Semantic Web 0 (0) 1 IOS Press A Review of Argumentation for the Social Semantic Web Editor(s): Harith Alani, The Open University, UK Solicited review(s): Fouad Zablith, American University of Beirut, Lebanon; Simon Buckingham Shum, The Open University, UK; Iyad Rahwan, Masdar Institute, Abu Dhabi Open review(s): Jodi Schneider*, Tudor Groza®, and Alexandre Passant * * Digital Enterprise Research Institute, National University of Ireland, Galway, Jirstname.lastname @ deri.org » School of ITEE, The University of Queensland, Australia, tudor.groza @ uq.edu.au Abstract. Argumentation represents the study of views and opinions that humans express with the goal of reaching a conclusion through logical reasoning. Since the 1950’s, sev- eral models have been proposed to capture the essence of in- formal argumentation in different settings. With the emer- gence of the Web, and then the Semantic Web, this model- ing shifted towards ontologies, while from the development perspective, we witnessed an important increase in Web 2.0 human-centered collaborative deliberation tools. Through a review of more than 150 scholarly papers, this article pro- vides a comprehensive and comparative overview of ap- proaches to modeling argumentation for the Social Seman- tic Web. We start from theoretical foundational models and investigate how they have influenced Social Web tools. We also look into Semantic Web argumentation models. Finally we end with Social Web tools for argumentation, including online applications combining Web 2.0 and Semantic Web technologies, following the path to a global World Wide Ar- gument Web. Keywords: Argumentation, Semantic Web, Social Web, Se- mantic Web, Ontologies 1. Introduction In recent years, the problem of representing large- scale argumentation on the Web has attracted the at- tention of scholars from fields such as artificial in- telligence [137], communication theory [5], business management [88] and e-government [107]. At the same time, argumentation researchers began establish- ing the foundations for a World Wide Argument Web (WWAW) as “a large-scale Web of interconnected ar- guments posted by individuals to express their opin- ions in a structured manner” [138]. Arguments on the Web can be used in decision- making contexts. Decision-making often requires dis- cussion not just of agreement and disagreement, but also the principles, reasons, and explanations driving the choices between particular options. Furthermore, arguments expressed online for one audience may be of interest to other (sometimes far-flung) audiences. It can be difficult to re-find the crucial turning points of an argumentative discussion, even one in which we have participated. Yet on the Web, we cannot subscribe to arguments or issues, nor are there tools that support searching for arguments. Nor can we summarize the rationale behind a group’s decision, even when the dis- cussion took place entirely in public venues such as mailing lists, blogs, IRC channels, and Web forums. By providing common languages and principles to model and query information on the Web (such as RDF [90], RDFS [1], OWL [3], SPARQL [2], Linked Data principles [18], etc.), the Semantic Web [20] is an appropriate means to represent arguments and ar- gumentation uniformly on the Web, and to enable, for instance, browsing distributed argumentation patterns that appear in various places on the Web. Indeed, re- searchers have shown that the Semantic Web can be used for visualization and comparison in decision ra- tionale [102]. In this context, this paper discusses research in mod- eling argumentation as it relates to the Social Seman- 1570-0844/0-1900/$27.50 ©) 0 — IOS Press and the authors. All rights reserved 2 Schneider et al. / A Review of Argumentation for the Social Semantic Web tic Web [10,70,27], focusing on foundational models of argumentation, their applications in the Social Web, and on ontologies (as in computer science [69]). In particular, our purpose is to investigate ontologies and tools which may be useful for argumentation on the Social Semantic Web, a field where the aforemen- tioned Semantic Web technologies support Social Web [123] applications, while at the same time Social Web paradigms are used to generate Semantic Web data col- laboratively and at large scale. This convergence aims at providing new and improved ways to integrate and discover data, following the vision of Social Machines provided by Berners-Lee [19], both on the Web and in the enterprise [127]. In the context of argumenta- tion, this could help to aggregate arguments from var- ious websites — for instance a discussion starting on Twitter and followed up on a mailing list, later frozen on a wiki once consensus is reached — thus providing new means to follow argumentative discussions on the Web. This would enable an argument-centric view of the Web. Moreover, the Social Web does not yet have widely- used argumentative ontologies, though this problem has been noted [67], along with the need for federation infrastructures [125]. Thus, in order to identify how different argumentation models and tools can be used for the Social Semantic Web, this paper offers a review of more than 150 research papers on the topic, from 1945 to 2011, from which we compare: — 14 theoretical models of argumentation — 14 Semantic Web models for argumentation (i.e. ontologies) — 37 tools for representing argumentation on the Web. As the focus is on human-centered argumentation [87], with the goal of improving access and providing overviews and visualizations, this article will briefly mention, but not analyze, the agent-based argumenta- tion domain! Following the introduction, we provide brief overviews of argumentation (Section 2.1) and of the Social Web (Section 2.2), then discuss requirements for support- ing argumentation on the Social Semantic Web (Sec- tion 3). We next present theoretical models of argu- 'See for instance the Argumentation in Multi-Agent Systems (ArgMAS) Workshop series, in its ninth year in 2012. In particular, in the social context, Heras has presented argumentation work from a social perspective using case-based reasoning and ontologies e.g. the ArgCBROntology ?. [76,73]. mentation (Section 4) from a variety of fields, compare them (Section 5), and present applications of these the- oretical models (Section 6). Subsequently we present (Section 7) and compare (Section 8) Semantic Web models of argumentation. Then we move on to review- ing tools: in Section 9 we highlight thirteen notewor- thy features of Social Web argumentation tools, based on a comprehensive analysis of thirty-seven relevant tools (see the Appendix for full details). Finally we conclude the paper in Section 10. 2. Background 2.1. Argumentation Argumentation theory is the study of agreement, disagreement, and of the dialogues and writing through which we convince ourselves and others of our points of view [65]. Informal argumentation occurs through- out conversations, online and offline, often in conjunc- tion with persuasion or with joint decision-making. Even logically sound decisions may involve choices based on values and preference judgements: people may agree on the facts of a situation yet disagree on the preferred outcome or decision to be taken. That is vitally different from disagreeing on the facts of a sit- uation (in which case more information is called for). We are concerned with argumentative discussions, which we take to be online, mainly textual messages and discussions, in which subjective perspectives or differences of opinion are important and relevant. Groups may use online conversations and social media to coordinate and support decision-making; such argu- mentative discussions can be found in many online dis- cussion fora such as standardization bodies’ listservs, Wikipedia editors’ wiki pages, and open source com- munities’ IRC channels, bug reports, and listservs. In- dividuals may also draw on online conversations and social media in order to form personal opinions and clarify their own preferences, based on sensemaking and analysis of others’ experiences; such argumenta- tive discussions can be found, for instance, on product reviews websites, political blogs, in patient advocacy and support group discussions, and in brief anecdotes shared via microblogs. There are a variety of common argument structures [134]. A single premise may directly support a conclu- sion (as in (i) of Figure 1 on the facing page), but more commonly, they are combined to come to a conclusion. Schneider et al. / A Review of Argumentation for the Social Semantic Web 3 ” a 77 B| cinghe nked one sem divergent Fig. 1. Common argument patterns, from [134]. Premises and conclusions may also be chained (as in (iv) of Figure 1 on the next page). Further, as we will see in this review, there are differ- ent ways of thinking about and modeling arguments. Argumentation theorists have variously modeled indi- vidual argument structure (e.g. Toulmin, discussed in Section 4.1, page 4), argument chaining (e.g. Arau- caria* [150,142,144]), and groups of arguments (e.g. Dung, discussed in Section 4.5, page 6). There is a sig- nificant difference between what is possible to analyze when looking at these different levels of argumenta- tion: they show micro- and macro- structures which are not commensurate, so it is important to have clarity about what kind of analysis suits a situation. We believe that for social media, the basic units of argumentation are claims and justifications. By a claim, we mean an assertion of fact or opinion’. Justifications—reasons for believing the claim—are of- ten elicited when a claim is questioned. Some justifi- cations take the form of explanations; opinions may be elaborated upon and explained even when no disagree- ment is expected. 2.2. Social Semantic Web The interaction of users around the Web has been shifting from individual siloed Web systems, towards more open and interlinked social applications>. In dis- cussion environments, such interlinkage is particu- larly important: the same community may discuss top- ics across multiple sites, and use multiple types of sites, such as blogs and microblogs, discussion forums, and wikis. Crosslinking the discussions of these sys- tems is a first step, which has been taken by SIOC — Semantically-Interlinked Online Communities [26]. Shttp://araucaria.computing.dundee.ac.uk/ 44 propositional commitment, in Walton and Krabbe’s terminol- ogy [198] Shttp://oreilly.com/web2/archive/ what-is-web-20.html Yet the internal structure of these discussions — such as whether the participants agree or disagree, are con- tributing diverse ideas, or debating in circles — is still not represented in SIOC. Capturing such underlying arguments would be valuable, and research is begin- ning to address this for instance by identifying argu- ment schemes used in Amazon reviews [75,208] and by modeling the speech acts in Twitter conversations [148]. Yet infrastructure for argumentation on the So- cial Semantic Web is still needed. 3. Requirements What are the requirements for supporting argu- mentation on the Social Semantic Web? Arguments must be identified, resolved, represented and stored, queried, and presented to users. Identification involves mining arguments, in the form of claims, from text (Section 6.12.2, page 17), eliciting them from users, or some combination of these approaches. Resolving involves indicating the relationships between the in- dividual claims that make up arguments: are they on the same topic? Do they agree or disagree? Represent- ing and Storing arguments requires a suitable ontol- ogy to represent claims and the relationships between them. This supports Querying and enables Presenting the Social Semantic Argument Web, i.e. using these ontologies to facilitate access to conversations, sum- marizing the contentious and agreed-upon points of a discussion. The representations chosen are key to this process, since they determine what stored information can be retrieved, and what information needs to be mined and resolved. Existing representations will need to be aug- mented, since the information we can retrieve depends on what information we store. The desired ontologies should encompass not only the structural features of posts (such as the date and author of a post) and of conversations (such as the reply structure of multi- ple posts), but also additional argumentative features, for instance to mark claims and to indicate the rela- tionships between them. The simplest relationships for representing argumentation indicate whether pairs of claims support or challenge each other. Yet in gen- eral, these relationships do not just pertain to pairs; in general, entire groups of argumentative messages may need to be considered together. The meaning of a dia- logue may be lost by chunking messages and treating them individually, out of the original context. 4 Schneider et al. / A Review of Argumentation for the Social Semantic Web Even simple scenarios may give rise to complex argumentation involving chains of statements (e.g. [178]), and context-dependent relationships in which the conclusion of one argument is premise of another [209]: this makes the graph structures of the Seman- tic Web a natural fit. Wyner et al. suggest that be- sides agreement and disagreement, the semantic types of arguments should at least include introduction of a premise or exception, refinement, and pronomial anaphora and call for a modular architecture “where different relationships or debate components may be added systematically” [209]. 3.1. Example Applications & Requirements We envision two main approaches to studying argu- mentation on the Social Semantic Web: 1. Focusing on the real-time, dialogical nature of the Social Web, i.e. by soliciting arguments from humans through conversation and real-time ex- change. 2. Focusing on the Social Web as a source of ar- tifacts, ie. by using existing natural language conversations and reconfiguring the traces and archives of these conversations. Examples of the first case would be a chatbot or an interactive webform; these could help populate a knowledge base or enable argumentative interaction between humans and intelligent agents. Examples of the second case would be discussion summaries or in- teractive conversation browsers; a discussion summary could highlight the agreement and disagreement about a topic expressed in a number of Social Web sources, or a review browser could enable faceted navigation through reviews based on the factors they mention, and the polarity and strength of the reviewer’s perspective on each such factor. Formal semantics will be needed in both cases, but for different functions. In the first case, formal semantics translate from natural language to agent- appropriate vocabularies, potentially enabling reason- ing over human input. Argumentation has long been used for planning between agents: agent-based ap- proaches to the Semantic Web are common [179], and there has been some work in mediating between hu- mans and agents [169,202]. In the second case, the se- mantics will mainly be useful for presenting humans with visualizations and overviews, therefore the ease of mining and presenting representations, and the suit- ability for human understanding, should be preferred. Another factor is how—by what process and agent— arguments are translated into the formal semantics. This may be the responsibility of a human or machine. If argumentation is annotated by humans-either the person posting a comment or other individuals—they will need sufficient understanding of the model as well as a suitable incentive or motivation for annotating. Meanwhile, machine-based (algorithmic) annotation is limited by our current understanding of informal ar- gumentation, and by the multilayered meaning of con- versations. Further, argumentation can be treated as static (for completed discussions) or dynamic (for on- going conversations). If participants annotate the dis- cussion, the requirement to annotate can distract from the discussion; yet non-participants and machines may be limited by lack of awareness of the context and of subtle language cues. Models supporting human navigation (e.g. to sup- port our second example above) should make psycho- logical sense; this is not a factor for internal represen- tations for machines. In either case, to ensure feasibil- ity (for either human or algorithmic entry), argumenta- tion classifications need to be clearcut. Thus, the gran- ularity of the model must be limited. 4. Theoretical Models of Argumentation This section discusses fourteen theoretical models. Seven are designed for capturing argument structure: Toulmin [180], Issue-Based Information Systems [96], Walton’s argumentation schemes and critical ques- tions [197], Walton and Krabbe’s dialogue types [198], Dung’s Argumentation Frameworks [51], Value-based Argumentation Frameworks [15], and Factor Analysis and Dimension Analysis [16]. An additional seven lin- guistic approaches deal with issues relevant to argu- ment structure or detection: Speech Act Theory [158], Language/Action Perspective [205], Pragma-dialectic [186], Metadiscourse and Structural Elements of Text [79], Rhetorical Structure Theory [110], Coherence [91], and Cognitive Coherence Relations [152]. 4.1. Toulmin The study of informal argumentation originated in philosophy in 1958 with Toulmin [180]. Toulmin sought to find a common underlying basis for argu- ments in every field of human activity. His model applies, for instance, to legal, scientific, and infor- mal conversational arguments. In Toulmin’s theory, Schneider et al. / A Review of Argumentation for the Social Semantic Web 5 evidence and rules called Warrants support Claims. Claims may also be qualified (i.e. with constraints or to indicate uncertainty); Rebuttals may be used to ar- gue against an argument. Toulmin’s argument pattern is shown in Figure 2: Data is supported by Warrants which have Backings, showing that a Claim holds with Qualifiers regarding the situation, unless there is a Re- buttal. Figure 3 shows Toulmin’s now-famous argu- ment, presented according to this structure. [QUALIFIER] a “ DATA | | CLAIM WARRANT | REBUT | Fig. 2. An interpretation of Toulmin’s argument pattern, from [29]. Harry was born) in Bermuda | [ Harry isa Sc sumably. os = — +0, pre ie bly, | British subject Since Unless A man be Both his parents were Bermuda aliens/ he has become a generally be a naturalised American/ .., British su t pect On account of The following statutes and other legal provisions: Fig. 3. Toulmin’s example argument from page 105 of [180]. 4.2. Issue-Based Information System (IBIS) IBIS, Issue-Based Information System, is a problem- solving structure first published in 1970 [96]. As the name suggests, IBIS centers around controversial is- sues which take the form of questions. Specialists from different fields may use the same words with differ- ent assumptions and intentions®, hampering commu- nication. IBIS is especially intended to support com- munity and political decision-making. In this scenario, there may be three separate groups—the participants in the discussion, the relevant experts, and the decision 6“Many central terms used are proper names for long stories spe- cific of the particular situation, with their meaning depending very sensitively on the context in which they are used." [96] makers—each of whom need to communicate with each other and who must also get information from existing records and documentation. IBIS, as originally designed, is a documentation sys- tem, meant to organize discussion and allow subse- quent understanding of the decision taken; this ex- plains the use of “Information System” in its acronym. The context of the discussion is a discourse about a topic. Issues may bring up questions of fact and be discussed in arguments. Hete, “Arguments are con- structed in defense of or against the different positions until the issue is settled by convincing the opponents or decided by a formal decision procedure,” [96]. IBIS also recognizes model problems, such as cost-benefit models, that deal with whole classes of problems. Several kinds of relationships exist between is- sues: direct successor, generalization, relevant anal- ogy, compatible, consistent, or inconsistent. The method also distinguishes issue content, as factual, deonic (“Shall X become the case?’”), explanatory, or instru- mental (“Shall we take approach X to accomplish Y?"). Originally implemented as a paper-based system, IBIS influenced several ontologies and numerous tools (see Section 6.2, page 12) as well as procedures such as dialogue mapping [47]. 4.3. Walton’s Argumentation Schemes and Critical Questions The Canadian philosopher Walton has written ex- tensively on argumentation for more than thirty years (e.g. [190,193,194]); a 2010 festschrift honoring his contributions [143] shows how his work has influenced and been applied to computer argumentation. Informal argumentation is one of Walton’s specialties [195], and in this section, we discuss two of his key theories, start- ing with argumentation schemes. According to Rahwan [134], while many taxonomies of argumentation have been proposed [129,63,185,85], Walton’s taxonomy [191] provides the point of depar- ture for computational models of argumentation. In his detailed classification from 1995 [191], Walton de- scribes each scheme with a name, a conclusion, a set of premises, and a set of critical questions. Critical ques- tions address the points where this argument scheme may break down, and suggest attacks against the argu- ment. For example, the following six critical questions 6 Schneider et al. / A Review of Argumentation for the Social Semantic Web are associated with the Argument from Expert Opinion [62]’: . How credible is £ as an expert source? . Is F an expert in the field that A is in? . Does £’s testimony imply A? . Is E reliable? . Is A consistent with the testimony of other ex- perts? 6. Is A supported by evidence? Walton’s 2008 book [197], coauthored with computa- tional argumentation researchers, presents 65 general argumentation schemes, presumably updating [191]. mM Bw N Ee 4.4. Walton and Krabbe’s Dialogue Types Discussion types, first developed by Walton and Krabbe [198], have also been influential. ® Seven types of dialogue are shown in Figure 4 on the next page. These types are Persuasion, Inquiry, Discovery, Negotiation, Information-Seeking, Deliber- ation, and Eristic. They are distinguished by the ini- tial situation, the individual goals of the participants, and the overall goal of the dialogue. For instance, an information-seeking dialogue and an inquiry have sim- ilar goals, but differ in the initial situation: one per- son is believed to have the answer in an information- seeking dialogue, while in an inquiry, no one has the answer. Persuasion and deliberation are distinguished by whose preferences are used: in a persuasion dia- logue, the outcome depends on the preferences of the individual to be persuaded, while in a deliberation, the group preferences are used. Understanding the goal of a conversation is impor- tant for determining the outcome, and for determining what conversational moves are relevant. We may be able to access to pragmatics [92] of a conversation by understanding its goals. This is also how we evaluate a conversation: What was A trying to achieve? What was B trying to achieve? Did they achieve it? In our own view, these types of dialogue can be classified based on whether knowledge plays a 7[62] attributes this to page 49, D. Walton, Appeal to Expert Opin- ion, Penn State Press, University Park, 1997. 8 walton has revised this taxonomy several times. ‘Discovery’ was not in several earlier formulations, such as [193](p. 183); it is moti- vated by choosing the best hypothesis for testing. Debate and Peda- gogical appeared in an earlier formulation [192] which provides de- scriptions of the goals of each dialogue. Other scholars have also suggested extensions and modifications, for instance Dunne et al. have proposed adding examination dialogues [52]. large, middling, or minor role. Inquiry, Discovery, and Information-seeking dialogues are almost entirely knowledge-based, while knowledge plays only a mi- nor role in Negotiation (aiming at a harmonious set- tlement) and Eristic (quarrels, beneficial mainly for venting emotions). Knowledge plays some role in the remaining two types: in Persuasion and Deliberation, opinion and belief also have a large role. Further com- plexity arises because dialogue types may shift in an actual discussion, and argument schemes may be em- bedded in one another [196]. 