can quote one or more source documents and include links to these source documents so no information need be lost and all voices can still be heard inclusively Grouping the different formulations of an argument together into a single node in the graph abstracting away details makes it possible to quickly obtain an overview of the arguments and to obtain a clearer picture of relationships among arguments A hypertext or map of the source documents directly without an argument graph would make it difficult to see the forest for the trees Argument graphs are also useful as a decision support tool when analysing problems and issues such as legal issues whether by an invidual or in a group When used this way the arguments in the graph represent inference steps and dependencies between propositions Original arguments can be entered or edited directly by users whether or not the arguments have first been formulated in some natural language text If the relevant knowledge of the application domain has been formally represented using a knowledge based system some arguments can be automatically found and entered into the graph Carneades provides a knowledge representation language for argumentation schemes and an infe
59. uired with beyond reasonable doubt requiring the most proof Whenever one of these standards is met all of weaker standards are also meet 6 2 5 Statement Weight The weight of a statement encodes the extent to which users agree or disagree with the statement on a scale of 0 0 to 1 0 where 0 0 represents disagree reject 0 5 means no opinion and to 1 0 is denotes agree accept Statements with a weight greater then or equal to 0 75 and less than 1 0 are assumed to be true until an argument pro or con the statement has been added to the argument graph Conversely statements with a weight less than or equal to 0 5 and greater than 0 0 are assumed to be false until an argument pro or con the statement has been added to the graph 6 2 6 Statement Metadata Statements can be annotated with metadata properties using the Dublin Core metadata elements The Dublin Core elements are briefly summarized in the Metadata section of this manual To add metadata to a statement go to the Metadata tab of the statement editor and complete the form Multiple values of an element can be entered by listing Separate the items in the list with semicolon characters You can use the description field of the Metadata tab of the statement editor to provide whatever background information you want about the statement You can structure and format the description including headers lists quotations hypertext links and other elements u
60. varied semantics The semantics of theories are defined by mapping instantiations of argumentation schemes to argument graphs and in turn evaluating these graphs using the computational model of structured argument presented in Gordon Prakken and Walton 2007 7 1 Modeling Argumentation Schemes In this section we illustrate how to use the language to represent a version of the scheme for arguments from practical reasoning A domain theory is represented by first defining a language dictionary of symbols denoting predicates and terms and then a set of inference rules called schemes using this language The language is represented as a map from symbols to predicates and individuals in the Clojure programming language Each symbol of the language is mapped to a structure with fields for the symbol of the predicate redundantly the 37 arity of the predicate i e the number of columns in a tabular representation of the relation denoted by the predicate and an optional number of forms for expressing statements and questions about this predicate in one or more natural languages To illustrate below are the definition of some of the predicates of the language used a version of the schemes for practical reasoning def L circumstances make predicate symbol circumstances arity 1 forms en make form positive The circumstances are fs negative The circumstances are not 7s question Are the circumstances
61. vel lists the premises of these arguments Deeper levels of the argument graph can be navigated to by first clicking on a statement or argument in the outline and then following the links on the next page Since argument graphs may contain cycles and are not restricted to trees some items may appear multiple times in the outline A list of references to the source documents used to construct the argument graph For documents available on the Web the reference will include a hyperlink to the source document Notice that the issues and items in the outline are prefixed with colored dots These indicate the current status label of the statements and arguments in the graph where Yellow means the statement or argument is undecided or unknown Green means the statement or argument is acceptable in true or presumably true Red means the statement or argument is not acceptable out false or presumably false Using Hypertext to Browse an Argument Graph There is a web page for each statement and argument in the argument graph providing detailed information about the element along with links to related statements and arguments in the graph You can use these pages to navigate from node to node in the argument graph by simply clicking on the links in the usual way To go back to previous pages use the back button of your web browser 15 D Carneades Web App x lt gt B localthost 3000 carneades projects copyright main outl
62. ws Not all the material digitised by publishers is scanned with OCR Optical Character Recognition with the purpose of making the resulting content searchable If the rights holders will not do this libraries should be able to offer this service It would have a transformative effect on research learning and teaching by opening up a mass of content to users which can be searched using search engines The interests of copyright holders will not be harmed because the resulting output will act as marketing material for their materials LIBER p 3 Figure 7 Metadata Tab of an Argument Page 19 5 3 Visualizing Argument Graphs in Argument Maps The menus of the argument graph page statement pages and argument maps include map and outline buttons Clicking on the map button generates a diagram called an argument map which visualizes the argument graph as a network directed graph of statement nodes and argument nodes connected by links Statement nodes are shown as boxes argument nodes with circles and boxes with rounded corners D Carneades Web App WESS lt C D localhost 3 arneades projects copyright main map Qi 00DA iii Apps G Cloud CJ News G Directories J Shopping G Projects J FOKUS C3 MARKOS G Politics C3 Carneades Other bookmarks Projects Project Outline Argument Statement Map Outline Current the lack of harmonizaton of copyrigh exceptions facliates the freedom
63. x html projects copyric ain statements nev 4x OD Q iii Apps G Cloud CJ News G Directories CJ Shopping G Projects D FOKUS C3 MARKOS Gj Politics CJ Carneades Other bookmarks Projects Outline Statement Argument New Argument New Statement Map Outline New statement Model Metadata Main issue True False Figure 12 New Statement Form A new statement should not be entered into the argument graph if the negation of the statement is already in the argument graph That is for every proposition P a single statement node should be entered into the graph to represent both P and P It doesn t matter whether the positive or negative form of the statement is included explictly in the argument graph Typically the positive form is used but in some cases you may prefer the negative form For example you may prefer to include the sentence The payment was illegal explicitly in the argument graph and then represent arguments pro the legality of the 27 payment as argument con the claim the payment was illegal The best choice may depend on the procedural context of the debate Only one form of the statement is required since arguments pro the negation of the statement are equivalent to arguments con the statement And premises of arguments can be explicitly negated in argument graphs and in the forms for entering and editing arguments This approach has the advantage of reducing the number of statements in the graph up to 5

Carneades 3.7 User Manual

Contents

Download Pdf Manuals

Related Search

Related Contents