4.5. Dung’s Argumentation Frameworks Dung provides a powerful graphical model of ar- gumentation frameworks in [51], which has been widely used in computational argumentation. Argu- mentation frameworks are defined as sets of arguments and attacks between them. Formally, an argumentation framework is a pair AF = (AR, attacks) where AR is a set of arguments, and attacks is a binary relation on AR, i.e. attacks C AR x AR. Fig. 5. Example of an argumentation framework Then the questions of interest are to find maximal sets of arguments that do not attack each other (these are called conflict-free), and to find arguments that are not defeated by a given set of arguments (these are called acceptable). A conflict-free set of arguments is then considered to be admissible if each argument is acceptable with respect to the set. Dung then finds maximal admissible sets, known as preferred extensions. Also important are the grounded extensions, which represent the least fixed point of the function mapping an argumentation framework to the Schneider et al. / A Review of Argumentation for the Social Semantic Web 7 Type of Initial Situation Participant’s Goal Goal Of Dialogue Dialogue Persuasion Conflict of Opinions Persuade Other Party | Resolve or Clarify Issue Inquiry Need to Have Proof Find and Verify | Prove (Disprove) Evidence | Hypothesis Discovery | Need to Find an Find and Defend a Choose Best Hypothesis Explanation of Facts Suitable Hypothesis for Testing Negotiation Conflict of Interests ~ Get What You Most Reasonable Settlement Want | Both Can Live With Information- Need Information - Acquire or Give | Exchan ge Information | Seeking Information Deliberation Dilemma or Practical, Co-ordinate Goals and Decide Best Available . Choice Actions Course of Action Eristic Personal Conflict Verbally Hit Out at Reveal Deeper Basis of | Opponent Conflict Fig. 4. Walton’s seven types of dialogue, from [196]. set of acceptable arguments of that framework. A sta- ble extension is a conflict-free set of arguments that at- tacks each argument that does not belong to the set. A simple argumentation framework is shown in Fig- ure 5. In this example, A and B attack each other; A also attacks C; and D is not attacked. Thus A,D and B,C,D are preferred extensions. In Dung’s theory, there is a notion of ‘defend’ — to defeat the attackers — but there is no direct notion of ‘support’. Arguments ‘support’ another one by not be- ing defeated, and by not attacking, a given argument. 4.6. Value-based Argumentation Frameworks Value-based Argumentation Frameworks [15], based on Dung’s argumentation frameworks, address persua- sion and preferences. It is not just differences about the facts, or failures in logic that can cause reason- able people to disagree: differences in values can also be to blame. For instance: “Despite the fact that the weather is beautiful, I choose to stay inside, because I have something important to do". In practical rea- soning, two people can come to different, consistently logical opinions, based on a difference of values: “A key element in persuasion is identifying the value con- flict at the root of the disagreement so that preference between values can explicitly inform the acceptance or rejection of the competing arguments." The the- ory of Value-based Argumentation Frameworks thus draws from Perelman’s notion of audience [129]: ar- guments are often addressed to particular audiences, and persuasive arguments are those aligned with the audience’s values and preferences. Value-based Argu- mentation Frameworks provide a method for logically calculating consistent approaches, distinguishing be- tween the facts of a situation and community mem- bers’ values. 4.7. Factor and Dimension Analysis There is a large business market in legal information retrieval, and one method for classifying and indexing legal cases has been the key factors of a case, for in- stance for the early legal argumentation system CATO [7]. Factor analysis can be helpful in supporting com- munity decision-making or in summarizing reviews. Factors are simplifications that are either present or ab- sent; when present, a factor “always strengthens the case for the same disputant" [16]. Further development along the lines of CATO yielded Ashley’s later system, HYPO [11], which uses dimensions rather than factors. Dimensions “capture the legal relevance of a cluster of facts to the merits of a claim". Dimensions such as “Obligation to aid the victim" or “Failure to heed traffic signs" contribute to determinations of culpability, and have been recorded in manually constructed databases [12]. Dimensions 8 Schneider et al. / A Review of Argumentation for the Social Semantic Web can be present to a greater or lesser degree and it may unclear which side they favor. 4.8. Speech Act Theory Several approaches to conversation and argumenta- tion have been derived from speech act theory. Searle’s Speech Act Theory [158] describes five categories of speech acts: assertives, directives, commissives, ex- pressives, and declaratives. Speech acts are about the force of a statement: what effect they seek to have on the hearer or the world. Assertives (‘The sky is blue’) assert that something is true. Directives (‘Clean your room’) order, permit, or request something. Commis- sives are vows or pledges (‘I swear to tell the truth’). Expressives offer thanks or congratulations, or express feelings (‘Great work!’). Declarations (‘I now pro- nounce you man and wife’) enact what they say, ef- fectively changing reality.” Speech act theory is not a complete model of argumentation, yet it is a relevant theory that has been widely influential. For example, earlier (Section 4.4, page 6) we discussed that under- standing the goals of a conversation could be important for interpreting it; the same speech acts can be used in different ways, depending on the goals of a dialogue 10 4.9. Language/Action Perspective The Language/Action Perspective (LAP) [205] em- beds Speech Act Theory in a task-based framework. Argumentation is found in each of the three types of conversations which accomplish goals in the Lan- guage/Action Perspective, according to de Moor and Aakhus: Conversations for action involve making commitments; conversations for possibility create a context for action; and conversations for disclosure al- low participants to share their views and concerns [49]. 4.10. Pragma-dialectic The pragma-dialectic approach is a complete argu- mentative theory, which has been developed over a number of years in numerous scholarly works (espe- cially [184,185,186]) and popularized in the authors’ textbooks (e.g. [187]). Like the Language/Action ° As with all speech acts, sincerity is a criterion, and social criteria, e.g. ceremony, may also hold. !0For a pertinent example see the persuasion and deliberation sce- narios discussed in Atkinson et al. [13]. Perspective, it uses speech acts, further developing Searle’s theory in order to model argumentation. Rather than focusing on the logical forms and patterns of reasoning, as Walton does, van Eemeren and Groo- tendorst’s pragma-dialectic theory views argumenta- tion as a social process, used to settle “a difference of opinion by verbal means” [187] (pp. ix-xii). I Confrontation Assertive Expressing a standpoint Commissive Acceptance or non-acceptance of a standpoint> upholding non-acceptance of a standpoint [Directive Requesting a usage declarative] [Usage declarative] Definition, specification, amplification, etc.] af Opening Directive Challenging to defend a standpoint Commissive Acceptance of the challenge to defend a standpoint Agreement on premises and discussion rules Decision to start a discussion [Directive Requesting a usage declarative] [Usage declarative] Definition, specification, amplification, etc.] mm Argumentation Directive Requesting argumentation Assertive Advancing argumentation Commissive Acceptance or non-acceptance of argumentation [Directive Requesting a usage declarative] [Usage declarative] Definition, specification, amplification, etc.] Iv Concluding Commissive Acceptance or non-acceptance of a standpoint Assertive Upholding or retracting a standpoint Establishing the result of the discussion [Directive Requesting a usage declarative] [Usage declarative] Definition, specification, amplification, etc.] Fig. 6. Distribution of speech acts in a critical distribution from [186]. Depending on the context, the same speech acts can function as an explanation, a piece of information, or as an argumentation: for argumentation, the context must include a difference of opinion. Depending on their order and position in the discussion, speech acts take on different meanings, as we will see in Figure 6. Further, Searle’s illocutionary speech acts, which func- tion at the sentence-level, combine in argumentation into a higher-order textual element: the argumentation is itself seen as a complex speech act. The main speech acts within an argumentation are assertives, commissives, and directives. Expressives— which express emotions—do not help resolve the differ- ence of opinion, but may affect how or whether the dis- cussion proceeds. Declaratives—which bring a state of affairs into being—are relevant for definitions, specifi- cations, amplifications, and explanations; van Eemeren and Grootendorst call these “usage declaratives”. The pragma-dialectic approach also stresses the principles of clarity, honesty, efficiency, and relevance, Schneider et al. / A Review of Argumentation for the Social Semantic Web 9 updating Grice’s Cooperation Principle [64]—which fo- cuses on the intention of language—with the Searlean focus on the communicative aspects of language use. Relevance, for example, can be global, local, subject matter-specific, or probative. An argument may be rel- evant at one phase, but irrelevant at another; for exam- ple an argument related to selecting the topic of dis- cussion is not relevant once the topic has been agreed upon. To understand the force of a speech act—-whether an assertive, commissive, or directive—, we must iden- tify where we are in the argumentation. van Eemeren and Grootendorst identify four dialectical stages of ar- gumentation: Confrontation, Opening, Argumentation, and Closing [186]. In the confrontation stage, the issue at hand is announced, agreed upon, or clarified. In the opening stages, the rules are agreed to (perhaps implic- itly). In the main stage (Argumentation), each party is expected to make a serious effort to support his point of view, while also allowing the other party to make his case. Finally, the argument closes when the goal is fulfilled or the parties agree to end the debate. The pragma-dialectic approach is far more exten- sive in attending not only to linguistic patterns, but also to the social context in which they are embedded (i.e. the linguistic pragmatics, see e.g. [92]). Argumenta- tive discourse starts with the assumption that the lis- tener does not (necessarily) agree with the speaker’s position, and aims at “coming to a reasonable agree- ment” [187], p 4. Speakers may anticipate objections, explaining their reasoning to account for expected (im- plicit) differences of opinion. Or, they may wait to hear their conversational partners’ standpoints or doubts, and then respond. The issue, according to the pragma-dialectic the- ory, may be single or multiple, and mixed or non- mixed. Single disagreement is about just one proposi- tion while multiple disagreement is about more than one proposition. If both a positive and a negative stand- point are taken on the issue, the disagreement is mixed, otherwise it is simple. This is a particularly useful dis- tinction for conversations in social media. Pragma-dialectic is also particularly useful for de- termining which parts of social media discourse can be considered argumentative, since it presents phrases that tend to mark argumentation, and since it treats speech acts from an argumentative perspective. For ex- ample, doubt is often implicit, but certain phrases mark it more explicitly, such as “I don’t know whether", “T’m not yet convinced that”, “Couldn’t it be that”, and “T’ll have to think about whether”. Similarly, set phrases often indicate the topics of a debate—-helping us detect a participant’s standpoint, which “expresses a certain positive or negative posi- tion with respect to a proposition" [186](p. 3). Stand- points may often be obliquely stated, yet they can sometimes be recognized by the appearance of partic- ular phrases, which can be baldly stated (“my stand- point is that”, “we are of the opinion that”) or explicit (“I think that”, “if you ask me”, “therefore”). Some phrases that may be used to indicate a standpoint also 27 468 permit alternate interpretations (“the way I see it”, “in other words”, “all things considered”). Other patterns, like “shouldn’t”, “you must never’, “‘that...is”, “ought to be”, commonly co-occur when a standpoint is ex- pressed [187](pp. 10-12). Such phrases, like the metadiscourse markers we discuss next, can potentially be applied in detecting and mining argumentation (which we later discuss in Section 6.12.2, page 17). However, phrase-detection alone does not suffice, and significant challenges re- main. Language is multivalent and context is needed for properly reconstructing argumentation. “In lan- guage use there is often the case that there is more than one purpose at the same time, and if language is used argumentatively, the argumentative function need not always be the most important," [186](p. 23), meaning that reconstructing argumentation must extract a single thread of meaning out of many. 4.11. Linguistic Markers of Argumentation: Metadiscourse and Structural Elements of Text We next discuss linguistic markers of argumenta- tion; strictly speaking these are not theoretical models of argumentation, yet they identify possible argumen- tation, essentially enabling the argumentation struc- tures to be annotated in and abstracted from the text. Metadiscourse refers to the “aspects of a text which explicitly organize a discourse or the writer’s stance to- wards either its content or the reader.” [79](p. 14). Ar- gumentative words and phrases such as ‘but’ and ‘ac- cording to X’ are prominent examples. Metadiscourse is used not only to structure text but also to influence the reader’s view. Hyland classifies metadiscourse into interactive and interactional types. Interactive metadiscourse, which organizes text, includes transitions (in addition, but, thus, and); frame markers (finally, to conclude, my purpose is); endorphoric markers (noted above, see Fig, in section 2); evidentials (according to X, Z states); and code glosses (namely, e.g., such as, in 10 Schneider et al. / A Review of Argumentation for the Social Semantic Web other words) [79](p. 49). Interactional metadiscourse, which makes the author’s views explicit and invites readers’ response, includes hedges (might, perhaps, possible, about); boosters (in fact, definitely, it is clear that), attitude markers (unfortunately, I agree, surpris- ingly); self mentions (J, we, my, me, our); and engage- ment markers (consider, note, you can see that). As markers of the persuasive and rhetorical ele- ments of texts, metadiscourse elements are likely to be useful signals for identifying claims and arguments in social media. 4.12. Rhetorical Structure Theory Rhetorical Structure Theory (RST) [110], a method for analyzing texts according to their structure and rhetorical role, was developed at the University of Southern California’s Information Sciences Institute to assist with computer-based text generation. In RST, structures such as ‘Concession’, ‘Evidence’, and ‘Jus- tify’, called ‘relations’, describe the relationship of two or more spans of text. Generally one span (the most important) is called the nucleus, while the less impor- tant spans are known as satellites. In some situations (such as sequences and contrasts), both spans are nu- clei of equal weight. Justifications and hedges are more likely to appear in satellites while the nucleus is more likely to contain claims; this has potential application in detecting arguments and in summarizing social web applications. 4.13. Coherence Coherence is another important concept in text and dialogue. Coherence is not an argumentation theory per se, but it is both an essential part of text, and an essential part of argumentation, since before an argu- ment can be understood (much less formally evalu- ated), how its parts hold together or interrelate must be understood. Knott compares the following two exam- ples [91]. 1. “Tim must love that Belgian beer. The crate in the hall is already half empty." 2. “Tim must love that Belgian beer. He’s six foot tall." While the first example is coherent the second exam- ple is more challenging to make sense of: the reader expects (but does not get) a sensible explanation or ev- idence for why Tim must love that Belgian beer. Argumentation relies on coherence: Adding ‘be- cause’ works in the first example but not in the sec- ond example. The word ‘because’ stresses the ex- pected causal relationship, making the informal argu- ment more evident. 3. “Tim must love that Belgian beer, because the crate in the hall is already half empty." In Sentence 3, the reader must still infer some informa- tion, such as that the crate in the hall contains Belgian beer, and that Tim is the main person drinking the con- tents of the crate. Such missing premises are typical in informal argumentation. 4.14. Cognitive Coherence Relations One actionable way of expressing coherence is by using specific signaling terms. The causal relationship (expressed in ‘because’) is one of the Cognitive Co- herence Relations which Sanders uses to explain how readers understand text [152]. The four Cognitive Co- herence Relations are: Basic Operation (causal or ad- ditive), Source of Coherence (semantic or pragmatic), Polarity (positive or negative), Order of Segments (for causal relations only: basic or non-basic, depending on whether or not the antecedent appears before the con- sequent). Based on this, the relationship signaled in be- cause in Sentence 3 (above) is causal, pragmatic, pos- itive, and basic. 5. Comparison of Theoretical Models Any of these models could be expressed in se- mantic formats (e.g. RDF) since they are compatible with a graph-based representation of argumentation. Yet for modeling argumentation on the Social Seman- tic Web, it is most meaningful to examine the chal- lenges and opportunities that might advantage any one model or framework over the others. As previously noted, various units have been modelled—individual argument structures (e.g. Toulmin, discussed in Sec- tion 4.1, page 4), argument chaining (e.g. Araucaria!! [150,142,144]), and groups of arguments (e.g. Dung, discussed in Section 4.5, page 6). We can make several further distinctions between models, for instance based on the community in which they originated, their purpose or use, the extent to which they focus on disagreement, the unit of analysis “http: //araucaria.computing.dundee.ac.uk/ Schneider et al. / A Review of Argumentation for the Social Semantic Web ll on which they focus, their granularity, and their suit- ability for automation or for aiding human reasoning. 5.1. Community of Origin Various communities have contributed models, par- ticularly the argumentation and linguistics communi- ties. The IBIS model comes from management and was later taken up by design rationale and human- computer interaction (HCI) communities. The Lan- guage/Action Perspective originated in artificial intel- ligence and HCI and was later adopted by communi- cation theorists. In some cases, models bridge commu- nities: the pragma-dialectic approach is an argumenta- tion model which has been heavily influenced by lin- guistics, and by the theory of pragmatics [92] in par- ticular. Models have been shaped by their originators and proponents, and the purposes for which they were intended. 5.2. Purpose or Use The intended purpose for each model depends largely on its origin. Models put forth by the argu- mentation community are generally designed to sup- port either analysis (e.g. to determine the reasoning patterns used and to identify fallacies) or formal rea- soning, in order to address questions such as compu- tational decisions of which argument won, what the deciding factors were, or what values and preferences were expressed in the discussion. Models of linguistic features may be used in discourse analysis, for summa- rization, and to support natural language generation by both machines and non-native speakers. Models from other communities are generally intended to support flow-based process analysis, for instance to organize information in order to avoid information overload, to speed human decision-making, and to provide a record of collaborative thought processes. 5.3. Agreement/Disagreement Focus Disagreement and the process of coming to con- sensus are the core of argumentation. While disagree- ment and agreement are central in models coming from the argumentation community, other models focus on this core to a greater or lesser extent. Most linguistic models are considerably broader and less focused on the argumentation aspects, yet in addressing conversa- tion, they provide valuable insights as well as analy- sis tools. HCI models focus on supporting collabora- tion and shared visions; disagreement is analyzed or understood only to the extent necessary for coming to consensus or providing an overview of viewpoints. 5.4. Unit of Analysis Different units of analysis have been used. At the language layer, models may focus on the relation- ship between different clauses (Metadiscourse, Coher- ence, Cognitive Coherence, RST) or the communica- tive function of different words, phrases, and sentences (Speech Act Theory and the Language/Action Per- spective). The pragma-dialectic approach focuses on the propositional level, while factors analysis looks at important attributes or dimensions. Other models focus on classifying individual arguments and their relation to a whole (IBIS), or studying the internal structure of arguments (Toulmin, Walton, Argumenta- tion Frameworks, Value-based Argumentation Frame- works). 5.5. Granularity Models of linguistic features are more granular, but sometimes less focused on the overall structure. Coarse-grained and simple models, such as IBIS, more common in application. Yet even IBIS is not gener- ally applied in its full complexity, but is rather reduced to focusing identifying issues, and then on identifying pros and cons for a particular issue. Fuller versions be- come more complex by looking at the relationships be- tween arguments: what responds to what. 5.6. Ease of Automated Application Mechanistic application is possible for some mod- els but not for others. In particular, classification for Walton’s model would be quite difficult due to the large number (65) of argument categories and the need for detailed reasoning. On the hand, in many cases language technologies can be mechanically analyzed. Identification and classification of argumentation via language technologies is still in its infancy, yet offers great potential to expand algorithmic understanding of language. 5.7. Support for Human Reasoning To aid human reasoning, however, linguistic models that mainly use cue words are probably too granular since they occur at the sentence level, probably more 12 Schneider et al. / A Review of Argumentation for the Social Semantic Web granular than needed. For this purpose, Walton’s crit- ical questions are very useful, because they can point humans to the questions that need to be addressed, opening the door to checklists for reasoning, which could be applied consistently by groups. Value-based frameworks also address the basic reasoning underly- ing social decisions: each person has their own rea- sons, which may be revealed in the course of a discus- sion. Focusing on what the values are, and being able to articulate them, can help both in developing em- pathy for dissenting viewpoints, and in making clear the rationale for group decisions when consensus is needed. 6. Applications of the Models to the Social and Semantic Web The theoretical models discussed have been quite in- fluential, and in many cases we can directly trace So- cial and Semantic Web applications of these models. In this section we describing some applications, review- ing each of the models in turn. 6.1. Applications of Toulmin Toulmin is cited frequently and in numerous fields, from rhetoric (e.g. [199]) to education (e.g. [37]) to computer argumentation (e.g. [103]). While his model is a useful abstraction, scholars have argued about whether people actually think in terms of Toulmin’s warrants [121]. One early hypertext system, SEPIA, drew from the Toulmin system [171]. A modified version of Toul- min’s argumentation schemes have been used to de- scribe a cooperative dialogue game, implemented in Prolog, in which the participants’ goal is to reach a claim on which they agree, while also producing a supporting argumentation structure for the claim [14]. Such dialogue games could be modified for use in mixed-initiative systems. In the Semantic Web, the Toulmin Argument Model has been implemented by an OWL 2 DL ontology that imports CiTO!? [130]. It follows Toulmin’s model closely, as shown in Figure 7. Pnttp: //www.essepuntato.it/2011/02/ argumentmodel =I Prefixes amo http: //www.essepuntato.it/2011/02/argumentmodel/ [=] Legend| Kinds of entity eeesasEt [ones fmeenmewe{ sum] |e Evidence amo:proves Claim Qualifier a amo:supports amo:leadsTo amozisValid Unless Fig. 7. The core Toulmin argument model ontology [130]. 6.2. Applications of Issue-Based Information System (IBIS) Although many tools are described as ‘using the IBIS model’ or ‘IBIS-like’, there is significant varia- tion in the underlying structure of these models [82]. In our view, these models use ‘IBIS-like’ to mean that they concern decision-making or design rationale, pro- vide graphical representations, and use some form of polarity. The IBIS model has a long history of use, partic- ularly with early hypertext systems. Early critiques of IBIS came from the design rationale community. One difficulty was that only deliberated issues were included; Procedural Hierarchy of Issues (PHI) mod- ifies IBIS to allow inclusion of subissues which are not deliberated [56]. PHI was adopted by another early system, the Author’s Argumentation Assistant [157], which also drew from the authors’ earlier Toulmin- based system, SEPIA [171]. Another difficulty, representing the relationships and interdependencies of issues [56], remains chal- lenging to resolve, though ideas such as nonfunctional requirements and dependencies (Section 7.6, page 20), might be relevant. IBIS has also been used outside of design ratio- nale. For instance, Gerosa et al. [59] discuss an e- learning message board system adopting a modifica- tion of IBIS, where message types are specified. In addition to the IBIS-analogues, Question, Argumen- tation, and Counter-Argumentation, the system adds two types: Seminar (a general topic for the week) and Clarification. IBIS has also influenced the design of modern ontologies, including the SALT Rhetorical Ontology, SIOC-Argumentation, DILIGENT, and the Schneider et al. / A Review of Argumentation for the Social Semantic Web 13 Change Ontology which we discuss after reviewing IBIS’ RDF representation. 6.2.1. IBIS RDF IBIS RDF® is an RDF representation of the IBIS model. refersTo is modelled as a subProperty of dcterms : reference with two subproperties, pro and con. The larger IBIS vocabulary provides Pub- lished Subject Indicators!+ for important terms, in- cluding pro, con, Idea, Question, Argument, Decision, and Reference. 6.2.2. SALT Rhetorical Ontology SALT [68] is a rhetorical ontology for scholarly communication. In SALT, opposing arguments can be connected together with the relation Semantically-Interlinked Online Communities [26] — a model that focuses on representing online communi- ties and the content shared within them. While SIOC simply focuses on the notion of replies (sioc:reply_to) to represent connections be- tween discussion items, the SIOC-Argumentation mod- ule goes further and provides finer-grained representa- tion of discussions and argumentations in online com- munities. So far a modification of SIOC that draws from DILI- GENT and OMDoc has been used in the math wiki system SWiM'® [99], a Semantic Wiki for Mathemati- cal Knowledge Management. The SIOC/OMDoc argu- mentation ontology (Figure 9 on the following page) is further described in Lange’s dissertation [97]. It in- hasCounterArgument, while a RhetoricalElementcorporates IBIS-style classes from SIOC (Position, can also be connected with what it argues for (Argument and hasArgument, for instance). SALT’s argu- mentation also includes Reason, which contains Argument (further specified to PositiveArgument or NegativeArgument) and CounterArgument. 6.2.3. DILIGENT ontology DILIGENT provides an argumentative structure for collaborative ontology construction; the acronym comes from the phrase Distributed, Loosely-controlled and evolvInG Engineering processes of oNTologies. DILIGENT draws from both RST (Section 4.12, page 10) and IBIS (Section 4.2, page 5), as shown in Fig- ure 8 on the next page. 6.2.4. Change Ontology (ChAO) in Collaborative Protégé DILIGENT [36] itself influenced the Change On- tology in Collaborative Protégé. Castro et al. distin- guish between an argument (which is well-focused and specific) and an elaboration (which provides support for the argument, possibly with file attachments) [36]. Positions become clear through the dispute-resolution process. With Protégé, various argumentation-related Annotations can be added, including Explanation, Proposal, and AgreeDisagreeVote [156]. 6.2.5. SIOC-Argumentation The SIOC-Argumentation!> model [98] expands on the IBIS model with terms such as Decision and Argument. It is provided as an extension of SIOC — Bhttp://purl.org/ibis 4nttp://www.topicmaps.org/xtm/index.html# def-published-subject-indicator 'Shttp://rdfs.org/sioc/argument Decision, Idea, and Issue), as well as domain- specific argumentation classes for math (e.g. Wrong, Keep_as_Bad_Example, Incomprehensible). As opposed to IBIS-RDF, SIOC-Argumentation (Figure 10 on page 15) provide the means to easily integrate argumentation modelling patterns with So- cial Web applications since it relies on SIOC, already used in various applications (Drupal7, etc.). However, SIOC-Argumentation has limitations: it does not rep- resent taxonomic, causal, or similarity relations, which prevents its use in contexts that require deeper analysis of arguments. 6.2.6. SWAN-SIOC SWAN-SIOC [128] harmonizes the argumentation aspects of two pre-existing ontologies: along with SIOC, it is based on SWAN — Semantic Web Appli- cations in Neuromedicine [44] — an ontology which focuses on scientific communication in neurology. SWAN/SIOC uses twelve terms, as shown in Fig- ure 11 on page 15. The most general term is relatedTo, which has five direct descendents or subterms. These, in turn, may have subterms, until we reach the base terms in the ontology: disagreesWith, agreesWith, and discusses. SWAN/SIOC provides a simple model for the relationships between items. Tools using SWAN-SIOC include PDOnline, which is discussed in Section A.31, page xiv. 6.3. Applications of Walton Walton’s model has been widely applied in compu- tational argumentation [143], and recent research has lSnttp://kwarc.info/projects/swim/demo. html Schneider et al. / A Review of Argumentation for the Social Semantic Web oborstesnissue Decision ksue hosStotus: onkssue mabedksue: ‘otrenBy vifsLiteral rifs:Liferall withates Bboration Ader responses Tokssue} eo | ivenBy | DaiobaiE hf-Literal formalizesidea slike rifs:Literal \gtvenBy eatecha ristted SummortzesArgumentation ‘orgementsOn s | Argument f - >| Argumentafion 1: Position Challenge Justification Humon-Argumentotion || Machine-Aurgumentatton pmovidesTenct: rdfs Literal || tia: rifs-Ltteral | sgremont | [ Agroement | | Arora | Couniréxample| [Evista] | ample | Fig. 8. An overview of the core DILIGENT ontology from [174]. S10c angumarntatoe moxiuie ieertty shown] suppa Decision sunClassOq Cntology Entity aaa f oe ea SubClass Whang: Inappropriate | Incormprehensiote | for Domain Provide Example Keep as So as Fig. 9. The argumentation ontology from SWiM extends SIOC-Argumentation and DILIGENT Argumentation [99]. demonstrated how argument schemes could be used to aid sensemaking in Amazon reviews [75,74,208]. Avicenna and ArgDF incorporate Walton’s schemes. The only Social Web application we are aware of is Parmenides!’ [32,33,34], which uses the following ar- gumentation scheme and value-based argumentation frameworks [15]: Argumentation Scheme AS1: — Inthe current circumstances R, — we should perform action A, "http: //cgi.csc.liv.ac.uk/~parmenides/ — to achieve new circumstances S, — which will realize some goal G, — which will promote some value V. The following sixteen critical questions are associ- ated with Argumentation Scheme AS1: CQI Are the believed circumstances true? CQ2 Assuming the circumstances, does the action have the stated consequences? CQ3 Assuming the circumstances and that the action has the stated consequences, will the action bring about the desired goal? CQ4 Does the goal realize the value stated? Schneider et al. / A Review of Argumentation for the Social Semantic Web 15 agrees_withy disagrees_with! eutral_towands Fig. 10. An overview of the SIOC-Argumentation module from [98]. refersTo(4) incansistentvith coOnsistentvith relevantTo alternativeTo related Tol 5) disagreesWith inResponseTo(3) agreesWith anousadrrom discusses motwatecBy cites: Fig. 11. SWAN-SIOC ontology from [42]. CQ5 Are there alternative ways of realizing the same consequences? CQ6 Are there alternative ways of realizing the same goal? CQ7 Are there alternative ways of promoting the same value? CQ8 Does doing the action have a side effect which demotes the value? CQ9 Does doing the action have a side effect which demotes some other value? CQ10 Does doing the action promote some other value? CQI11 Does doing the action preclude some other ac- tion which would promote some other value? CQ12 Are the circumstances as described possible? CQI13 Is the action possible? CQ14 Are the consequences as described possible? CQI15 Can the desired goal be realized? CQ16 Is the value indeed a legitimate value? 6.4. Applications of Walton & Krabbe’s Dialogue Types Walton & Krabbe’s Dialogue Types’s have been influential. Persuasion, in particular, has been exten- sively studied [132], and the wider area of dialogue games has been an active area of agent argumentation [113]. In applications to human argumentation, some work on deliberation (e.g. [189] and the Parmenides system Section A.30, page xiii as a whole) has used dialogue type to focus the discussion. 6.5. Applications of Dung’s Argumentation Frameworks Dung’s argumentation frameworks have been in- credibly influential: as of 2011, the original 1995 paper has over 450 citations in the ACM Digital Library, and over 1500 in Google Scholar. A Java reasoner called Dungine implements Dung’s acceptability semantics [170]. Dungine is part of ArgKit!’, a Java 5 develop- ment toolkit for applications that use argumentation; it is open source, licened under LGPL. Dung’s approach has driven computational research in argumentation and provided the basis for a large body of theoretical work. Among theoretical exten- http: //argkit.org/ 16 Schneider et al. / A Review of Argumentation for the Social Semantic Web sions of Dung’s work, we have focused on Value-based Argumentation Frameworks, and the relevant Social Web application of Dung is in fact an application of that extension, as we next explain. 6.6. Applications of Value-based Argumentation Frameworks Parmenides uses value-based argumentation frame- works in addition to the argument scheme and criti- cal questions discussed above. It pinpoints the source of the disagreement, by having participants respond to a series of questions in a survey format. The group’s preferences are revealed in the results, which are dis- played to administrators as graphical argumentation frameworks. 6.7. Applications of Factor Analysis Factor analysis has been applied in Ashley’s legal argumentation systems, but we know of no specific So- cial Web applications. The factors analysis approach is still used by com- mercial providers of legal information [207]. More re- cently, automatic text mining has been used to iden- tify these factors [207]. Generalizing factors, perhaps using argument schemes and critical questions, could provide another approach to argument mining; see for instance Heras’ manual application of argument schemes to Amazon reviews [75] and Schneider’s fac- tors analysis of Wikipedia deletion discussions [155]. 6.8. Applications of Speech Act Theory Speech acts are used to model the flow of on- line conversation in several recent works. Jeong et al. [83] use semi-supervised machine learning to iden- tify speech acts in email and forum posts. Ritter et al. [148] model Twitter conversations with Speech Act Theory in combination with topic modelling and show a Speech Act transition map with probabilities for each state. One central use is in provenance in the Semantic Web. For assertions modelled in RDF, Carroll et al. [30] use the idea of performative warrants, to describe assertions made legitimately by the authority signing a Named Graph. 6.9. Applications of the Language/Action Perspective Using the Language/Action Perspective and draw- ing from Speech Act Theory, Twitchell et al. [181] model online conversations to classify them and create visual maps, used for information retrieval: “Using current search engines, the searcher could search for the words Vietnam, war, and critique. However, many critiques of the war might not con- tain the word critique, and would thus be lost (or receive a low ranking) in such a search. If the searcher was able to issue a query such as Viet- nam war (critique) where critique is the purpose of at least one participant in the conversation, she would likely get better results. The search for the semantic meaning of the words Vietnam war using conventional searching techniques would then be combined with the search for the pragmatic force of the word critique, yielding a search result with higher precision than searching on semantic meaning alone.” [181](bold, underline added). Attending to Speech Acts can also help predict decep- tion, which uses ‘fewer assertions and more expres- sives’ [181]. 6.10. Applications of Pragma-dialectics We are not aware of any argumentation tools spe- cific to pragma-dialectics. de Moor, however, has taken a reconstructive approach to argumentation based on pragma-dialectics (e.g. [49]). Pragma-dialectics has also been very influential in the argumentation community, integrated into e.g. Walton’s textbook descriptions of argumentation [195] and discussed in at least one journal special issue [25] and edited collection [183]. 6.11. Applications of Metadiscourse and Structural Elements of Text Annotating argumentation in natural language often takes advantage of detecting metadiscourse and docu- ment structure. Two particularly promising approaches come from Teufel and Sandor. Teufel’s rhetorical com- ponent extraction uses machine learning to extract and classify text according to its rhetorical status [175]. S4ndor’s concept-matching framework detects con- trasting ideas linguistically, using metadiscourse and rhetorical markers [6]. Rather than determining the re- lations between text spans, S4ndor uses her concept- matching framework to infer contrasts, novel informa- tion, etc. from the author’s metadiscourse [6]. Teufel and Moens focus on the document-level con- text, rather than the relationship between text spans. In Schneider et al. / A Review of Argumentation for the Social Semantic Web 17 their argument zoning, machine learning is used to ex- tract and classify text from academic articles accord- ing to its rhetorical status [175]. Sandor and Teufel and Moens provide contributions in risk assessment, an- notation, and audience- and task-specific summariza- tion. Reuse of their work has included an application to find rhetorical features of related work sections, first using classification algorithms, and then applying on- tologies [9]. However, these techniques are of particu- lar interest because of further work in argument mining drawing on these ideas, which we soon discuss (Sec- tion 6.12.2, page 17). 6.12. Applications of Rhetorical Structure Theory RST has been widely used for a variety of pur- poses and in 2006 a paper summarizing its applica- tions [173] was published. Recently, Mentis et al. [114] used RST to analyze group decision rationale, compar- ing new and established groups using relations such as ‘Interpretation & Evaluation’, ‘Evidence’, ‘Elabo- ration’, ‘Concession’, and ‘Antithesis’. Summarization research has frequently drawn upon RST [111,112]. Some further applications of RST and related ap- proaches are discussed in a recent survey of work on discourse structure [200]; argumentation might be found in several genres they discuss, such as in the es- say analysis and scoring and opinion mining applica- tions. 6.12.1. DILIGENT DILIGENT, briefly discussed above as an applica- tion of IBIS (Section 4.2, page 5), also draws from RST (Section 4.12, page 10) as shown in Figure 8 on page 14. To improve the agreement, clarity, and sat- isfaction [174] of discussions for ontology creation and refinement, DILIGENT restricts the argumenta- tion. Five argumentative relations — alternative, . . Lee : Ta: evaluation and justification, counterExamplé elaboration, and example — were drawn from RST [99], based on the arguments that advanced the ontology creation process in an experiment [131]. 6.12.2. Argumentation Mining Drawing on rhetorical parsing, argument mining is a new area of study which seeks to detect and ex- tract arguments from texts algorithmically. Mochales- Paulau’s dissertation work [116] focused on mining ar- guments [117,207,115] from the European Court of Human Rights and from the Araucaria annotated cor- pus!, based on Context Free Grammars [207] as well as techniques from Teufel and Moens. Earlier Grover et al. [66] adapted Teufel and Moens’ approach to determine the argumentative role of sentences drawn from a corpus of legal judgements. In “Automatic Argumentation Detection and its Role in Law and the Semantic Web" [117], Mochales- Paulau and Moens suggest that argument mining could contribute to the World Wide Argument Web [138], by extracting argument structures without human an- notation. As they point out, automatic argument detec- tion is needed at multiple levels: the inner structure of each argument as well as the overall structure of how multiple arguments are combined to contribute to the argumentative discourse. 6.13. Applications of Coherence Related notions of coherence are used in Thagard’s explanatory coherence theory, a computational theory “an explanatory hypothesis is accepted if it coheres better overall than its competitors” [176]. This theory has been applied to analyze scientific reasoning and le- gal trials. Another theory it influenced-that of cognitive co- herence relations, which we discuss next—has been in- fluential. 6.14. Applications of Cognitive Coherence Relations Cognitive Coherence Relations contributed to the development of ScholOnto [109]. In separate work, Mancini’s cinematic hypertext [108] used Cognitive Coherence Relations to develop a visual language for structuring hypertext links, increasing the coherence of argumentation conveyed in non-linear hypertext by clearly expressing the rhetorical relationships between chunks of text. Meanwhile, agent-based argumentation has used Cognitive Coherence Relations as a theory of gmatics [126]. ‘he ScholOnto [166,17] project, which ran at Open University’s Knowledge Media Institute from 2001- 2004, focused on modeling claims and arguments in scholarly communication. ClaiMapper, ClaiMaker, and ClaimSpotter were among the tools”! developed '9 AraucariaDB™ , the Arucaria corpus of arguments, draws in part from discussion fora (BBC Talking Point, Christian Apologetics & Research Ministry Discussion Boards, MSNBC discussion forum, NPR discussion boards, and Global Greens) [86]. 2lhttp: //projects.kmi.open.ac.uk/scholonto/ software. html 18 Schneider et al. / A Review of Argumentation for the Social Semantic Web in the project, which was seen as part of sense- making research. An open source web publishing tool called the Digital Document Discourse Environ- ment”, or D?F [164] was also developed in related research. ScholOnto made an RDF Schema avail- able, but database queries with SQL were preferred to querying based on this RDF Schema (SPARQL was first released as a working draft in 2004). The under- lying ScholOnto ontology for these projects is shown in Figure 6.14 on the facing page. This ontology now underlies one of the applications we later discuss: Co- here is argument mapping software for sensemaking that integrates annotation and argumentation for the general public [162]. 7. Ontologies incorporating argumentation In addition to the seven ontologies discussed above, which implement or follow from particular theories— ChAO, DILIGENT, IBIS RDF, SALT, ScholOnto, SIOC-Argumentation, and SWAN/SIOC-—we now re- view seven ontologies relevant to argumentation. These include both dedicated argumentation ontologies (the Argument Interchange Format) and ontologies de- signed with substantial input from the argumentation community (the Legal Knowledge Interchange For- mat) as well as ontologies that incorporate small num- bers of argumentative elements (the Annotation On- tology, bio-zen-plus, the Citation Typing Ontology, the Non-functional requirements and Design Rationale Ontology, and the Semantic Annotation Vocabulary). 7.1. Argument Interchange Format (AIF) The Argument Interchange Format [120] is a pow- erful, dedicated ontology for argumentation, originally designed to ensure interoperability of argumentation software such as ArgDF, ArgKit, Carneades, and On- line Visualisation of Argument. AIF would be chal- lenging to apply to the Social Web because it requires argumentation schemes to be specified. In fact, even arguments themselves are not necessarily clearly spec- ified in the informal argumentation found in the So- cial Web! Thus, for example, enthymemes make for- mal specification of arguments challenging [22]. The original core ontology, shown in Figure 7.1 on page 20 consists of two disjoint sets of nodes: informa- tion nodes (I-nodes) holding the content of the argu- http: //d3e.sourceforge.net / ment and scheme nodes (S-nodes) holding the relation- ships between arguments. Scheme nodes are further di- vided into three main types, for representing logical in- ference (RA nodes), preferences or values (PA nodes), and conflicts between I-nodes (CA nodes). More re- cent work on AIF envisions classification of argumen- tation schemes, enabling automated reasoning over the schemes [136]. In Spring 2012, new specifications for AIF were released in OWL and RDF; an SQL database definition is also in progress”. AIF forms the foundation for the World Wide Ar- gument Web (WWAW). The WWAW is “a large-scale Web of interconnected arguments posted by individu- als to express their opinions in a structured manner” [138], where RDFS and OWL were suggested to be used for AIF. The foundations of the World Wide Ar- gument Web have been further discussed by Rahwan and others (e.g. [139,134,135]). AIF has continued to develop, and several published extensions of AIF exist. Rahwan adds form nodes (F- nodes) [138] in order to more fully represent generic argument schemes (as opposed to the instantiations of those schemes). Then Walton’s argument schemes can be represented, using ConflictSchemes to capture exceptions/Critical Questions. With AIF-RDF™*, Rah- wan et al. [138] add RDFS extensions to an AIF im- plementation. In this implementation, edges are explic- itly typed. Letia and Groza add a Context Node, used to evaluate the same argument in different contexts [100]. Rahwan et al. [136] present a new formalization of AIF in OWL-DL, implemented in Avicenna (Sec- tion A.9, page iii). Work on this area continues, largely published in the Computational Models of Argument (COMMA) conference series, with a large body of promising research in the 2012 edition of this biannual conference. While AIF was intended to model monological ar- guments, dialogue has been another area of inter- est in AIF extensions, with work from Modgil and McGinnis [118] and Reed et al. (e.g. [146]). Ear- lier work began the process of extending monological AIF for use in representing dialogical argumentation [141,145,147,140]. Bretp: //www.arg.dundee.ac.uk/?page_id=197 4nttp://argdf.org/source/ArgDF_Ontology. rdfs Shttp://www.comma-—conf.org/ Schneider et al. / A Review of Argumentation for the Social Semantic Web 19 Problem Related Causal Similarity Supports/ Challenges + t General | Taxonomic | is similar is differs f causing is the oppo: cr nar has to affect ts anal Fig. 12. Class structure of the Scholarly Discourse Ontology from [166] . 7.2. Annotation Ontology Argumentation enters into the Annotation Ontol- ogy’s*® [43] curation use case. In that use case, a hu- man curator reviews an annotation created by a text mining service, and first rejects it. This curator sub- sequently changes her mind after a discussion with a second curator, and finally accepts the annotation af- ter all. The statuses Rejected , Discusses, and Accepted express an argumentative workflow in this situation. The Annotation Ontology has been used by existing tools such as the SWAN Annotation Framework and Utopia PDF reader [177]. It is currently being incorpo- tated into a new W3C standard for Open Annotation 7 7.3. bio-zen-plus ontology framework The bio-zen-plus ontology*® [151] is an ontology for biology; as the name suggests, it is an extension of the bio-zen ontology’. It includes two argumentative http: //code.google.com/p/ annotation-ontology/ 7hetp://www.openannotation.org/spec/core/ Bhttp://neuroscientific.net/bio-zen-plus. owl Mhetp://neuroscientific.net/index.php?id=43 properties, supported-—by and in-conflict-with, augmenting the argumentation-related correlation-concepts, suchasPositive correlation (unsigned), Positive correlation (signed), Negative correlation (unsigned), and Negative correlation (signed), whichare found within the bio-zen ontology’. 7.4. Citation Typing Ontology (CiTO) CiTO*! [160,161] is an ontology for citation net- works in scholarly publications. Its argumentative terms include corrects, confirms, gives support to,is agreed with by, is ridiculed by, qualifies, and refutes. Papers can thus be semantically enhanced.*? For ex- ample, an author could indicate in a paper that it updates a previous publication, and critiques a piece of related work, while using evidence from an- other paper (citesAsEvidence). Readers can as- semble bibliographies using CiTO properties, for in- stance with the bibliographic management software Mhtetp: //neuroscientific.net/bio-zen.owl hetp://purl.org/spar/cito 2hetp://imageweb.zoo.ox.ac.uk/pub/2008/ plospaper/latest/ 20 Schneider et al. / A Review of Argumentation for the Social Semantic Web Graph [Informally: Argument metuork] is-a | Scheme node (S-node) for: scheme application node [Informally: Rule] is-a has-a ] Information node (I-node) or: data node [Informally : Node] | Derived And Rule of inference Conflict Preference eritati price PEE rer application application application aii bolt al) Se eae oat a nade (RA-nade) mode (CA-node)) |node ¢PA-node) te Fey ‘ CP ee application defeat) mode Conflict achene|—is a—>] Scheme }¢——— is eles [__————* Rule of inference scheme [also: rule scheme] s-a isa S=8 Logical Presumptive | inference inference scheme scheme Ca. Logical | Logical Presumptive! | Le is a is conflict aoe preference) |preference scheme scheme scheme Fig. 13. Concepts and relations of the Argument Interchange Format CiteULike*’, showing the possibilities of semantic an- notation. 7.5. Legal Knowledge Interchange Format Legal Knowledge Interchange Format (LKIF)*, developed as part of the ESTRELLA project, is an OWL ontology [3] for the legal domain. Its Rules & Argumentation Module deals with Exceptions, Rules, Arguments, and Assumptions [133]. It also imports the LKIF Expression Module, which pro- vides “‘a vocabulary for describing, propositions and propositional attitudes (belief, intention), qualifica- tions, statements and media” [133]. It includes terms for various PropositionalAttitudes, as well as Intention and Lie, for instance. Brttp://www.citeulike.org/ http: //www.estrellaproject .org/lkif-core/ lkif-rules.owl# 7.6. The NDR Ontology The Non-functional requirements and Design Ra- tionale (NDR) Ontology [102] addresses the visual- ization of non-functional requirements as Softgoal In- terdependency Graphs. While some classes (such as Softgoal) are specific to this domain, the NDR Ontology introduces useful argumentative labels and causal relationships. For example, the label prop- erty can be used to indicate the extent to which goals are met (i.e. whether they are adequately ‘satis- ficed’): Denied, Weakly denied, Undecided, Weakly satisficed, Satisficed,orConflict. NDR also has classes for Argument ation, Claim, Contribution, and Interdependency (includ- ing a subclass, Correlation). The Cont ribution of child goals to the parent goal can be labelled as Break, Hurt, Unknown, Help, Make, etc. Schneider et al. / A Review of Argumentation for the Social Semantic Web 21 7.7. Semantic Annotation Vocabulary The Semantic Annotation Vocabulary [60] was de- veloped for the Trellis system (Section A.35, page xvi). They used various dimensions: pertinence, reli- ability, credibility, causality (e.g. contribute to, indi- cate), and temporal ordering, as well as structural rela- tionships (such as part/whole, example-of, describes). 8. Comparison of Semantic Web Models In Figure 14 on the following page we present a comparison of the Semantic Web models discussed. Topics addressed include whether each ontology is centered on relations or concepts as well as whether it is IBIS-like (i.e. does it contain concepts function- ally equivalent to IBIS’ ‘Statement’, ‘Issue’, ‘Posi- tion’, and ‘Argument’?). We also cover what types of relations it contains, drawing from ScholOnto’s types: causal, similarity, generic, supporting, challeng- ing, taxonomic (e.g. hierarchical categorization), and problem related. Further, we describe whether polarity (e.g. positive vs. negative) and weights are explicit or implicit and whether the ontology specifies other on- tologies to use for content provenance and authorship provenance (such as from FOAF, SIOC, or PAV-the Provenance & Authoring and Versioning ontology**) and domain knowledge (such as from DOLCE, SKOS, or the PRotein Ontology). We have used a ‘?’ to indi- cate that we were not able to find this information in publications, or when information was ambiguous, to reconcile it. Some models provide a shallow view of arguments yet are situated within a larger (perhaps social) con- text. Yet other models, originating in the argumenta- tion community, focus on representing the arguments themselves, often including the internal structure of the arguments. The argumentation community’s inter- est in the Semantic Web has been motivated in part by the idea of The World Wide Argument Web (WWAW) [138], while the semantic web community’s interest has centered on communication structures, rather than the details of argumentation or rhetoric. Shtetp://swan.mindinformatics.org/spec/1.2/ pav. html 9. Features of Social Web Tools We conducted a thorough review of argumentation tools for the Social Web, attending especially to exist- ing Social Web sites, tools using the Semantic Web, and prototypes from the research community in Social Web and Semantic Web. In this section we describe re- lated reviews of argumentation tools, the scope of our review, and some highlights concerning thirteen par- ticular features of these systems—visualization, ease of use, collaboration, user engagement, balancing contri- butions, deliberative polling, distributive and federated systems, annotation, incremental formalization, pop- ulating knowledge bases from user input, mixed ini- tiative, search, reasoning and querying. These provide ideas of the aspects that may be need to be consid- ered for argumentation systems on the Social Seman- tic Web. For further detail about each system we re- viewed, consult the Appendix, Section A, page i. 9.1. Related Reviews Argumentation tools have been reviewed and overviewed in various publications, including two contemporary books. Visualizing Argumentation [87] presents eight chapters which cover the history and cognitive foun- dations of argumentation tools; describe tools for col- laborative learning and deliberation; provide insight into map-based facilitation of in-person meetings; and describe mapping scholarly debates. Of particular in- terest in this exceptional volume is the chapter on “The Roots of Computer Supported Argument Visu- alization" [165]. Knowledge Cartography: Software Tools and Mapping Techniques [122] provides seven- teen case studies of using mapping and argumentation tools, primarily in education, but also in science, poli- tics, and organizational knowledge transfer. Argumentation tools have also gained attention in e-government (e.g. [107], or for a tools review see ‘Opinion gathering in e-democracy’, Chapter 3 of Cartwright’s dissertation, [35]) and education ([{153]). Crossover interest in politics from the IEEE commu- nity is evidenced by a “Trends & Controversies’ sec- tion “AI, E-government, and Politics 2.0" [39]. Scheuer et al. [153] review 45 argumentation sys- tems*® used in Computer Supported Collaborative Learning and discuss 13 empirical studies involving 36Six of these systems are also discussed in our review below: Argunet, ConvinceMe, CoPe_it!, Debategraph, Debatepedia, and SEAS. 22 Schneider et al. / A Review of Argumentation for the Social Semantic Web & 2 Fy 3 3 : 5 3 8 < 8 -= a Ee -— = 5 6 fg & e $ z § = nN oO o e ¥ : = fo w ec 6 = 5 Yu =z °o e 5S oo o 2 = o s = - 2 © 4 gg a é zs $ - § gs 3 §& a Concept/Relation centered] R Cc R R R? R Cc Cc R R R R Cc ? OWL/RDF] OWL RDFS OWL OWL OWL RDFS RDFS RDFS OWL OWL RDFS RDFS OWL RDFS Statement] N Y Y N Y Ne N Y N N N N Y N Issue] N Y N N Y Y N Y N N N N N N Position} N Y N N Y Y N Y N N N N Y N Argument] N Y N N Y Y Y Y N N N N N N Causal} N N N N N N N N N N Y Y Y¥ N Similarity} N N N N N N N N N N N Y N N Generic] N N Y N Y Y N N Y Y N Y Y N Supporting] Y N N Y Y N Y Y Y Y Y Y Y Y Challenging] Y N N Y Y Y Y Y Y Y Y Y Y Y Taxonomic] Y N N N N N N N N N N Y Y N Problem-related| Y N N N Y Y Y Y N N N Y N N (a) Polarity Weights oe Pens bomen Design context Provenance Provenance knowledge NDR Ontology Implicit None None None None Sets engineering IBIS RDF Explicit None None None None oo ae solving Annotation Ontology Implicit For correlations FOAF, PAV SIOC, PAV Open Annotation . oe DOLCE, SKOS, . Bio-zen plus Implicit None SIOC FOAF Bio model Biology DILIGENT Implicit None ? Actor name None See ontology design . System: author, System: author, Collaborative chao ee None date, time date, time Coe ontology design Scholarl SALT Implicit None Implicit None Open eno ar Ys communication SIOC-Arg Implicit None SIOC ? None Social Web SWAN/SIOC Implicit Implicit SIOC ? None Scientific Discourse ciTO Implicit None None None None Scholarly citations Sem. Ann. Voc. Implicit Implicit ? ? None Web of trust ScholOnto Implicit Implicit None None None Digital libraries LKIF Explicit None Agents ? ? Legal AIF Explicit None ? ? ? Argumentation (b) Fig. 14. A comparison of Semantic Web models for argumentation in terms of (a) structure; and (b) features. Schneider et al. / A Review of Argumentation for the Social Semantic Web 23 the use of argumentation systems in education. Inter- esting results from their work are that arguments are constructed in learning applications in five main ways: free-form arguments, argumentation based on back- ground materials, arguments rephrased (e.g. reworded and rekeyed) from a transcript, arguments extracted (e.g. copied/pasted) from a transcript, and system- provided units, with combined approaches also used in some applications. Further, they compare the advan- tages and disadvantages of user-controlled and system- controlled layouts for education. Their discussion of ontologies is limited. This tools coverage in this section differs from pre- vious coverage in its scope of collaborative, Web- based tools with argumentation components, and in its attempt at comprehensiveness. A further bias has been software aimed at use by the public, rather than exclusively for government consultation, enterprise decision-making, or learning argumentation and criti- cal thinking skills. However, we have deliberately in- cluded several research prototypes which focus on Se- mantic Web approaches to argumentation on the Web and on supporting the nascent World Wide Argumen- tation Web. 9.2. Scope: Collaborative, Web-based tools with argumentation components Tools were considered in-scope if they were collab- orative (i.e. involved sharing information among mul- tiple parties who could build upon each others’ work in some way), Web-based (i.e. allowed display of in- formation on the Web), and had argumentative dis- cussion components. By argumentative discussion, we mean discussion around disagreements, explanations, and reasons, coming from or including a rational (i.e. reason-based) standpoint. Some prospective tools were excluded due to failing one or more of these conditions. Tools failing the ‘collaborative’ criterion included the EUProfiler*’, and the HealthCentral/Washington Post Poligraph 2008°*. Users of these tools viewed per- sonalized visualizations, based on their answers to a questionnaire, however they are not asked (or able) to share their comments on others’ views, to interact with other users (adding to a larger debate), or to contribute to sensemaking or analysis of existing argumentation. Thetp://euprofiler.eu/ Bhttps ://www.washingtonpost.com/wp-srv/ health/interactives/poligraph/ Tools failing the ‘Web-based’ criterion included email tools such as WIT and Zest, SAIC’s SIAM and Causeway, and the argumentation tools Carneades, Araucaria, and Convince Me*?. WIT*® [19] and Zest [210] focused on argumentation in email. SLAM*! and Causeway"? are Windows-based software for influence net modeling, designed for analyst use and primarily for collaboration inside the firewall; although HTML can be exported, Web-based collaboration is not sup- ported. Similarly, Carneades*? maps can be shared in LKIF, but not directly visualized online. Araucaria [150,142,144] offers a searchable online argument cor- pus, but not online display of its arguments. While Convince Me* offers a Java applet for display, argu- ments cannot be saved or published via the applet. Tools failing the ‘argumentative discussion’ crite- tion included general Web2.0 tools as well as Anek- dotz, and Vox Populi. General Web2.0 tools (e.g. Twit- ter“, Facebook*’) and social software (generic mail- ing lists, forums) were excluded since their argumen- tation support is peripheral. Anekdotz*® failed because the sites currently using it focus on the emotional, rather than the rational, aspects of argumentation. For example, the breakups section of When You Knew asks commenters to click on either ‘Put their stuff on the curb’ or ‘Give em another shot’ to solicit feed- back, which is marked as positive, negative, or neutral. Vox Populi*? [24] supports documentary filmmakers in generating argumentative film sequences based on annotated interviews. Further, tools were treated differently depending on their origin and availability; for instance, it was con- sidered helpful to include many contemporary research systems even though we were not able to interactively explore Web-based demo versions for some of those systems. We have inevitably missed some relevant sys- tems, and would appreciate the reader’s assistance in fixing this flaw. 3°Note that we do include the similarly named ‘ConvinceMe’ site in Section A.17, page viii. Mnttp://www.w3.org/WIT/ ‘http: //www.inet.saic.com/inet-public/siam. htm 2http://www.inet.saic.com/inet-public/ causeway.htm Shttp://carneades.berlios.de/ “http: //araucaria.computing.dundee.ac.uk/ Shttp://codequild.com/convinceme/ http: //twitter.com http://www .facebook.com/ Shttp://www.anekdotz.com/ http: //homepages.cwi.nl/~media/demo/IWA/ 24 Schneider et al. / A Review of Argumentation for the Social Semantic Web 9.3. Classifications of social tools Aakhus and collaborators [49,4] classify argumen- tation software by use: issue networking, funnelling, or reputation (Figure 15 on the next page). Shum says that each tool is ‘tuned’ to a different task: “foraging for material, classifying and linking it, discussing it in meetings and online, and evaluating specific points in more depth" [163]. We later use this categorization, as ‘Functional type’ in our comparison of tools. Scheuer et al. [153] compare the visualization and representation styles of argumentation tools used in computer-supported collaborative learning. They sum- marize the pros and cons of 5 representation styles, as shown in Figure 16 on the facing page. We later use this categorization, as “Representation style’ in our comparison of tools. Scheuer’s representation styles are typically used for discussions (linear represen- tation), modeling (container), or both (threaded, graph). For instance, graph representations are highly expressive, with explicit labelling of relationships, but make it hard to see temporal sequences. 9.4. List of social and Semantic Web tools considered In this section we draw from our review of thirty- seven online argumentation tools: AGORA: Partici- pate - Deliberate, ArgDF, Arguehow, Argument Blog- ging, Argumentum, Argumentations.com, Argunet, Avicenna, bCisiveOnline, Belvedere, Cabanac’s an- notation system, Climate CoLab, Cohere, Compet- ing Hypotheses, Considerlt, ConvinceMe, CoPe_it!, CreateDebate, Debate.org, Debategraph, Debatepe- dia, Debatewise, DiscourseDB, Dispute Finder, Hy- pernews, LivingVote, Opinion Space, Online Visual- isation of Arguments, Parmenides, PDOnline, REA- SON, Riled Up!, SEAS, Trellis, TruthMapping, and Videolyzer. These systems are described and discussed individually, in alphabetical order, in the Appendix, Section A, page i. 9.5. Visualization Overviews and visual representations aid under- standing. Here we point out visualization features of some Social Web argumentation tools, along with screenshots; further details are available in the ap- 50The container approach uses discrete visual areas to group re- lated items. For example in Debatepedia each question is contained. in a frame with pro and con arguments on that question. pendix. Argument maps are one classic representation, which continues to be popular with a variety of tools, including Argunet, Cohere, and Climate CoLab. On Argunet, users have significant control over the presentation of arguments, such as colors and descrip- tions of different argument families. Related maps can be published in series, as shown in Figure 17(a) on page 27. In the argument map representation, each node can be opened up to reveal a matrix listing which other arguments support, attack, are supported by, and are attacked by the given node (Figure 17(b) on page 27). Phase 2 Nodes: 16 The substantiation that the fossilized plants Edges: 17 lie at the top of the Greywacke triggered a new anomaly. On the one hand, the fossil criterion (H1) seemed to imply that the tap fr layers of the Greywacke are Carboniferous (Fossil-clock-argument). On the other hand, all British Carboniferous coal ... Inferential Density: 3 Phase 3 Nodes: 20 Murchison and his colleagues assembled Edges: 23 empirical evidence from all over Europe, hoping to obtain clues about the correct dating of the Devon Greywacke. During an a expedition to Russia in 1840, they found, (a) Coe ee a ed or Inferential Density: Press 9 i . oasits etsewhore found in Carboniterous stata opeut (only) at the top of the Devon Greywacke | Me napeeee | cara | (Steet terete a ] tel | (b) Fig. 17. Argunet can show an (a) overview of several related argu- ment maps; and (b) in each individual map, nodes can be opened up to show arguments they support, attack, are supported by, and are attacked by. In Argumentum®! arguments are colored to indi- cate the supporting (green) and opposing (red) argu- ments (Figure 18 on page 27). Comments, but not their replies, are similarly colored to indicate agree- ment or disagreement. Pro and con arguments are dis- tinguished by green and red lines to the left of a com- ment, posted linearly, rather than in two columns. Competing Hypotheses™ supports breaking down information into hypotheses, evidence, and analysis, SIhttp: //arg.umentum.com/ hetp: //competinghypotheses.org/ Schneider et al. / A Review of Argumentation for the Social Semantic Web discussions - Addresses issue of sequential incoherence - Modeling Graph (e.g., Belvedere, Digalo) - Discussions - Modeling - Intuitive form of knowledge modeling (Suthers et al. 1995) - Highly expressive (e.g., explicit relations) - Many graph-based modeling languages exist Container (e.g., - Modeling SenseMaker, Room 5) - Easy to see which argument components belong together and are related Matrix (¢.g., Belvedere) - Modeling - Easy systematic investigation of relations - Missing relations between elements are easily seen (Suthers 2003) A clash of claims Enables a full exploration of the race aes mecnorlicie where the interaction agreements, disagreements,and participants to share gE is organized into a web rationales for position bur explore, and learn about of Issues with relevant typically provides no each others’ positions and «Hyper jansandreasons functionalities for sectiing reasons and where these = QuestMap developed for differences. will be held accountable SConpardiien each issue. to the doubes of others. Funneling ‘Treat argumentation Sequences interaction Into a A situation calls for a z 35 ACCVE CONSENSUS series of activities thar commitment by all ; formation where successively Narrow a Gispute or participants to 3 particular = \GroupSye interaction is decision toward the most course of action. This -SAMM organized into a flow acceptable conclusion. The Stands in concrast to = Other GDSS from broad differences functionalities such as issue-networking, which - Voting taals toward an acceptable = brainstorming, categorizing,and = alms primarily at conclusion. votng provide means to remove self-correcthon, resistance to collective action. Ot At CONSENSUS Reputation ‘(Creace a Provides means like ratings for A situation calls for Ennripless base for action by individual discovering and cultivating pooling and refining with each other over experts. It SEgeee actrees fee erce who the bestanswers stands in contrast co both Sai eas a community of pss cinapracko ghee funneling, because it does ~ Webi participants. most wants answered, Th not aim at consensus, and orchestration Cee issue- because anda repository of answers. people are mot from their standpoints. Fig. 15. Issue-networking, funnelling, and reputation, from [49]. Representation style Typical uses Pros ‘Cons Linear (e.g., chat) - Discussions —- Familiar and intuitive to most - Risk of sequential incoherence (especially users, easiest to use (McAlister et al. 2004) synchronous) _ Rest to see temporal sequence - Not suited to represent the and most recent contributions conceptual structure of arguments ~ Lack of overview Threaded (e.g., forums, - Discussions - Familiar and intuitive to most - Moderately hard to see temporal Academic Talk) (especially users, easy to use sequence (because of multiple asynchronous) - Easy to manage large threads) as compared to Linear ~ Limited expressiveness (only tree-like structures) - Hard to see temporal sequence - Lack of overview in large argumentation maps (need a lot of space, can lead to “spaghetti” images (Hair 1991; Loui et al. 1997) ~ Limited expressiveness (e.g., only implicit relations, only tree-like structures) ~ Lack of overview in large argumentation maps because of missing relations ~ Limited expressiveness (e.g., supports only two element types (row, column), no relations between relations) - Uncommon (Non-intuitive) way of making arguments Fig. 16. Comparison of the visualization and representation styles of CSC. 'L argumentation tools, from [153]. 26 Schneider et al. / A Review of Argumentation for the Social Semantic Web poWade Abii) 0 ovis Henan | Healthcare is a human right You rt oppose ‘Top | Rew nied) EG eo On Fave (ws to ot) 27 rear sori by relevance #) (any positon #) EEE ] | Kevinconner t 1 ‘yinal are “human ronis"7The ony noms Wwe shows De conceMmed about are ihe Monts promised to us by Our | | Constnton. We do NOT need a larger goverment Lat tree markat capitalism oo it's job, The government | pawing contrat over neat. ‘comment Pon & e100 Spearman t 1 ; | can not Be refused. What I maan by this i a "right" & lntaly dependent en your own acton, | have the : to free speech, My actions may result in my Geath but no-one can stop me from doing It you are on @ - 4 | seserted land an AGomment Prema ® ¢76 { 1 ne has the right fo fa, iberty and the pursuit of happiness, that should not ba circumscribed by the ability to pay the extreme fees charged by tha fascists (I mean the insurance companies). The LS is the ely industrial naan that do. believe 40 #1 vas sities god = Fig. 18. In Argumentum, the left-hand color bars indicate the sup- porting (green) and opposing (red) arguments Fig. 21. At CreateDebate, users add and comment on pro and con arguments. which are entered into a matrix as shown in Fig- ure 36(a) on page vii. The matrix can help visually in- dicate the most likely and least likely scenarios.>* Mul- tiple analyses can be combined to provide a group view (Figure 36(b) on page vii), or compared pairwise. ConsiderIt™* [94,93] powers the Living Voters’ Guide®. What is unique is the possibility to drill down to understand other voters’ perspectives. In addition to seeing pros and cons on an issue from all voters, re- gardless of their stance, (Figure 20(a) on the follow- ing page), the Living Voters’ Guide can show the key points for a particular group of voters (Figure 20(b) on the next page), such as those undecided on the issue or strongly supporting it. This can help users understand Fig. 22. Opinion Space maps comments in a constellation view. what makes an issue controversial. Users indicate how they feel about an issue before and after reading an SEAS [105,104] structures arguments as templates, argument (deliberative polling), which could also be showing a colored tree view (Figure 23 on the facing used to find the most convincing arguments. page). SEAS visualization features are also consider- CreateDebate (Figure 21 on page 27) offers nu- able: to visualize multiple dimensions, SEAS uses star- merous statistics for each debate, such as the lan- burst, constellation, and table views. guage grade level, average word lengths, and vocab- ulary overlap, as well as a wordcloud. Some debates have more than two sides. Replies can ‘support’, ‘dis- pute’, or ‘clarify’ a given point. In Opinion Space [55], opinions are mapped in a constellation, using principal component analysis, to show a user where they stand compared to other re- spondents, as shown in Figure 22 on page 27. Each point in the visualization represents a perspective; larger points represent more popular perspectives. 9.6. Ease of use By ‘ease of use’, we mean how easy an interface is to use, based on our own perception. ConvinceMe® lets users add arguments to pro or con columns, or add a rebuttal by clicking a button. Arguments can be voted up or down. Various other systems (discussed in the appendix) have similar functionality. Debatepedia®’, a wiki-based system, provides an intuitive editing environment, where users can edit the entire page or just the relevant section, such as the pro *3More sophisticated ACH-based software uses matrices as input or con for a topic. to Bayesian probabilistic reasoning. “http: //engage.cs.washington.edu/ considerit/ SShttp: //www.convinceme.net/ Shttp://www. livingvotersguide.org/ http: //debatepedia.idebate.org/ Schneider et al. / A Review of Argumentation for the Social Semantic Web 27 Edit your enasistency scores Duplicate personal matrix ‘Show Data Columns | | Sart Hypotheses || Sort Evidence | ee ee rs ery Inconniniant Nestea! Sen Neutral Inconsistent ee oe ered Inconsistent Incoresistent 1B veh: (arian tans 28) Compare the rabiogs of { tester ‘view others’ personal matrices; (Manew surton iS) (Views) (a) Group Matrix Consensus Guage: . Unsnieniny [Show Data Columns || Sort Hypothases || Sort Evidence | Pe eg LL Consensus (@) Consensus | Consensus No | Mild Dispute: Mild Dispute Consensus () Consensus 6 Consensus 1 ‘Compare the ratings of [ vester TE) wen; Cwatbew eurton — TS) (a ‘new others’ personal matrices: | Mattheew furton 589 (View. (b) Fig. 19. In Competing Hypotheses, (a) each individual’s analysis is represented in a consistency matrix; (b) multiple analyses can be combined to create a group matrix. In the group view, darker shades of purple indicate more disagreement. = Explore what other voters think about 1053. ii] Cooren uber reseed! he eypata Baise uur betes Pe =—oO ‘om your rat a come PED INE ou ins (ohatte ts gaancn chloe wacanend aor (a) Explore what other voters think about 1053. all ‘GGK on a Bere Bee seme ete hey poss Bue ine Group Eaves aco ree Ce eran Ss a cvclavaani cham Oe eet (b) Fig. 20. The Living Voters’ Guide compiles pro and con lists on each issue. They give (a) an overview of what all voters think about the issue; as well as (b) the key points for undecided voters. Arg ° “| 1 1.3.1 UNUSUAL TRANSACTION: ts there a large, unusual, of questionable | ansaction? John Lowrance On: 15 Jan 2007 16:45:30 @ Yes, almost certainly Likely, more likely than not Even, about as likely as not Unlicely, more unlikely than not No, almost certainly not Rationale: The bank transaction is sufficiently large to warrant strong suspicion. eo Bank Tr ci Record - 15 May 2006 6 mtered Dy: Ti Boyce 17 Dee 2006 93243 : 8; This contains a large transaction that is out of the ordinary for a business of this size, 4 Computer of Employee Jahn Doe - 13 Mar 2006 bee Erie Yeh On: 14 Nov 2006 10:19:43 je e: Deleted records on his computer included references to. a known problematic company. Email Messape tram John Doe to Jane Ooe - 5 Apr 2006 ip Janel Murdock On: 18 Jan 2007 1:2223 Exhibits | "| Fig. 23. SEAS visualizes arguments as a tree; colors indicate the credibility level [105]. Living Vote*® asks participants to vote on argu- ments in a tree. “To vote on an argument, you must first prove you’ve read it by answering a simple question.” Users can also add arguments to the tree. 9.7. Collaboration Systems approach collaboration in various ways. bCisive Online is intended for real-time collabora- tion in conjunction with with audio conferencing. One person edits the map at a time, adding nodes and con- nections between nodes while others can point with their cursor or request editing control. Climate CoLab uses moderation to help review comments and deal with the learning curve for argu- ment mapping. While any user can add Pros, Cons, and Issues directly to an argument map, moderators are ex- Shttp://www.LivingVote.org/ 28 Schneider et al. / A Review of Argumentation for the Social Semantic Web pected to keep track of the conversation, adding new ideas to the argument map as needed. Competing Hypotheses*’ has persistent chat (es- sentially a comment thread) for the entire project as well as message boards for each hypothesis, evidence item, and evidence-hypothesis pair. On Debatewise™, everyone can collaborate in cre- ating the strongest case both for and against a given issue, using teams. 9.8. User Engagement Debatewise also makes it easy to get involved by providing suggestions of 5-minute, 20-minute and 1- hour tasks and showing “7 things you should have an opinion on" in rotating images on the homepage. Social networking is one part of the engagement at Debate.org, which allows users to search for people with particular profile attributes, such as income, lo- cation, ideology, gender, president, religion, and who they are interested in and looking for. Along with adding comments, on many sites, users can vote for arguments that convinced them (e.g. Cre- ate Debate). A user’s reputation is generally based on the success of their arguments. Votes may be weighted, for instance after 10 up or down votes, further votes have less influence on ArgueHow. Competition helps engage users at other sites, which can be quite explicit about this aspect; for instance, Riled Up!’s tagline is “Like Raising Cain? So Do We." Other sites, like ConvinceMe treats debates as games; in one such game, the debater whose idea is most pop- ular is crowned “King of the Hill”. Competitive de- bating environments may use point schemes and user rankings to motivate contributions. 9.9. Balancing Contributions There are several approaches to balancing contribu- tions. For example, by removing authorship markers, argument maps may increase the neutrality of a con- versation. Another approach to balancing is taken by TruthMap- ping®', which focuses on having a persistent con- versation which can not be drowned out by a single opponent. Users can leave feedback in critiques at- tached to each premise and conclusion (Figure 53(b) Shetp: //competinghypotheses.org/ ®nttp://debatewise.org/ Slhttp: //www.truthmapping.com/ on page xvii), users can continually modify each con- tribution, but can only post one critique on each node. Anyone can leave a rebuttal, but only one user, the original arguer, can modify the map. The system in- dicates when comments are out of sync with the map, and a wiki-style history is available to better trace the conversation. 9.10. Deliberative Polling One measure of arguments is how persuasive they are. Measuring how users feel about an issue before and after reading an argument—known as deliberative polling [57]-is used in some systems, such as Con- siderIt, Living Vote, and OpinionSpace. Debate.org uses deliberative polling but places more importance on other factors besides agreement when scoring debates, giving the most importance to using reliable sources and making convincing argu- ments. 9.11. Distributed and Federated Systems Systems can be distributed, like Argument Blog- ging, in which JavaScript code is placed on blogs to link back to a server which provides access to a larger conversation. Federation is an exciting direction: Argunet [154] uses an open source federation system for sharing ar- gument maps from a desktop tool®. Uses can run their own server or use a public server, Argunet.org™, which allows authors to make maps public or restrict viewing and/or editing to a specified group. These mechanisms also impact the privacy of a system—whether work can be saved privately, used col- laboratively with a small group, or shared publicly with the world. 9.12. Annotation Annotating materials or commenting on existing discussions can be an important way to interact on the Social Web. Annotation can also be used to clas- sify messages. For instance, Hypernews™ [21] asks users to indicate what kind of message they are post- http: //www.argunet.org/editor/ Snttp://www.argunet.org/debates/ “http: //www.hypernews.org/HyperNews/get/ hypernews/reading.html Schneider et al. / A Review of Argumentation for the Social Semantic Web 29 ing (None, Question, Note, Warning, Feedback, Idea, More, News, Ok, Sad, Angry, Agree, Disagree). Videolyzer® [50] provides an integrated discussion forum for annotating and challenging the claims a video makes. Segments likely to be of interest are iden- tified ahead of time by processing both the transcript and the video. Annotation has also been treated in from an argu- mentation perspective in the research of Cabanac et al., who study social validation of argumentative de- bates through collective annotations [28]. 9.13. Incremental Formalization With incremental formalization, representations are useful before they are fully interpretable by the com- puter. Incremental formalization can be helpful since people find it difficult to make structure, content, or procedures explicit [159]. As the user’s understanding (or goals) change, some systems facilitate systematic additions or changes. With Argunet, arguments can be quickly sketched or reconstructed as premises and conclusions, support- ing incremental formalization. CoPe_it! transforms the user-created informal spa- tial hypertext view (Figure 39(a) on page ix) into an issue chart (Figure 39(b) on page ix) according to rules shown in Figure 39(c) on page ix). Users can also cus- tomize the transformation rules. 9.14. Populating Argumentative Knowledge Bases From User Input Other approaches are also iterative, similar to incre- mental formalization but focusing specifically on the input phase. For instance, user input may be processed, and the output presented to the user, who can then cor- rect it. Trellis introduced a language processing technique called “Annotation Canonicalization through Expres- sion synthesis" [23], which applied an ontology to a user-supplied sentence, checked the computer’s ontol- ogy application by presenting a paraphrase to the user, and solicited additions to the ontology from unknown or misunderstood words. Controlled natural languages, which adopt a more restrictive grammar and vocabulary in order to facili- tate parsing, have also been used to take in informa- tion, formalizing it for reasoning. For instance, Wyner Shttp://videolyzer.com/ et al. [206] propose using a controlled natural lan- guage called Attempto Controlled English [46] for high-stakes argumentative discussions, in order to gen- erate a first-order-logic representation of the discus- sion. Dispute Finder® [53,54] provides just-in-time in- formation, alerting users when information they read is disputed, based on its database of disputed claims. This relies on a disputes database which was first pop- ulated by hand-annotation by activists (interested in in- forming or convincing others) and then extended al- gorithmically. The algorithm, which can be applied on the Web at large, uses a set of 54 patterns to identify possible disputed claims. 9.15. Mixed Initiative Another possibility is to use Mixed Initiative sys- tems, wherein the actions of both humans and ma- chines are important. Online Visualisation of Argu- ment (OVA) is part of a pipeline of argumentation tools [168] which starts to bridge the gap between human-oriented argumentation tools and calculation- based agent argumentation. Mixed initiative discus- sions are enabled by the argument maps created by OVA or any other AIF-based tool. Thus, instead of rep- resenting one’s point of view countless times in a fo- rum or FAQ, it would be possible to delegate these conversations to a machine agent using an underly- ing argument map, as prototypes like MAgtALO® [145,202] and the Google Wave discussion bot Arvina show. 9.16. Search Semantic search focuses not on mere keyword matches but on retrieving structured data, such as whether an opinion is argued for or against. Seman- tic search is possible with several tools. ArgDF® uses the AIF-RDF ontology described above [139,138,212] and Sesame RDF”. In ArgDF, it is possible to display all the arguments in which a claim is involved (e.g. where it is used as a conclusion or as a premise) or all the arguments using, say, the Argument from Expert Opinion reasoning scheme. Snttp://ennals.org/rob/disputefinder.html http ://ova.computing.dundee.ac.uk Shttp: //www.arg.dundee.ac.uk/?page_id=61 nttp://argdf.org/ Various other tools tools export AIF without directly imple- menting semantic search. 30 Schneider et al. / A Review of Argumentation for the Social Semantic Web DiscourseDB’! uses Semantic MediaWiki [95] with the SemanticForms” extension. This makes it possi- ble to list all commentary written by particular person, published in a particular venue, and so forth. Further, since items indicate the position they take on a topic, DiscourseDB can list all commentary for or against a given position. When a topic has multiple positions (e.g. Darfur’), DiscourseDB is especially helpful in summarizing the discussion. Semantic search uses a simple syntax: for instance, on the Darfur conflict is- sue, to search for commentary opposing the position that the U.N. should send peacekeepers, this code is used: [fis against::Darfur conflict / United Nations should send peacekeepers]]. Since the results are al- ready displayed on summary pages, most end users would not need to create or modify queries, but Dis- courseDB’s semantic search is a powerful tool for cre- ating summaries. 9.17. Reasoning and Querying As previously mentioned, Attempto Controlled English can generate first-order-logic representations, which allow inference and consistency-checking, and can be translated into OWL and RuleML [58]. Mean- while, Open Vocabulary Executable English can be used for simple reasoning in the Internet Business Logic System”. Parmenides” [32,33,34] is a structured survey tool for gathering public opinion on a proposal. Based on argument schemes and critical questions from argu- mentation theory, Parmenides can pinpoint the source of the disagreement, by having participants respond to a series of questions. At the end of the survey, users are offered the choice of submitting an alternative pro- posal, and are shown the answers they chose. Admin- istrators can then analyze the group’s responses, which are displayed in graphical argumentation frameworks [51]. More advanced reasoning and querying is enabled by Avicenna, an OWL-based argumentation system on the Web which uses Jena [31], ARQ”, and Pel- let [167]. Since OWL supports inference over transi- Mhttp://discoursedb.org/ Pyttp://www.mediawiki.org/wiki/Extension: Semantic_Forms Bhttp: //discoursedb.org/wiki/Darfur_ conflict http://www. reengineeringllc.com/ Bhtetp://egi.csc.liv.ac.uk/~parmenides/ Mhttp://jena.sourceforge .net/ARQ/ tive properties, Avicenna can support argument chain- ing, such as retrieving all arguments that directly or in- directly support a given conclusion. Avicenna is also used to infer the classification hierarchy of argument schemes: for example, an appeal to expert opinion is a specialization of an argument from position to know. 10. Discussion & Conclusion We have reviewed argumentation theory, existing ontologies with argumentative aspects, and Web tools for argumentation. We now discuss three main gaps based on our observations. First, the ontologies given need further adaptation to meet the existing variety of social tools and purpose. In particular, arguing is a social activity. The varieties of argument tools on the Social Web need distinct types of interface support and social engineering. The remaining question is what appropriate Social Web ontologies for argumentation would be. Rather than a single ontology, we envision modular, interoperable components. 10.1. Arguing is a Social Activity As argumentation scholars have long realized, hu- mans argue for a variety of reasons, not always to solve “wicked problems". Rather, arguing is a social activ- ity people may use to position and establish them- selves. This kind of arguing is important in the Social Web, where people play by arguing such as with Con- vinceMe’s the ‘King of the Hill’ game, or create net- works of friends and enemies, such as on Riled Up! and Create Debate. Arguing can also be used to con- nect people such as on Debate.org. Ontologies for the Social Semantic Web will need to respect these social aspects, and may need to incorporate emotive indica- tors such as the heat of the debate as well as the manner in which the outcome will affect the participants. The notion of debate, where two parties face off, is also well-represented in existing social tools. Debate may allow individuals to show their expertise, to find the best arguments, or simply to practice their rhetor- ical skills. Debate topics may be reused, for ongoing issues with two or more defendable positions, espe- cially when a topic is controversial. This suggests two opportunities. First of all, future Social Semantic Web prototype tools for sensemaking and argument map- ping could be tested with for argumentation for some common debate topic in order to find a large audience of potential evaluators. Second, providing meaningful Schneider et al. / A Review of Argumentation for the Social Semantic Web 31 ways to discover new debate topics, and potentially record and share the outcome of these debates, could be helpful. Frequent debaters may also provide an in- teresting class of users since we might expect them to be more familiar with fallacies and argument diagram- ming, making them potentially more savvy about ar- gumentation schemes and similar abstractions. 10.2. Current Use While argumentation support has become more mainstream, it is still a niche. While there is a desire for public discussion systems, especially in areas such as e-government, social discussion systems and social networks are driven by network effects (e.g. you are persuaded to use them by the ability to communicate with your friends and colleagues) and by ease of use. Argumentative elements in generic social media tools are very basic: Facebook and Google Plus use ‘Like’ and ‘+1’ buttons, which imply a semantics of agree- ment; YouTube adds a ‘dislike’ button, and flagging posts for moderation (e.g. on Craigslist) or downvoting posts (e.g. on StackOverflow or Reddit) also implies dislike. With existing systems, discussed in the Appendix, there is a continuum from those with little use to those with wide use. Some (non-research) sites have few users and seem to have been abandoned. Some re- search prototypes are not accessible at all (and have been discussed based on papers and screenshots). Other research prototypes are available, and some seem to have users. Some are widely (or at least some- what) used — showing (perhaps) what’s needed to build a Social Web infrastructure for argumentation. Argumentation support has not yet moved firmly from the academic lab, into the mainstream. While dis- cussion is widespread, argumentation needs are often specific to the reasoning schemes used — which vary by discipline and area. Such constraints simplify the reasoning process for humans as well as for argumen- tation support. Further, in dialogue, most argumenta- tion happens informally: we can count on our conver- sational partners to indicate what is missing and to de- mand that we explain what is unclear to them. It is difficult to systematically indicate assumptions and to make reasoning explicit; while this is needed for ideal reasoning support, it is not feasible or reasonable to ex- pect in everyday discourse. This leads into a discussion of usability. Usability needs depend on the task at hand and the target audience. Tools for in-depth analysis by experts can be more complex and involved than those for ca- sual use by the general public. E-government and de- liberation tools have the strictest usability needs for this reason. 10.3. Bridging the Social Web and the Semantic Web to Manifest the World Wide Argument Web To conclude, we discuss the obstacles to manifesting the Social Semantic Argumentation Web, along with a research agenda. 10.3.1. Problems Despite candidate technical architectures for a World Wide Argument Web, the WWAW is not yet viable on the Social Web at large. We notice several interrelated obstacles: first, the existing ontologies are not meant for integrating wide-scale informal discussion; second, current approaches to supporting argumentation gen- erally require substantial human effort; and third, de- termining the appropriate uses and re-uses for social media posts depends on their context (e.g. the type of discourse). First, the very informality of social media can make discussions more difficult to integrate. Argumentation is used in many contexts and while formal argumen- tation can be represented with ontologies such as AIF, argumentation on the Social Web can be quite infor- mal, with missing premises and unexpressed argument schemes. While human analysis can sometimes bridge the gap between AIF and the Social Web, (facilitated for instance by tools such as OVA, Section A.29, page xiii), more scalable solutions are needed. Several ap- proaches will be needed to more routinely express the existing argumentation on the Social Semantic Web. Second, most current approaches to supporting ar- gumentation still require substantial human effort; little automatic processing is available. Issues and stances can be categorized by hand, mapped and ana- lyzed, or voted up and down. Tools for certain tasks— situational analysis, argumentation reconstruction, and argument mapping—are highly developed. These tools have become, while not mainstream, widely accepted for certain communities. The value of spatial hypertext visualization systems cannot be discounted, and some automation does exist: leveraging human-devised ma- trices using Bayesian networks (in more advanced ver- sions of ACH tools like Competing Hypotheses), or summarizing human-answered surveys with argumen- tation frameworks (Parmenides). Yet since these value- added tools still require substantial human effort to en- 32 Schneider et al. / A Review of Argumentation for the Social Semantic Web ter information in particular formats, their very use is an encouraging sign that some forms of argumentation support would be beneficial. Third, context is important for integrating conversa- tions and claims. One strength of the Semantic Web is in bringing conversations together; this has been very powerful for the Social Semantic Web in general. Yet the rhetorical effect of an argument depends on certain contextual information, such as the surrounding con- versation, its participants, and the medium. Extract- ing and summarizing conversations without this con- text has risks (i.e. potentially presenting misleading or overstated arguments). 10.3.2. Research Agenda To overcome these obstacles and manifest the So- cial Semantic Argumentation Web, we see a need for various approaches. First, we need ontologies suitable for representing informal Social Web arguments, and to map to these ontologies. Second, to address and re- duce the human effort required we also need to moti- vate participation, and find ways to infer argumenta- tive relationships. Third, we need to further investigate context. First, we need ontologies that map between the social world and the argumentative world. A modu- lar approach will be needed, reusing existing work, both in domain knowledge and in Social Web mod- eling, for instance by importing existing ontologies (particularly SIOC). Maximal benefit from a Social Web ontology for argumentation will come from align- ing and crosswalking to popular ontologies (especially AIF). Many tools can already output AIF, and analyst- oriented tools can be brought onto the Argument Web with comparatively little effort. Motivated users and defined argumentation schemes ease this process. Mapping to these ontologies will also leveraging the existing human effort already used in argumentation tools. SEAS, for example, already uses argument tem- plating. Such templates appear to be specialized ar- gument schemes, which could be expressed in shared repositories and even classified (for instance using OWL as Avicenna does). Once the argument schemes can be referenced, SEAS might provide another source of AIF data, as well as point to further enrichment needed. The ACH process underlying Competing Hy- potheses seems to use a narrower set of reasoning; its data, similarly, might be encompassed by understand- ing and expressing the ACH argument scheme. The an- alyst community is also a good place to start with in- terface interventions such as using Controlled Natural Language (CNL); whereas on the general Social Web, CNL would restrict input, in analysis tools, CNL might open the vocabulary. Second, to further address the human effort re- quired, we need to motivate wide-scale participation and improve automatic argument detection. To encour- age ongoing human effort, it would be helpful to create a virtuous cycle—-by which people benefit from the Ar- gument Web, thereby motivating increased participa- tion. Understanding the niches filled by existing tools, and whether these needs could be fulfilled better by a larger Argument Web, would help in this regard. For instance, while abstract argument schemes may not be well understood by users, Parmenides shows that stepwise processes based on these schemes can be powerful. Opening up the analysis tools, so that a group could view aggregate responses, would take Par- menides to a new level of collaboration. While Par- menides focuses on gathering multiple responses on the same set of issues, a different approach would be to crowdsource work based on an argument scheme. Many groups already do this informally with check- lists and procedures, for instance in Wikipedia’s article promotion process. Providing templates where users could indicate which critical questions they have asked and answered, and at what point in time, might help to distribute and share this work, while making the un- derlying process more transparent. Automatic detection of arguments might help fur- ther bootstrapping the existing Social Web into the Ar- gument Web. In the scholarly communication and le- gal fields, argument detection relies on rhetorical fea- tures. Argumentative markers would also help in mod- ifying these argument detection approaches for use on the Social Web. Analyzing existing Social Web cor- puses, such as DisputeFinder’s claims database and the Discussion Fora from the Aracaria corpus may help in determining such markers. Various corpus-processing techniques and approaches may be useful for detecting argumentation, which shares rhetorical features with other sorts of speech. Linguistic pragmatics dominate in much argumentation, so one form of progress would be to find unassailable features which mark argumen- tative contexts on the Social Web. Relevant approaches may come from opinion mining [124], question an- swering and explanation [119], contradiction detection [149], controversy detection [72], persuasion detection [211,8], stance detection [188] and automatically typ- ing links [45]. Third, we need to investigate how to preserve the thetorical effect of an argument even when it is di- Schneider et al. / A Review of Argumentation for the Social Semantic Web 33 vorced from its original context. Some aspects of con- text are straightforward: for instance, items can be con- fusing or non-sensical when stripped of context. The canonical example from metadata scholars is the “on a horse” [204] problem: in the context of a Theodore Roosevelt collection, the description “on a horse" ad- equately describes a photo; yet outside of that context it is unclear and diffuse. Other contextual factors in- clude the type of argumentative discourse. The vari- ous types of argumentative discourse, shown in Fig- ure 4 on page 7, vary in the amount of interest and value they generate, outside the immediate context of the discussion. Eristic dialogue, of personal conflicts, for instance, is generally not worth reusing (though it could be used to establish, for instance, who started an argument, or that a dyad should avoid further discus- sions). Discovery dialogue, on the other hand, which seeks to find and defend a hypothesis, can be useful both for understanding the process undertaken and the outcome. Support may also depend in part by who reuses discussions—the participants, outside parties, or both— and how much support they need at various points in time. Reviews, for instance, are mainly written for an external audience. Blog and microblog posts may be read by others but also searched by the author as a form of externalized memory. There may also be a temporal component: standards bodies use their own discussions in order to make decisions, but after the fact, these discussions may be of considerable interest to non-participants trying to understand why a partic- ular decision was taken. The decision-making associ- ated with discussions may be a particular point where support is needed. Decisions can be taken by groups (e.g. standards bodies) or by individuals (e.g. from re- views) and may depend on a centralized discussion or widely distributed pieces of information. Overall, there is significant potential for support- ing argumentation on the Social Semantic Web, but a large amount of work remains to be done, in creat- ing ontologies, easing human annotation of arguments, improving techniques for detecting and mining argu- mentation, and in marking the context, pragmatics, and provenance of dialogues. Given the vast amount of commentary on the social web — much of it argu- mentative, persuasive, or opinionated — there is a great need for further research and technical developments to search and organize these discussions with semantic web technologies. Acknowledgements This work was supported by Science Foundation Ireland under Grant No. SFI/09/CE/11380 (Lion2). We are grateful for the extensive comments of the reviewers, Iyad Rahwan, Simon Buckingham Shum, and Fouad Zablith. Thanks also to Agnes Sandor for her comments which have prompted further clarifica- tions and furthered our thinking. The re-revision of this paper also benefited significantly from conversa- tions with and feedback from Katie Atkinson, Trevor Bench-Capon, and Adam Wyner, thanks to a visit to the University of Liverpool sponsored by the COST Action ICO801 on Agreement Technologies, STSM 1868. References [1] RDF Vocabulary Description Language 1.0: RDF Schema. W3C Recommendation 10 February 2004, World Wide Web Consortium, 2004. http://www.w3.org/TR/ rdf-schema/. [2] SPARQL query language for RDF. W3C Recommenda- tion 15 January 2008, World Wide Web Consortium, 2008. http://www.w3.org/TR/rdf-spargl-query. [3] OWL 2 Web Ontology Language: Document Overview. W3C Recommendation 27 October 2009, World Wide Web Consortium, 2009. http://www.w3.org/TR/ owl2-overview/. Mark Aakhus. Pragmatic Web as communication design practice. In PragWeb 2006. Proceedings of the First Inter- national Conference on the Pragmatic Web 21-23 September 2006 Stuttgart, Germany. Ges. fiir Informatik, Bonn, 2006. Mark Aakhus and Miriam Greenfeld Benovitz. Argument re- construction and socio-technical facilitation of large scale ar- gumentation. In Proceedings of the 3rd International Confer- ence on the Pragmatic Web: Innovating the Interactive Soci- ety, pages 77-81, 2008. Agnes Sandor. Using the author’s comments for knowledge discovery. In Semaine de la connaissance, Atelier texte et connaissance, Nantes, France, 2006. Vincent Aleven and Kevin D. Ashley. Evaluating a learning environment for case-based argumentation skills. In Proceed- ings of the 6th international conference on Artificial Intelli- gence and Law, ICAIL 1997, pages 170-179, New York, NY, USA, 1997. Pranav Anand, Joseph King, Jordan Boyd-Graber, Earl Wag- ner, Craig Martell, Doug Oard, and Philip Resnik. Believe me-—we can do this! Annotating persuasive acts in blog text. In AAAI Workshop on Computational Models of Natural Ar- gument at the Twenty-Fifth AAAI Conference on Artificial In- telligence, 2011. [9] M.A. Angrosh, Stephen Cranefield, and Nigel Stanger. Ontology-based modelling of related work sections in re- search articles: Using CRFs for developing semantic data based. information retrieval systems. In iSemantics 2010, 2010. [4 [5 [6 [7 [8 34 [10] Qi [12] [13] [14] [15] [16] 117) [18] [19] [20] [21] [22] [23] [24] [25] [26] Schneider et al. / A Review of Argumentation for the Social Semantic Web Anupriya Ankolekar, Markus Krétzsch, Duc Thanh Tran, and Denny Vrandecic. The two cultures: Mashing up Web 2.0 and the Semantic Web. Journal of Web Semantics, 6(1):70-75, 2008. Kevin D. Ashley. Modeling legal arguments: Reasoning with cases and hypotheticals. MIT Press, Cambridge, MA, USA, 1991. Kevin D. Ashley and Edwina L. Rissland. But, see, accord: Generating blue book citations in HYPO. In Proceedings of the Ist International Conference on Artificial Intelligence and law, pages 67-74, 1987. Katie Atkinson, Trevor J. M. Bench-Capon, and Doug Wal- ton. Distinctive features of persuasion and deliberation dia- logues. Argument & Computation, 2012. in press. Trevor J. M. Bench-Capon. Specification and implementa- tion of Toulmin dialogue game. In Proceedings of JURIX, volume 98, pages 5-20, 1998. Trevor J. M. Bench-Capon. Persuasion in practical argument using Value-based Argumentation Frameworks. Journal of Logic and Computation, 13(3):429-448, 2003. Trevor J. M. Bench-Capon and Edwina L. Rissland. Back to the future: Dimensions revisited. In Proceedings of JURIX 2001, 2001. Neil Benn, Simon Buckingham Shum, John Domingue, and Clara Mancini. Ontological foundations for scholarly debate mapping technology. In Second International Conference on Computational Models of Argument (COMMA ’08), May 2008. Tim Berners-Lee. Linked Data. Design issues for the world wide web, World Wide Web Consortium, 2006. http:// www.w3.org/DesignIssues/LinkedData.html. Tim Berners-Lee and Mark Fischetti. Weaving the Web: The Original Design and Ultimate Destiny of the World Wide Web by its Inventor. Harper Collins Publishers, New York, 1999. Tim Berners-Lee, James A. Hendler, and Ora Lassila. The Semantic Web. Scientific American, 284(5):3443, 2001. David Alan Black. Introducing HyperNews. Linux Journal, (27), July 1996. Elizabeth Black and Anthony Hunter. Using enthymemes in an inquiry dialogue system. In Proceedings of the 7th interna- tional joint conference on Autonomous agents and multiagent systems-Volume 1, pages 437-444. International Foundation for Autonomous Agents and Multiagent Systems, 2008. Jim Blythe and Yolanda Gil. Incremental formalization of document annotations through ontology-based paraphrasing. In Proceedings of the 13th International Conference on World Wide Web, pages 455-461, 2004. Stefano Bocconi, Frank Nack, and Lynda Hardman. Sup- porting the generation of argument structure within video se- quences. In Proceedings of the 16th ACM conference on Hy- pertext and hypermedia, pages 75-84, 2005. David Braybrooke. Introduction to special issue of argumen- tation originating in a conference at The University of Texas at Austin. Argumentation, 17(4):361-363, 2003. John G. Breslin, Andreas Harth, Uldis Bojars, and Stefan Decker. Towards Semantically-Interlinked Online Commu- nities. In The Semantic Web: Research and Applications, Proceedings of the second European Semantic Web Confer- ence (ESWC ’05), number 3532 in LNCS, pages 500-514. Springer, 2005. [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] John G. Breslin, Alexandre Passant, and Stefan Decker. The Social Semantic Web. Springer, 2009. Guillaume Cabanac, Max Chevalier, Claude Chrisment, and Christine Julien. Social validation of collective annotations: Definition and experiment. Journal of the American Soci- ety for Information Science and Technology, 61(2):27 1-287, 2010. Daniela V. Carbogim, David Robertson, and John Lee. Argument-based applications to knowledge engineering. The Knowledge Engineering Review, 15(2):119-149, 2000. Jeremy J. Carroll, Christian Bizer, Patrick Hayes, and Patrick Stickler. Named graphs, provenance and trust. In Proceed- ings of the 14th International Conference on World Wide Web, 2004. Jeremy J. Carroll, Ian Dickinson, Chris Dollin, Dave Reynolds, Andy Seaborne, and Kevin Wilkinson. Jena. In Proceedings of the 13th international World Wide Web con- ference - Alternate track papers & posters - WWW Alt. ’04, page 74, 2004. Dan Cartwright and Katie Atkinson. Political engagement through tools for argumentation. In Computational models of argument: Proceedings of COMMA 2008, pages 116-127. IOS Press, 2008. Dan Cartwright and Katie Atkinson. Using computational ar- gumentation to support e-participation. [EEE Intelligent Sys- tems, 24(5):42-52, 2009. Dan Cartwright, Katie Atkinson, and Trevor Bench-Capon. Supporting argument in e-democracy. In Electronic Gov- ernment: Third International Electronic Democracy (EDEM 2009), pages 151-160, 2009. Daniel Robert Cartwright. Digital Decision-Making: Using Computational Argumentation to Support Democratic Pro- cesses. PhD thesis, University of Liverpool, 6 2011. Alexander Garcia Castro, Angela Norena, Andres Betancourt, and Mark A. Ragan. Cognitive support for an argumentative structure during the ontology development process. In Pro- ceedings of the 9th International Protégé Conference, Stan- ford, California, July 2006. Andy R. Cavagnetto. Argument to foster scientific literacy. Review of Educational Research, 80(3):336-371, 2010. Steve Cayzer. Semantic blogging and decentralized knowl- edge management. Communications of the ACM, 47(12):47- 52, 2004. Hsinchun Chen. AI, e-government, and politics 2.0. [EEE Intelligent Systems, 24(5):64-86, 2009. Timothy Chklovski, Yolanda Gil, Varun Ratnakar, and John Lee. TRELLIS: supporting decision making via argumenta- tion in the Semantic Web. In Proceedings of the Second In- ternational Semantic Web Conference, October 2003. Timothy Chklovski, Varun Ratnakar, and Yolanda Gil. User interfaces with Semi-Formal representations: a study of de- signing argumentation structures. In Proceedings of the 10th International Conference on Intelligent User Interfaces, pages 130-136, 2005. Paolo Ciccarese. Scientific Discourse Relationships Ontol- ogy Specification. Technical report, Massachusetts General Hospital, December 2008. Paolo Ciccarese, Marco Ocana, Sudeshna Das, and Tim Clark. AO: an open Annotation Ontology for science on the Web. In Bio-Ontologies 2010, 2010. [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] Schneider et al. / A Review of Argumentation for the Social Semantic Web 35 Paolo Ciccarese, Elizabeth Wu, Gwen Wong, Marco Ocana, June Kinoshita, Alan Ruttenberg, and Tim Clark. The SWAN biomedical discourse ontology. Journal of Biomedical Infor- matics, 41(5):739-751, October 2008. Chip Cleary and Ray Bareiss. Practical methods for automat- ically generating typed links. In Proceedings of the Seventh ACM Conference on Hypertext, pages 31-41, 1996. Juri Luca De Coi, Norbert E Fuchs, Kaarel Kaljurand, and Tobias Kuhn. Controlled english for reasoning on the Seman- tic Web. In Semantic Techniques for the Web-The REWERSE Perspective, volume 5500 of Lecture Notes in Computer Sci- ence, pages 276-308. Springer, 2009. Jeffrey Conklin. Dialogue Mapping: Building Shared Under- standing of Wicked Problems. John Wiley and Sons, Ltd., 2005. Sudeshna Das, Michael Rogan, Holli Kawadler, Samuel Brin, Stéphane Corlosquet, and Tim Clark. PD Online: A case study in scientific collaboration on the Web. In Workshop on the Future of the Web for Collaborative Science, 19th Inter- national World Wide Web Conference, 2010. Aldo de Moor and Mark Aakhus. Argumentation support: From technologies to tools. Communications of the ACM, 49(3):93-98, 2006. Nicholas Diakopoulos, Sergio Goldenberg, and Irfan Essa. Videolyzer: quality analysis of online informational video for bloggers and journalists. In Proceedings of the 27th Interna- tional Conference on Human Factors in Computing Systems, pages 799-808, 2009. Phan Minh Dung. On the acceptability of arguments and its fundamental role in nonmonotonic reasoning, logic program- ming and n-person games. Artificial Intelligence, 77(2):321- 357, 1995. Paul E. Dunne, Sylvie Doutre, and Trevor J. M. Bench-Capon. Discovering inconsistency through examination dialogues. In Leslie Pack Kaelbling and Alessandro Saffiotti, editors, L/- CAI, pages 1680-1681. Professional Book Center, 2005. Rob Emnals, Dan Byler, John Mark Agosta, and Barboara Rosario. What is disputed on the Web? In WWW 2010, 2010. Rob Ennals, Beth Trushkowsky, and John Mark Agosta. Highlighting disputed claims on the Web. In WICOW 2010 at WWW 2010, 2010. Siamak Faridani, Ephrat Bitton, Kimiko Ryokai, and Ken Goldberg. Opinion Space: A scalable tool for browsing on- line comments. In Proceedings of the 28th International Con- ference on Human Factors in Computing Systems, CHI 10, pages 1175-1184, 2010. Gerhard Fischer, Andreas C. Lemke, Raymond McCall, and Anders I. Morch. Making argumentation serve design. Human-Computer Interaction, 6(3):393-419, 1991. James S. Fishkin. Democracy and Deliberation: New Direc- tions for Democratic Reform. Yale University Press, 1991. Norbert E. Fuchs, Kaarel Kaljurand, and Gerold Schneider. Attempto controlled english meets the challenges of knowl- edge representation, reasoning, interoperability and user in- terfaces. In FLAIRS Conference, pages 664-669, 2006. Marco Aurélio Gerosa, Mariano Gomes Pimentel, Hugo Fuks, and Carlos Lucena. Analyzing discourse structure to coordinate educational forums. In Intelligent Tutoring Sys- tems 2004, number 3220 in LNCS, pages 23-36. Springer- Verlag, 2004. [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] (71) [72] [73] [74] [75] Yolanda Gil and Varun Ratnakar. TRELLIS: An Interac- tive Tool for Capturing Information Analysis and Decision Making. Internal Project Report, USC Information Sciences Institute, 2001. http: //www.isi.edu/expect /web/ trellis/Trellis-—8-01.pdf. Yolanda Gil and Varun Ratnakar. Trusting information sources one citizen at a time. In Proceedings of the First Inter- national Conference on The Semantic Web, pages 162-176. Springer-Verlag, 2002. Thomas F. Gordon, Henry Prakken, and Douglas Walton. The Carneades model of argument and burden of proof. Artificial Intelligence, 171(10-15):875-896, July 2007. Wayne S. Grennan. Informal Logic: Issues and Techniques. McGill Queens University Press, Montreal, Quebec, 1997. H.P. Grice. Logic and conversation. In Syntax and semantics, Vol.3, pages 41-58, 1975. Leo Groarke. Informal logic. In Edward N. Zalta, editor, The Stanford Encyclopedia of Philosophy. Spring 2012 edition, 2012. Claire Grover, Ben Hachey, and Chris Korycinski. Summaris- ing legal texts: Sentential tense and argumentative roles. In Proceedings of the Text Summarization Workshop at HLT- NAACL 2003, Human Language Technology Conference of the North American Chapter of the Association for Computa- tional Linguistics Volume 5, pages 33-40, 2003. Tudor Groza, Siegfried Handschuh, John G. Breslin, and Ste- fan Decker. An abstract framework for modeling argumenta- tion in virtual communities. International Journal of Virtual Communities and Social Networking, 1:35—-47, 2009. Tudor Groza, Siegfried Handschuh, Knud Mller, and Stefan Decker. SALT - Semantically Annotated LaTeX for scien- tific publications. In The Semantic Web: Research and Appli- cations, Proceedings of ESWC 2007, LNCS, pages 518-532. Springer, 2007. Thomas R. Gruber. Towards principles for the design of on- tologies used for knowledge sharing. International Journal Human-Computer Studies, 43(5—6):907-928, 1995. Thomas R. Gruber. Collective knowledge systems: Where the Social Web meets the Semantic Web. Journal of Web Semantics, 6(1):4-13, 2008. Ali Giirkan, Luca Iandoli, Mark Klein, and Giuseppe Zollo. Mediating debate through on-line large-scale argumenta- tion: Evidence from the field. 180(19):3686—3702, 2010. Karl Gyllstrom and Marie-Francine Moens. Clash of the typings: finding controversies and children’s topics within queries. In Proceedings of the 33rd European conference on Advances in information retrieval, ECIR’ 11, pages 80-91. Springer-Verlag, 2011. Stella Heras. Case-Based Argumentation in Agent Societies. PhD thesis, Universitat Politécnica de Valéncia, 2011. Stella Heras, Katie Atkinson, Vicente Botti, Floriana Grasso, Vicente Juliana, and Peter McBurney. Applying argumenta- tion to enhance dialogues in social networks. In Proceedings of the IJCAI Workshop on Computational Models of Natural Argument (CMNA), 2010. Stella Heras, Katie Atkinson, Vicente Botti, Floriana Grasso, Vicente Juliana, and Peter McBurney. How argumentation can enhance dialogues in social networks. In Computational Models of Argument - Proceedings of COMMA 2010, number 216 in Frontiers in Artificial Intelligence and Applications, Information Sciences, 36 [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] Schneider et al. / A Review of Argumentation for the Social Semantic Web pages 267-274. IOS Press, 2010. Stella Heras, Vicente Botti, and Vicente Juliana. On a Compu- tational Argumentation Framework for Agent Societies, vol- ume 6614 of LNCS, pages 123-140. Springer-Verlag New York Inc, 2011. Michael Hoffmann. Logical Argument Mapping (LAM) - A Manual. Working paper, Georgia Tech, 2011. http: //www.spp.gatech.edu/faculty/ workingpapers/wp50.pdf. Michael H. G. Hoffmann. Logical argument mapping: A cognitive-change-based method for building common ground. In Proceedings of the 2nd international conference on Pragmatic Web, ICPW ’07, pages 41-47, 2007. Ken Hyland. Metadiscourse. Continuum, 1st edition, 2005. Joshua E. Introne. Adaptive Mediation in Groupware. PhD thesis, Brandeis University, 2008. Joshua E. Introne. Supporting group decisions by mediat- ing deliberation to improve information pooling. In Proceed- ings of the ACM 2009 International Conference on Support- ing group work - GROUP ’09, page 189, 2009. Alex P. J. Jarczyk, Peter Loffler, and Frank M. Shipman LL. Design rationale for software engineering: a survey. In Pro- ceedings of the Twenty-Fifth Hawaii International Confer- ence on System Sciences, 1992, volume ii, pages 577-586, 1992. Minwoo Jeong, Chin- Yew Lin, and Gary Geunbae Lee. Semi- supervised speech act recognition in emails and forums. In EMNLP: Proceedings of the Conference on Empirical Methods in Natural Language Processing, pages 1250-1259, 2009. David R. Karger and Dennis Quan. What would it mean to blog on the Semantic Web? In The Semantic Web — ISWC 2004, pages 214-228. Springer, 2004. Joel Katzav and Chris Reed. On argumentation schemes and the natural classification of arguments. Argumentation, 18(2):239-259, June 2004. Joel Katzav, Chris Reed, and Glenn Rowe. Argument research corpus. In Proceedings of the 2003 Conference on Practical Applications in Language and Computers, pages 229-239. Peter Lang, 2003. Paul A. Kirschner, Simon J. Buckingham Shum, and Chad S. Carr. Visualizing Argumentation: Software tools for collab- orative and educational sense-making. Springer, | edition, 2003. Mark Klein. Achieving collective intelligence via large-scale on-line argumentation. Technical Report Research Paper No. 4647-07, CCI Working Paper No. 2007-001, MIT Sloan School of Management, April. Mark Klein. The MIT Collaboratorium: Enabling Effective Large-Scale Deliberation for Complex Problems. SSRN eLi- brary, (Research Paper No. 4679-08), December 2007. Graham Klyne and Jeremy J. Carroll. Resource De- scription Framework (RDF): Concepts and abstract syn- tax. W3C Recommendation 10 February 2004, World Wide Web Consortium, 2004. http: //www.w3.org/ TR/rdf-—concepts/. Alistair Knott and Ted Sanders. The classification of coher- ence relations and their linguistic markers: An exploration of two languages. Journal of Pragmatics, 30:135—-175, 1998. Kepa Korta and John Perry. Pragmatics. In Edward N. Zalta, editor, The Stanford Encyclopedia of Philosophy. Summer [93] [94] [95] [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] 2011 edition, 2011. Travis Kriplean, Jonathan Morgan, Deen Freelon, Alan Born- ing, and Lance Bennett. Supporting reflective public thought with Considerlt. In Proceedings of the ACM 2012 conference on Computer Supported Cooperative Work, CSCW °12, pages 265-274, New York, NY, USA, 2012. Travis Kriplean, Jonathan T. Morgan, Deen Freelon, Alan Boring, and Lance Bennett. Considerlt: improving struc- tured public deliberation. In Proceedings of the 2011 annual conference extended abstracts on Human factors in comput- ing systems, CHI EA 11, page 1831-1836, 2011. Markus Krétzsch, Denny Vrandeci¢é, Max Vélkel, Heiko Haller, and Rudi Studer. Semantic Wikipedia. Journal of Web Semantics, 5(4):251-261, 2007. Werner Kunz and Horst W.J. Rittel. Issues as elements of information systems. Working Paper 131, Center for Plan- ning and Development Research, University of California at Berkeley, Berkeley, CA, 1970. Christoph Lange. Semantic Web Collabora- tion on Semiformal Mathematical _ Knowledge. Ph.D., Jacobs University, School of Engineering and Science, Bremen, Germany, Germany, 2010. https://svn.kwarc .info/repos/swim/doc/phd/phd.pdf. Christoph Lange, Uldis Bojars, Tudor Groza, John G. Breslin, and Siegfried Handschuh. Expressing argumentative discus- sions in social media sites. In Proceedings of the ISWC2008 Workshop on Social Data on the Web (SDoW2008), October 2008. Christoph Lange, Tuukka Hastrup, and Stephane Corlosquet. Arguing on issues with mathematical knowledge items in a semantic wiki. In LWA: Lernen, Wissensentdeckung und Adaptivitdt, volume 448 of Technical Report, pages 41-48. Department of Computer Science, University of Wiirzburg, Germany, 2008. Ioan Alfred Letia and Adrian Groza. Contextual exten- sion with concept maps in the Argument Interchange For- mat. In Argumentation in Multi-Agent Systems, volume 5384. Springer, 2009. Anna De Liddo and S. Buckingham Shum. Cohere: A proto- type for contested collective intelligence. In Collective Intel- ligence In Organizations - Toward a Research Agenda, 2010. Claudia Lépea, Pablo Inostroza, Luiz Marcio Cysneiros, and Hernan Astudillo. Visualization and comparison of architec- ture rationale with Semantic Web technologies. The Journal of Systems and Software, 82(8):1198—1210, 2009. Rui Pedro Lourengo and Jo ao Paulo Costa. A note on democ- racy and Group Support Systems concepts. In CISTI: 2010 Sth Iberian Conference on Information Systems and Tech- nologies, pages 1-6, 2010. John Lowrance, Ian Harrison, Andres Rodriguez, Eric Yeh, Tom Boyce, Janet Murdock, Jerome Thomere, and Ken Mur- ray. Template-based structured argumentation. In Knowledge Cartography, Advanced Information and Knowledge Process- ing, pages 307-334. Springer, 2008. John D. Lowrance. Graphical manipulation of evidence in structured arguments. Law, probability and risk, 6(1-4):225, 2007. Ann Macintosh. Moving toward “Intelligent” policy develop- ment? IEEE Intelligent Systems, 24(5):79-82, 2011. Ann Macintosh, Thomas F. Gordon, and Alastair Renton. Providing argument support for e-participation. Journal of [108] [109] [110] {111) [112] [113] [114] [115] [116] (117) [118] [119] [120] [121] [122] Schneider et al. / A Review of Argumentation for the Social Semantic Web 37 Information Technology & Politics, 6(1):43-59, 2009. Clara Mancini and Simon Buckingham shum. Cognitive Co- herence Relations and hypertext: from cinematic patterns to scholarly discourse. In Proceedings of the 12th ACM confer- ence on Hypertext and Hypermedia, pages 165-174, Arhus, Denmark, 2001. Clara Mancini and Simon J. Buckingham Shum. Modelling discourse in contested domains: A semiotic and cognitive framework. International Journal of Human-Computer Stud- tes, 64(11):1154-1171, November 2006. William C. Mann and Sandra A. Thompson. Rhetorical Struc- ture Theory: Toward a functional theory of text organization. Text, 8(3):243-281, 1988. Daniel Marcu. Building up rhetorical structure trees. In AAAL 96 Proceedings, 1996. Daniel Marcu. The Theory and Practice of Discourse Parsing and Summarization. The MIT Press, November 2000. Peter McBurney and Simon Parsons. Dialogue games for agent argumentation. Argumentation in Artificial Intelligence, pages 261-280, 2009. Helena M. Mentis, Paula M. Bach, Blaine Hoffman, Mary Beth Rosson, and John M. Carroll. Development of de- cision rationale in complex group decision making. In Pro- ceedings of the 27th International Conference on Human Fac- tors in Computing Systems, pages 1341-1350, 2009. Raquel Mochales and Marie-Francine F. Moens. Argumen- tation mining. Artificial Intelligence and Law, 19(1):1—-22, 2011. Raquel Mochales Palau. Automatic detection and clas- sification of argumentation in a legal case. PhD thesis, K.U.Leuven, Faculty of Engineering Science, July 2011. Moens, Marie-Francine and De Schreye, Daniel (supervi- sors). Raquel Mochales-Palau and Marie-Francine Moens. Auto- matic argumentation detection and its role in law and the Semantic Web. In Law, Ontologies and the Semantic Web - Channelling the Legal Information Flood, volume 188 of Frontiers in Artificial Intelligence and Applications, pages 115-129. IOS Press, 2009. Sanjay Modgil and Jarred McGinnis. Towards characterising argumentation based dialogue in the Argument Interchange Format. In Argumentation in Multi-Agent Systems, volume 4946. Springer, 2008. Bernard Moulin, Hengameh Irandoust, Micheline Bélanger, and Gaélle Desbordes. Explanation and argumentation capa- bilities: Towards the creation of more persuasive agents. Ar- tificial Intelligence Review, 17:169-222, 2002. Carlos Ches nevar, Jarred McGinnis, Sanjay Modgil, Iyad Rahwan, Chris Reed, Guillermo Simari, Matthew South, Ger- ard Vreeswijk, and Steven Willmott. Towards an Argument Interchange Format. 21(4):293-316, 2006. Susan E. Newman and Catherine C. Marshall. Pushing Toulmin too far: Learning from an argument representa- tion scheme. Technical Report Technical Report SSL-92- 45, Xerox PARC, 1992. http://www.csdl.tamu.edu/ mar- shall/toulmin.pdf. Alexandra Okada, Simon J. Buckingham Shum, and Tony Sherbome. Knowledge Cartography: Software Tools and Mapping Techniques. Springer, 2008. The Knowledge Engineering Review, [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] Tim O’Reilly. What Is Web 2.0: Design Patterns and Business Models for the Next Generation of Software, 30 September 2005. http: //www.oreillynet.com/lpt/a/6228. Bo Pang and Lillian Lee. Opinion mining and sentiment anal- ysis. Foundations and Trends in Information Retrieval, 2(1- 2):1-135, 2008. Jack Park. Boundary infrastructures for IBIS federation: De- sign rationale, implementation, and evaluation. Ph.D. The- sis Proposal KMI-10-01, The Open University, Milton Keyes, 2010. Philippe Pasquier, Iyad Rahwan, Frank Dignum, and Liz Sonenberg. Argumentation and persuasion in the cogni- tive coherence theory. In Proceedings of the 2006 confer- ence on Computational Models of Argument: Proceedings of COMMA 2006, pages 223-234, Amsterdam, The Nether- lands, The Netherlands, 2006. IOS Press. Alexandre Passant. Semantic Web Technologies for Enter- prise 2.0. IOS Press, 2010. Alexandre Passant, Paolo Ciccarese, John G. Breslin, and Tim Clark. SWAN/SIOC: aligning scientific discourse representa- tion and social semantics. In Proceedings of the Workshop on Semantic Web Applications in Scientific Discourse (SWASD 2009), at (ISWC-2009), volume 523. CEUR-WS.org, 2009. Chaim Perelman and Lucie Olbrechts-Tyteca. The New Rhetoric: A Treatise on Argumentation. University of Notre Dame Press, June 1969. Silvio Peroni and David Shotton. Fabio and cito: Ontologies for describing bibliographic resources and citations. [EEE Intelligent Systems, 2011. Preprint. H. Sofia Pinto, Christoph Tempich, and Steffen Staab. DILI- GENT: towards a fine-grained methodology for DIstributed, loosely-controlled and evolvInG engingeering of oNTologies. In Proceedings of the 16th European Conference on Artificial Intelligence (ECAI 2004), August 22nd - 27th, pages 393— 397, August 2004. Henry Prakken. Formal systems for persuasion dialogue. The Knowledge Engineering Review, 21(2):163-188, 2006. ESTRELLA Project. Deliverable 1.4: OWL ontology of basic legal concepts (LKIF-Core). Deliverable IST-2004-027655, University of Amsterdam, January 2007. lyad Rahwan. Mass argumentation and the Semantic Web. Web Semantics: Science, Services and Agents on the World Wide Web, 6(1):29-37, February 2008. lyad Rahwan and Bita Banihashemi. Arguments in OWL: a progress report. In Computational models of argument: Pro- ceedings of COMMA 2008, pages 297-310. IOS Press, 2008. lyad Rahwan, Bita Banihashemi, Chris Reed, Douglas Wal- ton, and Sherief Abdallah. Representing and classifying ar- guments on the semantic web. The Knowledge Engineering Review, 26(04):487-511, 12 2011. lyad Rahwan and Guillermo Simari. Argumentation in Artifi- cial Intelligence. Springer, 2009. lyad Rahwan, Fouad Zablith, and Chris Reed. Laying the foundations for a World Wide Argument Web. Artificial In- telligence, 171(10-15):897—-921, July 2007. lyad Rahwan, Fouad Zablith, and Chris Reed. Towards large scale argumentation support on the Semantic Web. In Pro- ceedings of the 22nd national conference on Artificial Intelli- gence - Volume 2, pages 1446-1451. AAAI Press, 2007. Andrew Ravenscroft, Simon Wells, Musbah Sagar, and Chris Reed. Mapping persuasive dialogue games onto argumen- 38 [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] Schneider et al. / A Review of Argumentation for the Social Semantic Web tation structures. In Symposium on Persuasive Technology & Digital Behaviour Intervention at AISB: Artificial Intelli- gence and Simulation of Behaviour, April 2009. Chris Reed. Representing dialogic argumentation. Knowledge-Based Systems, 19(1):22-31, March 2006. Chris Reed and Glenn Rowe. Araucaria: Software for argu- ment analysis, diagramming and representation. /nternational Journal on Artificial Intelligence Tools, 13(4):961-979, 2004. Chris Reed and Christopher W. Tindale. Dialectics, dialogue and argumentation: An examination of Douglas Walton’s the- ories of reasoning and argument. College Publications, Lon- don, 2010. Chris Reed, Douglas Walton, and Fabrizio Macagno. Argu- ment diagramming in logic, law and Artificial Intelligence. The Knowledge Engineering Review, 22(01):87-109, 2007. Chris Reed and Simon Wells. Dialogical argument as an inter- face to complex debates. [EEE Intelligent Systems, 22(6):60— 65, 2007. Chris Reed, Simon Wells, Katarzyna Budzynska, and Joseph Devereux. Building arguments with argumentation: the role of illocutionary force in computation models of argument. In Computational Models of Argument - Proceedings of COMMA 2010, number 216 in Frontiers in Artificial Intelli- gence and Applications, pages 415-426. IOS Press, 2010. Chris Reed, Simon Wells, Joseph Devereux, and Glenn Rowe. AIF+: dialogue in the Argument Interchange Format. In Computational models of argument: Proceedings of COMMA 2008, pages 311-323. IOS Press, 2008. Alan Ritter, Colin Cherry, and Bill Dolan. Unsupervised modeling of Twitter conversations. In Proceedings of Hu- man Language Technologies: The 11th Annual Conference of the North American Chapter of the Association for Computa- tional Linguistics, 2010. Alan Ritter, Doug Downey, Stephen Soderland, and Oren Et- zioni. It’s a contradiction—no, it’s not: a case study using functional relations. In EMNLP:Proceedings of the Confer- ence on Empirical Methods in Natural Language Process- ing, pages 11-20. Association for Computational Linguistics, 2008. Glenn W. A. Rowe, Chris A. Reed, and Joel Katzav. Arau- caria: Marking up argument. In European Conference on Computing and Philosophy, 2003. Matthias Samwald and Klaus-Peter Adlassnig. The bio-zen plus ontology. Applied Ontology, 3(4):213-217, 2008. Ted J.M. Sanders, Wilbert P.M. Spooren, and Leo Noordman. Towards a taxonomy of coherence relations. Discourse Pro- cesses, 15:1—35, 1992. Oliver Scheuer, Frank Loll, Niels Pinkwart, and Bruce McLaren. Computer-supported argumentation: A review of the state of the art. International Journal of Computer- Supported Collaborative Learning, 5(1):43—102, 2010. David C. Schneider, Christian Voigt, and Gregor Betz. ArguNet-a software tool for collaborative argumentation analysis and research. In Proceedings of the LJCAI Workshop on Computational Models of Natural Argument (CMNA), 2007. Jodi Schneider, Alexandre Passant, and Stefan Decker. Dele- tion discussions in Wikipedia: Decision factors and outcomes. In 8th International Symposium on Wikis and Open Collabo- ration, WikiSym 2012, 2012. [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] Daniel Schober, James Malone, and Robert Stevens. Practical experiences in concurrent, collaborative ontol- ogy building using Collaborative Protégé. In Pro- ceedings of the International Conference on Biomedi- cal Ontology, Buffalo, NY, 2009. Nature Precedings. http://precedings.nature.com/documents/35 17/version/2. Wolfgang Schuler and John B. Smith. Author’s argumenta- tion assistant (AAA): a hypertext-based authoring tool for ar- gumentative texts. In Hypertext: Concepts, systems and appli- cations (Proceedings of the ECHT’90 European Conference on Hypertext), pages 137-151. Cambridge University Press, 1992. John R. Searle. Speech Acts. Cambridge University Press, New York, 1969. Frank M. Shipman and Catherine C. Marshall. Formality con- sidered harmful: Experiences, emerging themes, and direc- tions on the use of formal representations in interactive sys- tems. Computer Supported Cooperative Work, 8(4):333-352, December 1999. David Shotton. CiTO, the citation typing ontology, and its use for annotation of reference lists and visualization of citation networks. In Bio-Ontologies 2009, a Special Interest Group meeting at ISMB 2009, 2009. David Shotton. CiTO, the citation typing ontology. Journal of Biomedical Semantics, \(Supp1 1):S6, 2010. Simon Buckingham Shum. Cohere: Towards web 2.0 argu- mentation. In Computational models of argument: Proceed- ings of COMMA 2008, volume 172 of Frontiers in Artificial Intelligence and Applications. IOS Press, 2008. Simon Buckingham Shum. AIF — use Iraq debate. AIF 2.0 meeting contribution, http://projects.kmi.open.ac.uk/hyperdiscourse/docs/AIF- UseCase-v2.pdf, 2010. Simon Buckingham Shum and Tamara Sumner. JIME: an interactive journal for interactive media. First Monday, 6(2), February 2001. Simon J. Buckingham Shum. The roots of computer sup- ported argument visualization. tion: Software tools for collaborative and educational sense- making, pages 3-21. Springer, 2003. Simon J. Buckingham Shum, Victoria Uren, Gangmin Li, Bertrand Sereno, and Clara Mancini. Modeling naturalistic argumentation in research literatures: Representation and in- teraction design issues. International Journal of Intelligent Systems, 22(1):17-47, 2007. Evren Sirin, Bijan Parsia, Bernardo Cuenca Grau, Aditya Kalyanpur, and Yarden Katz. Pellet: A practical OWL-DL reasoner. Web Semantics: Science, Services and Agents on the World Wide Web, 5(2):51—53, 2007. Mark Snaith, Joseph Devereux, John Lawrence, and Chris Reed. Pipelining argumentation technologies. In Computa- tional Models of Argument - Proceedings of COMMA 2010, number 216 in Frontiers in Artificial Intelligence and Appli- cations. IOS Press, 2010. Mark Snaith, John Lawrence, and Chris Reed. Mixed ini- tiative argument in public deliberation. In Proceedings of the Fourth International Conference on Online Deliberation, 2010. M. South, G. Vreeswijk, and J. Fox. Dungine: A JAVA Dung reasoner. In Proceeding of the 2008 conference on Computa- tional Models of Argument (COMMA 2008), pages 360-368. case: In Visualizing Argumenta- (171) [172] [173] [174] [175] [176] [177] [178] [179] [180] [181] [182] [183] [184] [185] [186] [187] Schneider et al. / A Review of Argumentation for the Social Semantic Web 39 IOS Press, 2008. Norbert A. Streitz, Jorg Hannemann, and Manfred Thiiring. From ideas and arguments to hyperdocuments: Travelling through activity spaces. In Proceedings of the second an- nual ACM conference on Hypertext, HYPERTEXT ’89, pages 343-364, 1989. Daniel D. Suthers. Empirical studies of the value of conceptu- ally explicit notations in collaborative learning. In Knowledge Cartography, Advanced Information and Knowledge Process- ing, pages 1-24. Springer, 2008. Maite Taboada and William C. Mann. Applications of Rhetor- ical Structure Theory. Discourse Studies, 8(4):567-588, Au- gust 2006. Christoph Tempich, H. Sofia Pinto, York Sure, and Steffen Staab. An argumentation ontology for DIstributed, loosely- controlled and evolvInG engineering processes of oNTologies (DILIGENT). In The Semantic Web: Research and Applica- tions, pages 241-256. Springer, 2005. Simone Teufel and Marc Moens. Summarizing scientific arti- cles: Experiments with relevance and rhetorical status. Com- putational Linguist., 28(4):409-445, 2002. Paul Thagard. Explanatory coherence. Behavioral and Brain Sciences, 12(3):435-502, 1989. Terri Attwood Anita de Waard Tim Clark, Paolo Ciccarese and Steve Pettifer. A round-trip to the annotation store: Open, transferable semantic annotation of biomedical publications. In Beyond the PDF workshop, University of California San Diego, 2011. Francesca Toni and Paolo Torroni. Bottom-up argumentation. First International Workshop on the Theory and Applications of Formal Argumentation, 2011. Paolo Torroni, Federico Chesani, and Marco Gavanelli. Ar- gumentation in the Semantic Web. IEEE Intelligent Systems, 22(6):66-74, 2007. Stephen Edelston Toulmin. The Uses of Argument. University Press, Cambridge [Eng.], 1958. Douglas P. Twitchell, Mark Adkins, Jay F. Nunamaker Jr., and Judee K. Burgoon. Using speech act theory to model conversations for automated classification and retrieval. In (LAP’04):Proceedings of the 9th International Working Con- ference on the Language-Action Perspective on Communica- tion Modelling, pages 121-129, 2004. Manolis Tzagarakis, Nikos Karousos, Nikos Karacapilidis, and Dora Nousia. Unleashing argumentation support systems on the Web: The case of CoPe_it! In Web Science 2009, Athens, Greece, March 2009. F. H van Eemeren, editor. Advances in Pragma-Dialectics. SicSat/Vale Press, 2002. Frans H. van Eemeren and Rob Grootendorst. Speech Acts in Argumentative Discussions: A Theoretical Model for the Analysis of Discussions Directed Toward Solving Conflicts of Opinion. de Gruyter Mouton, June 1984. Frans H. van Eemeren and Rob Grootendorst. Argumenta- tion, Communication and Fallacies: A Pragma-Dialectical Perspective. Erlbaum, 1992. Frans H. van Eemeren and Rob Grootendorst. A systematic theory of argumentation: the pragma-dialectical approach. Cambridge University Press, 2004. Frans H. van Eemeren, Rob Grootendorst, and Fran- cisca Snoeck Henkemans. Argumentation: analysis, evalua- tion, presentation. Erlbaum, 2002. [188] [189] [190] [191] [192] [193] [194] [195] [196] [197] [198] [199] [200] [201] [202] [203] [204] [205] [206] [207] [208] Marilyn A. Walker, Pranav Ananda, Rob Abbott, Jean E. Fox Tree, Craig Martell, and Joseph King. That’s your evidence?: Classifying stance in online political debate. Decision Sup- port Systems, in press 2012. Doug Walton, Katie Atkinson, Trevor Bench-Capon, Adam Wyner, and Daniel Cartwright. Argumentation in the frame- work of deliberation dialogue. Arguing Global Governance: Agency, Lifeworld and Shared Reasoning, page 210, 2010. Douglas Walton. Plausible argument in everyday conversa- tion. SUNY Press, 1992. Douglas Walton. Argumentation Schemes for Presumptive Reasoning. Routledge, 1995. Douglas Walton. How can logic best be applied to arguments? Logic Journal of the IGPL, 5(4):603-614, July 1997. Douglas Walton. Fundamentals of Critical Argumentation. Cambridge University Press, October 2005. Douglas Walton. Dialog Theory for Critical Argumentation. John Benjamins, 2007. Douglas Walton. Informal Logic: A Pragmatic Approach. Cambridge University Press, 2nd edition, 2008. Douglas Walton. Types of dialogue and burdens of proof. In Computational Models of Argument - Proceedings of COMMA 2010, number 216 in Frontiers in Artificial Intelli- gence and Applications, pages 13-24. IOS Press, 2010. Douglas Walton, Chris Reed, and Fabrizio Macagno. Argu- mentation Schemes. Cambridge University Press, 1st edition, August 2008. Douglas N. Walton. Commitment in dialogue : basic concepts of interpersonal reasoning. State University of New York Press, Albany, 1995. Barbara Warnick. Rhetoric Online: Persuasion and politics on the World Wide Web. Peter Lang, 2007. Bonnie Webber, Markus Egg, and Valia Kordoni. Discourse structure and language technology. Natural Language Engi- neering, pages 1-54, FirstView Article 2011. Simon Wells, Colin Gourlay, and Chris Reed. Argument blog- ging. In Proceedings of the IJCAI Workshop on Computa- tional Models of Natural Argument (CMNA), 2009. Simon Wells and Chris Reed. MAgtALO: An agent-based system for persuasive online interaction. In AJSB 2008: Arti- ficial Intelligence and Simulation of Behaviour, 2008. Simon Wells and Chris Reed. A domain specific language for describing diverse systems of dialogue. Journal of Applied Logic, 2011. to appear. Robin Wendler. The eye of the beholder: Challenges of image description and access at harvard. In Metadata in Practice, pages 51-69. ALA Editions, 2004. Terry Winograd. A Language/Action perspective on the de- sign of cooperative work. Human-Computer Interaction, 3(1):3-30, March 1987. Adam Wyner, Tom Van Engers, and Kiavash Bahreini. From policy-making statements to first-order logic. In Electronic Government and the Information System Perspective, volume 6267 of LNCS, pages 47-61. Springer, 2010. Adam Wyner, Raquel Mochales-Palau, Marie-Francine Moens, and David Milward. Approaches to text mining ar- guments from legal cases. In Semantic Processing of Legal Texts, number 6036 in Lecture Notes in Computer Science, pages 60-79. Springer, 2010. Adam Wyner, Jodi Schneider, Katie Atkinson, and Trevor Bench-Capon. Semi-automated argumentative analysis of on- 40 [209] [210] [211] [212] Schneider et al. / A Review of Argumentation for the Social Semantic Web line product reviews. In Proceedings of the 2012 conference on Computational Models of Argument (COMMA 2012). Adam Wyner and Tom van Engers. A framework for enriched, controlled on-line discussion forums for e- government. In eGOV 2010, pages 1-11, 2010. Ka-Ping Yee. Zest: Discussion mapping for mailing lists. In Proceedings of the 2002 ACM conference on Computer- Supported Cooperative Work, 2002. J. Young, C. Martell, P. Anand, P. Ortiz, and H. T. Gilbert IV. A microtext corpus for persuasion detection in dialog. In Workshops at the Twenty-Fifth AAAI Conference on Artificial Intelligence, 2011. Fouad Zablith. ArgDF: Arguments on the Seman- tic Web. Master’s, The British University in Dubai Jointly with The University of Edinburgh, February 2007. http://www.argdf.org/source/. Schneider et al. / A Review of Argumentation for the Social Semantic Web i Appendix A. Tools A.l. List of social and Semantic Web tools considered In this section we draw from our review of thirty- seven online argumentation tools: AGORA: Partici- pate - Deliberate, ArgDF, Arguehow, Argument Blog- ging, Argumentum, Argumentations.com, Argunet, Avicenna, bCisiveOnline, Belvedere, Cabanac’s an- notation system, Climate CoLab, Cohere, Compet- ing Hypotheses, Considerlt, ConvinceMe, CoPe_it!, CreateDebate, Debate.org, Debategraph, Debatepe- dia, Debatewise, DiscourseDB, Dispute Finder, Hyper- news, Living Vote, MAgTALO, Opinion Space, Online Visualisation of Arguments, Parmenides, PDOnline, REASON, Riled Up!, SEAS, Trellis, TruthMapping, and Videolyzer. For further details about our inclusion criteria, see Section 9, page 21. A.2. AGORA: Participate - Deliberate Michael Hoffman’s system, AGORA: Participate - Deliberate [78], uses Logical Argument Mapping [77](Figure 24 on page ii), providing support for rep- resenting deductively valid arguments, using one of seven schemes: modus ponens; modus tollens; disjunc- tive syllogism; not-both syllogism; conditional syllo- gism; equivalence; and constructive dilemma. It relies on concept mapping software called CmapTools’’. A.3. ArgDF ArgDF’® is a Semantic Web-based argumentation system using the AIF-RDF ontology described above [139,138,212]. ArgDF uses Sesame RDF for storage and querying and Phesame for communicating with the Sesame through PHP pages. A.4. Arguehow ArgueHow” (Figure 25) is a argument-based dis- cussion board aimed at a general audience. Its purpose is to help find the best points supporting a position. Discussion points are sorted by votes for (“Creds’) and against (‘Cruds’) them. ArgueHow has a unique way ™Mhttp://cmap.ihme.us/ Bhttp://argdf.org/ PMretp://arguehow.com/ of handling reputation: users start with a reputation of 50, which increases or decreases according to the votes their points accrue. Votes are weighted: for instance, points with 10 ‘cred’ or ‘crud’ votes change less in response to further votes, and votes on users’ first 20 discussion points affect their reputation less than later contributions, allowing them to learn the system. HOW. ajpssnt ior ad geist login Seatbelt Laws FOR heer A thd Fig. 25. Arguhow offers structured discussion. A.5. Argument Blogging The idea of argument blogging was proposed by Wells, Gourlay and Reed [201] as a way to bring blogs into the WWAW, based on standard Web technolo- gies, and augmented by argument specific technolo- gies. In addition to AIF, argument blogging relies on the AIF Database (AIFDB) and Dialog Game Descrip- tion Language (DGDL). AIFDB is a MySQL database for storing AIF documents which can be serialized as RDF and accessed via a RESTful Web service. DGDL [140,203] is a grammar for describing the rules of dia- logue games. Argument blogging uses text from the current Web as a departure point for the WWAW. When browsing the Web, users select text and click a JavaScript book- marklet, to indicate whether they will attack an infer- ence, support or refute the selected text. This gener- ates a fragment of embeddable JavaScript the user can paste onto his/her blog. Once a blogger opts in to the WWAW by adding JavaScript to a webpage, the page displays a badge which links back to argument blog- ging server, where the distributed dialog can be visual- ized or exported as text. @ wall protects Israel from attacks launched in Gaza ¢— therefore ——| regarding protection by a wall, the situation in the West Bank would be exactly the same as in Gaza therefore the fact that there was not a single successful attack into Israel from Gaza is caused by the fact that there is if there are terrorists who try to destroy Israel in Gaza and in the West Bank, situation for Israel is the same with regard to both places Schneider et al. / A Review of Argumentation for the Social Semantic Web although Gaza is "home to more terrorist organizations <@—therefore— than the West Bank,” they did "not launch a single | successful attack into Israel in three years of fighting." objects a wall around Gaza there were more than 1000 rocket attacks from Gaza against Israel since 2000 re clarifies and supports although the situation is not “exactly” the same, there |¢— supports —) were more than 250 "terrorist attacks entering Israel from then the security the West Bank since 2001" Fig. 24. A sample LAM map, from [78]. Earlier work on semantic blogging predates the WWAW but focused more attention on the visualiza- tion of reply graphs of messages from multiple blogs [84] or the possibilities for inference [38]. A.6. Argumentum Argumentum® is an argument-based discussion site aimed at airing discussions. Debaters add topics and their arguments are colored to indicate the support- ing (green) and opposing (red) arguments (Figure 26). Comments, but not their replies, are similarly colored to indicate agreement or disagreement. Users some- times want to agree or disagree without leaving com- ments; currently this leaves a default comment that says “Type the reason why you oppose..." Argumentum’s most unique feature is that users can put their “2 cents" in literally: credibility, earned with good arguments, is measured in ‘cents’ and can be spent to influence a debate result. Users can also con- tribute arguments without starting from the Argumen- tum website, using bookmarklets*! or through Gmail and Facebook*”. Further, loggers and publishers can also contribute using Argumentum buttons or widgets. http: //arg.umentum.com/ SInttp://arg.umentum.com/share http: //arg.umentum.com/wiki/ more-ways-to-argue Argumentun »......; | Healthcare is a human right nf Coppase Gowonr |_| Fep || sew trmcking| | matateet (ood to pos search for. ‘Sort by relevance | #) | any pasitics | [Xavinconner [LOpponn] ‘Wiha are “human rights”7The only rights we should be concerned about are the rigivts promised to us by our Constitution. We do NOT need a kaeger government Let free market capstiliam do it's job. The goverrs having control aver heal: 6 e100 | | Tenyspeaman (Oppo | | | Acnger can nat oe refuses, What | moan by this is a “rghit” # tomy dependent on your awn acter. have me] “right to froe speech, My ectons may reault in my Seath but na ong can slap me from doing It you are on a sesened land an Comment Femaink 0 670 | | Nelicitsin t 1 Since averyone has the right to Me, Eberty and the pursuit of hapsiness, that should not be crcumscribed by the abilty to pay the extreme fees charged by the fascists {I mean the insurance companies). The LES is the ‘only Indhcatrial nation that do. nt Peemannk 40 ¢1 Fig. 26. In Argumentum, users can indicate support for an argument with money. The left-hand color bars indicate the supporting (green) and opposing (red) arguments. A.7. Argumentations.com Argumentations®* serves analysts who want to de- velop arguments collaboratively. Arguments, which are classified as either claims or open-ended issues, can be added or edited; an example is shown in Fig- ure 27 on page iv. To help suggest topics and build arguments, users can import news stories and extract statements (declarative sentences) from stories. Argumentations offers several unique features. First, arguments—whether claims or open-ended issues—are evaluated depending on their type. Claims are evalu- Shttp://www.argumentations.com/ Schneider et al. / A Review of Argumentation for the Social Semantic Web ili ated with a truth value and confidence. Open-ended is- sues are evaluated based on Desirability, Importance, Volatility, Likelihood, and Confidence. Second, along with tag clouds, Argumentations uses ‘tag spheres’ (Figure 28). Further, arguments can be opened in Sil- verlight. Finally, they offer some interesting tutorials which display mindmaps™. United Nations USA- KGI - Energy . . Water Climate . _ Change Environment Oil a “China Economy Fig. 28. The global warming ‘tag sphere’ from Argumentations. A.8&. Argunet Argunet [154] is a desktop tool®> coupled with an open source federation system for sharing argument maps. A public server, Argunet.org®*, allows authors to make maps public or restrict viewing and/or edit- ing to a specified group. Connecting to other servers is also possible; this focus on federation, makes Argunet unique. Argunet also has other unique features. Argunet is a multi-lingual environment which records the language of the map. Maps published at Argunet.org, must be released under the CC-BY license. An extensive online manual provides instruction, and they promote embed- ding debates. Users also have significant control over the presentation of arguments, such as colors and de- scriptions of different argument families. Related maps can be published in series, as shown in Figure 29(a). In the argument map representation, each node can be opened up to reveal a matrix listing which other ar- guments support, attack, are supported by, and are at- tacked by the given node (Figure 29(b)). Argunet ap- pears to support incremental formalization since ar- guments can be quickly sketched or reconstructed as premises and conclusions. “http: //www.argumentations.com/ Argumentations/Help/Tutorials/Tutorials.aspx Shttp: //www.argunet.org/editor/ http: //www.argunet .org/debates/ Phase 2 Nodes: 16 The substantiation that the fossilized plants Eqges: 17 lie at the top of the Greywacke triggered a Inferential Density: new anomaly. On the one hand, the fossil criterion (H1) seemed to imply that the top ~ layers of the Greywacke are Carboniferous (Fossil-clock-argument). On the other hand, all British Carboniferous oaal ... Phase 3 Nodes: 20 s Murchison and his colleagues assembled Edges: 23 empirical evidence from all over Europe, Inferential Density: hoping to obtain clues about the correct dating of the Devon Greywacke. During an — expedition to Russia in 1840, they found, (a) Coe ee a ed or Preset i © Fossils etsemtere found in Carboniferous strata oceur (only) at the top of the Davon Greywacke os (ern tn ns teu Goreme, soxcrs emen oe pees een) on) | (b) Fig. 29. Argunet can show an (a) overview of several related argu- ment maps; and (b) in each individual map, nodes can be opened up to show arguments they support, attack, are supported by, and are attacked by. A.9. Avicenna Rahwan and Banihashemi’s OWL-based argumen- tation system Avicenna (Figure 30 on page iv) was demonstrated at COMMA 2008 [135] and recent de- scriptions and screenshots appear in [136]. Extending the work of ArgDF, Avicenna is a Web-based system using Jena[31], ARQ*’, and Pellet [167]. Since OWL supports inference over transitive properties, Avicenna can support argument chaining, such as retrieving all arguments that directly or indirectly support a given conclusion. Avicenna is also used to infer the classi- fication hierarchy of argument schemes: for example, an appeal to expert opinion is a specialization of an argument from position to know. A.10. bCisive Online bCisive Online®® is an online argument mapping and spatial hypertext environment for real-time col- laboration and team problem-solving (Figure 31(a) on http: //jena. sourceforge.net /ARQ/ Shttp://www.bcisiveonline.com/ Water oy ArgumentatiOAgs per, EE eae) Schneider et al. / A Review of Argumentation for the Social Semantic Web war on l!érror War Crimes rrorism Argumentations [ Statements | Stories | Tags | Time! lizatior Rus Palestinian- Israeli ‘Conflict fghanista Climate ch lorth Korera ange Pee ne bce What is Argumantations.com? Contribute (© voter 2 osition #0 Develop Your Argument impor » story Manoge Your Quicklist Give Feedback Zimbabwe Argument Healthcare Is A Human Right -- According te Sameer Dessani, Amnesty International Background and Context (show) Argument Tree [mse Healthcare Is A Human Right View bn Sibvarfight = Current single-payer proposals come much closer to fulfilling the human right to health care than the market-based reform plans. -- According to two of Amnesty Internationalé€""s ally organizations, the National Economic and Social Rights Initiative and the National Health Law Program >> @aCoeEveryone has the right to a standard of living adequate for the health and well-being of himseif and of his family, including food, clothing, housing and medical care and necessary social services, and the right to security in the event] of unemployment, sickness, disability, widowhood, old age or other lack of Report this... Tags ‘Amnesty International, Finance, Health, Healthcare, Human Rights, Obama, Politics, US Healthcare, USA, Other Argumentations by this Author... Other suggested Argumentations... Embed this Argumentation Map on your website... livelihood in circumstances beyand his control 1.4€ -- Universal Declaration of | Human Rights, Article 25, United Nations >> | MiHealth care should not be considered a human right. >> Fig. 27. In Argumentions, colored dots indicate the supporting (green) and opposing (red) arguments. Details of argument Tipping lowers self esteem This argumet ie an instance of Appeal te Expert Opinion and has been crested by Mona Haidar a See zw io eee S ins ac 1 Ph yee png pe ne oa pean) Z ‘=o none el pPremiso: oS (Dr Phil 8 an expert wn 3 Paychology oa Al a = Or Poa acnate gt SP —> SEITEN tg ie a ete “ Al Fig. 30. Avicenna uses Walton’s critical questions and argument schemes [136]. page v). Aimed at the business market and individual decision-makers, bCisive Online is a commercial prod- uct from AusThink, the makers of the Rationale desk- top tool; the free option allows up to three users to collaborate, or users can upgrade with a monthly sub- scription fee. bCisive Online is unique in that it is in- tended for real-time use with audio conferencing. One person edits the map at a time, adding nodes and con- nections between nodes (Figure 31(b) on page v) while others can point with their cursor or request editing control. Maps can be embedded in blogs (which allows viewers to pan, zoom, hide and show parts of the map) or exported as PowerPoint. Snapshots can be saved as history items, to allow restoring to or reviewing a pre- vious map. A.ll. Belvedere Belvedere®’ is open source software for problem- based collaborative learning. It provides multiple http: //belvedere. sourceforge.net/ Schneider et al. / A Review of Argumentation for the Social Semantic Web v Situation buy the Suzuki == Ta (b) Fig. 31. (a) Collaborative maps for bCisive Online can be used for decision-making and requirements analysis. (b) bCisive Online’s node types show the kinds of discussions that it facilitates. views, such as tables, graphs, and argument maps, of the same topic (Figure A.11). It has been extensively investigated in studies of computer-supported collabo- rative learning [172]. themes | climate change f coral Bleaching | Ser xa fel) =| | gt ad a A Fa (Fetintios or via c themes | climate change [ coral bleaching | (“Delete ) (Tadd Data) {CAdd Hypothesis) (7 Print Preview) hemmal stress duet. Pallution or irritant Disease UV Stress z / oS / Sr | snecies specific = = = Ca # - # temperature increas + Ozone depletion has... « Fig. 32. Belvedere has both argument maps and tables to help orga- nize evidence in collaborative learning. A.12. Cabanac’s annotation system Cabanac used a Java-based system” to research so- cial validation of the arguments in comments [28]. Users did not contribute new content to an ongoing public debate, but analyzed the argumentative status of document comments. Uniquely, sliders were used to indicate the extent to which items were refuted, neutral, or confirmed (Figure 33 on page vi). In ef- fect, users were asked to synthesize the discussion. Aggregated information was not viewed by the users, but held by the experimenter. However, in principle, this approach could be used to promote collaborative sensemaking not just of annotations but also of debate. A.13. Climate CoLab The Climate CoLab?! is a deliberation platform un- der development at MIT, building on former projects such as the Deliberatorium and the ClimateCollabato- rium [88,89,71]. The community runs an annual con- test to gather proposals for mitigating global warming from the general public; once proposals are filtered by experts, everyone is invited to discuss the finalists. Users deliberate in the Positions tab, which facili- tates constructing an argument map, voting, and com- menting on each of five key topics. Moderators are ex- pected to review comments and add new ideas to the argument map; users can also add Pros, Cons, and Is- nttp://www.irit.fr/~Guillaume.Cabanac/ expe/ Unttp://climatecolab.org/ vi Schneider et al. / A Review of Argumentation for the Social Semantic Web According to its replies, this tem is. Alice : "In the near future, digital documents will completely replace paper.” | ae Ges) Ges) Oo? OA OF Fig. 33. Cabanac had users flag items (refuted, neutral, confirmed) and indicate their types (question, modification, example). sues directly to an argument map. The Climate CoLab is unique for integrating argument maps into a larger debate, and for its moderator support, which allows users to benefit from argument maps without necessar- ily needing to understand how to edit them. A.14. Cohere Cohere is open source software for sensemaking which integrates annotation and argumentation for the general public [162,101]. At the Cohere website”, users can view and create maps, or import them from the Compendium desktop software. Maps consist of ideas, which users can add directly on the site (Fig- ure 35), draw from Cohere’s global pool of public ideas, or clip via a Firefox plugin while browsing. aunty ‘Seee) (Caren Fig. 35. Adding an idea to Cohere. Cohere is unique for its integration with the Com- pendium desktop software, its incorporation of social bookmarking, and the ability to mark information as private, public, or shared with a group. Cohere also of- fers an API?. 2nttp://cohere.open.ac.uk/ Bhttp://cohere.open.ac.uk/help/code-doc/ A.15. Competing Hypotheses Competing Hypotheses” is open source analysis software based on the CIA methodology “Analysis of Competing Hypotheses" (ACH). The software sup- ports breaking down information into hypotheses, evi- dence, and analysis, which are entered into a matrix as shown in Figure 36(a) on page vii. The matrix can help visually indicate the most likely and least likely scenar- os.?> Multiple analyses can be combined to provide a group view (Figure 36(b) on page vii), or compared pairwise. Competing Hypotheses has persistent chat (essentially a comment thread) for the entire project as well as message boards for each hypothesis, evidence item, and evidence-hypothesis pair. We excluded ear- lier ACH implementations such as PARC ACH”. Un- like these systems, Competing Hypotheses has a test- ing server’ which allows online collaboration. It is unique for its visualization structure and its use of both individual and group information. A.16. Considerht Considerlt® [94] is a new open source deliberation platform under development at the University of Wash- ington. It powers the Living Voters’ Guide”, a delib- eration and voter-information platform for Washington State voters. What is unique is the possibility to drill down to un- derstand other voters’ perspectives. In addition to see- ing pros and cons on an issue from all voters, regard- “http: //competinghypotheses.org/ °5More sophisticated ACH-based software uses matrices as input to Bayesian probabilistic reasoning. Shttp://www2.parc.com/istl/projects/ach/ ach. html Thetp: //groups.google.com/group/ach-users/ browse_thread/thread/d87abec4ddf 8be6c0 Shttp://www.livingvotersguide.org/ considerit Snttp://www.livingvotersguide.org/ Schneider et al. / A Review of Argumentation for the Social Semantic Web Proposals: +show help vii ? Should developed countries provide funding to help developing nations address climate change? Position Q ©. Yes: Developed countries should 1 oe © provide funding. 9. Rich countries created the 0 eo problem Yq Costto developed world 0 will be relatively small ~ Public in developed nations 0 © will not support such transfers ef, No: Developing countries should 0 wy Vo pay theirown way Each nation should take 0 responsibility for itself ? Question Should developed countries provide funding to help developing nations address climate change? A key issue that emerged at the Copenhagen climate talks was whether developing countries would provide financing to help developing nations defray the cost of emission reductions and adaptation. For more, see Financial transfers in climate negotiations References: 1. Project Catalyst, Briefing Paper: Overall Financing Needs, December 2009 (4 page pdf file) OC sehrow hap Omments It is difficult to figure out exactly how much developed nations should compensate “Developed countries. 0 paid their way — Without financing, 0 ping nations for and There is starting to be more research on the exact costs of damages to the economy from climate change, though. Here's a paperon Temperatures and cyclones strongly associated with economic production in the Caribbean and Central America http://swww.pnas.org/content/107/35/15367,full By yangr on 10/25/10 2:36. AM (a) In debates, an argument map appears on the Left, Argument maps have four elements: + P issue + positions on theissue * @ arguments for * & arguments against Click on any item to view more detail, You can vote (9 ) on one position per issue. To add to the summary,log in and click the Advanced button. At the right, you may comment ((—}) on any item.See more. Hide this message (b) Fig. 34. At Climate CoLab, (a) the positions tab shows an argument map which users can edit or comment on. (b) argument maps are introduced with contextual help. Personal Matrix 3 Edit your consistency scores Duplicate personal matrix Alrplane o emg Credible Eanalstent Create Consistent Credivle ——Consistemt Se ting ‘Comraits don't widen ax ‘missile’ climbs (Compare the ranngs of { tester 1B) wath: ( Marthew ourton 1) (Compare) ‘View others’ personal matrices: | Matthew aurtan i) ( View...) (a) Group Matrix 3 Consensus Guage: Ce] a Ce a Consensus 1 Consensus § Miké Dispuse Mild Dispute Plumes are dispersing while source is still kn view LU ere head ke aes Masta cuued ‘Consensus N Consensus [iC Consensus 1 : — ‘Compare the ratings of (Compare. ) ‘ew others’ personal matrices: | Matthew Burton 689 (| View. (b) Fig. 36. In Competing Hypotheses, (a) each individual’s analysis is represented in a consistency matrix; (b) multiple analyses can be combined to create a group matrix. In the group view, darker shades of purple indicate more disagreement. viii Schneider et al. / A Review of Argumentation for the Social Semantic Web Explore what other voters think about 1053. ii] Explore what other voters think about 1053. ‘Cohen 1 ate sec he Ray pe Ina Ee Gro Deve ‘Cilek oe a bar is ne scene ct Poe hry paints that the group beskeves ig oe ec PED alee Si eesti emeseieien eo basemen ae (a) ‘om your rat a come My en scarce Se 1053 operons of rosa nee we undectaed on innate 908 (b) Fig. 37. The Living Voters’ Guide compiles pro and con lists on each issue. They give (a) an overview of what all voters think about the issue; as well as (b) the key points for undecided voters. less of their stance, (Figure 37(a)), the Living Voters’ Guide can show the key points for a particular group of voters (Figure 37(b)), such as those undecided on the issue or strongly supporting it. This can help users understand what makes an issue controversial. Users indicate how they feel about an issue before and af- ter reading an argument (deliberative polling), which could also be used to find the most convincing argu- ments. A.17. ConvinceMe ConvinceMe! is a competitive debating environ- ment which uses a point scheme and user rankings to motivate contributions to several types of debates. In the King of the Hill game, the most popular choice (and the debater who suggested it) wins. Battles are one- on-one debates between two users, who add arguments and evidence in hopes of getting readers’ votes; the debate ends when one side gets a pre-agreed number of votes. Open debates (Figure A.17) are ongoing and accept pro or con arguments from any registered user, as well as rebuttals to existing arguments; users con- vinced by an argument vote for it. These various types of debate games make ConvinceMe unique. A.18. CoPe_it! CoPe_it!!®! [182] is a spatial hypertext environ- ment for collaboration, aimed at the learning and e- Mnttp://www.convinceme.net/ lnttp://copeit.cti.gr/ Ninjas versus Pirates Te COL eas Add an Argument Add an Argument para conase moses (on) ETO. on H a ‘st cominced 1 tan 03. 2007 are en, 2007 Lcinanal Ninjas have foe advarsage on land, ba Although, raped ta mci thet pais tse matoars. A Unt fs 2 ins, and than youll navee sea fem Becauka you hive.na people Not caly fst but they ate nar bound by marality Uke ninja aaa. And Pair eagas. Foran wil Kl ARYEAR, BI anys, fr By MNES Buon saa, crams are dangerous. But ainja could anes on board [And pirates are too loud. You and say yar, and ate usually dra, it ‘al ninyat zz 0) Wibur: 61 convinced: Rebuttal dan 02, 2007 An, tee anachroniam that in te ninjal Are thar aver: any ninja taf in 1 gh as pina? | Am pews Bee ALS PUR PinRE ath samyrni (epaaking of which, in she masta TH ninjas and saemuraia Guke it out, and Lam gewty teat prates aw and'wore historical horntes poopie kllng and ‘kure Mat was anachmnisec wa) | aay NAY! stealing from innccent peopie. Ninjas were people Shi fot what waa ight Drair medio and farina, And Seay are way Rote (Gohnny upp ame excenton Rebuttal es, 2087 ‘Yet mogem pirates abound! ee The Gutiaw Sea:.A Wosd of Freedom, Chace. and Cee by Wihiam Langowibsche. Pirie st hk aan, aven in a word of amm bombe and piacnemas. foes, piiwies peobably ose Biackbemes Backbumes, eyupaiches and Fig. 38. In ConvinceMe’s Open Debates, users can vote for an argu- ment that convinced them government domains. Users can form groups to share maps, but communicate only through email on the site. Maps can be imported from Compendium, and entire discussions from external webforums in phpNuke for- mat can be imported using a URL. One unique aspect of in CoPe_it! is its approach to incremental formalization. CoPe_it! transforms the user-created informal spatial hypertext view (Fig- ure 39(a) on page ix) into an issue chart Figure 39(b) on page ix according to rules shown in Figure 39(c) on page ix. Users can also customize the transformation tules. Schneider et al. / A Review of Argumentation for the Social Semantic Web ix Coan Seeoes C60, O06 Workspace Issue Title Idea Alternative Items Linked => er ) Other a ¢ I : @& Ignore (b) (c) Fig. 39. CoPe_it! has (a) an informal spatial hypertext view; and (b) a formalized view, created by (c) automatically transforming items. A.19. CreateDebate CreateDebate!” is a social debate community, aimed at the general public as well as primary and sec- ondary school classes'*?. The highest-rated arguments are shown at the top, based on user votes (and ignoring the down votes), which are also used to determine a point score for the user. They offer bookmarklets and promote JavaScript buttons to webmasters!%*. Some unique features are that the debate moderator can add a ‘Topic Research’ section with RSS feeds from other sites, and that, in addition to pro/con debates, Cre- ateDebate has Perspective debates, which generally have more than two sides, are scored based on user- applied tags. A wordcloud and various statistics (Fig- ure 40), including the language grade level, average word lengths, and vocabulary overlap are calculated for each debate. A.20. Debate.org Debate.org!® is a social networking site for debate lovers. Debates take place between two members and have four cycles: the challenge period, debating pe- tiod, voting period, and post voting period. The de- bating period consists of 1-5 time-limited rounds in which debaters post arguments. While comments can be added at any time, votes are only accepted dur- ing the voting period. Voting involves choosing one of the debators (or ‘tied’) for each of the following six http: //www.createdebate.com/ 1Bnttp://www.createdebate.com/about/sites/ school M4http://www.createdebate.com/share/buttons IShttp://debate.org/ relieve god Fig. 40. At CreateDebate, users add and comment on pro and con arguments. questions: (1) Agreed with before the debate: (worth 0 points) (2) Agreed with after the debate: (worth 0 points) (3) Who had better conduct: (worth 1 point) (4) Had better spelling and grammar: (worth 1 point) (5) Made more convincing arguments: (worth 3 points) (6) Used the most reliable sources: (worth 2 points) . Points are awarded, with the most importance given to using reliable sources and making convincing argu- ments. Another unique feature is Debate.org’s focus on user profiles, where various user details are displayed in- cluding information such as income, location, ideol- ogy, gender, president, religion, and who they are in- terested in and looking for. These can be used to search for people with particular profile attributes, and ag- gregate user demographics! are also available. De- bate.org also determines the percentage to which other members agree with you on “the big issues" (cultural, \6nttp: //www.debate.org/about/demographics / x Schneider et al. / A Review of Argumentation for the Social Semantic Web religious, and political hot topics). Individual members are also ranked by their percentile, based on the out- comes of previous debates. The Instigator Mucear Pree themostvabis wubettute from foasl The Contender Vote Here Pee Tied tem forest Seamed Winning Losing a ar (Ail ap | arene teen meer pon some peed eu ts Pn a te | Fig. 41. Debate.org is a social networking site promoting debate. A.21, Debategraph Debategraph!”’ [106] is a wiki debate visualization tool which has been adopted for use at the Kyoto cli- mate change summit and is being tested by EU projects such as WAVE)’. Debategraph offers several visual- izations, including the Debate Explorer view shown in Figure 42(a) on page xi and a text-based outline shown in Figure 42(b) on page xi. Visualizations can be em- bedded in other websites, and Debategraph encourages users to add links to related webpages within graphs. A.22. Debatepedia Debatepedia! bills itself as the “the Wikipedia of pros and cons". Sponsored by the International Debate Education Association, Debatepedia is a collaborative community effort to summarize arguments. Each ar- gument page provides an overview, then a list of is- sues, with pros and cons supported by news articles and similar sources. It provides an intuitive editing en- vironment, where users can edit just the relevant sec- tion, such as the pro or con for a topic. Debatepedia is unique for providing an easily-editable wiki of pros and cons. http: //debategraph.org/ http: //www.wave-project.eu/ Snttp://debatepedia.idebate.org/ A.23. Debatewise On Debatewise!"®, everyone can collaborate in cre- ating the strongest case both for and against a given issue. As part of a partnership with the International Debate Education Association (iDebate), they provide links to Debatepedia and iDebate’s reference site De- batabase!''. Karma, teams, and lists of recent partici- pants and new editors help motivate participation. USE GEOENCINEERING TO SHIELD THE EARTH FROM SUNLIGHT, Uiee gouengincering to whielif Phe eueth fran sunlight. Yeu, Become No, because Mipht os self» orking Ina grouncworn stowed ba aang ad tr geoargeerng Fig. 43. Debatewise offers an executive summary, followed by a de- tailed pro/con debate. There are several unique features. The site makes it easy to get involved by providing suggestions of 5- minute, 20-minute and 1-hour tasks and showing “7 things you should have an opinion on" in rotating im- ages on the homepage. Edit histories are available for each pro and con point. Debates are structured as ad- judicated debates between two teams; other users can make comments, vote, and subscribe to debates. A.24. Discourse DB DiscourseDB!" is used to collaboratively collect policy-related commentary. Opinion pieces (Figure 44(a) on page xi) are collected from notable sources, news- papers and websites with at least 50,000 circula- tion/unique visitors per month. Users categorize these opinion pieces, selecting a quote, indicating the topic NOnttp: //debatewise.org/ http: //www.idebate.org/debatabase/intro. php N2nttp://discoursedb.org/ Schneider et al. / A Review of Argumentation for the Social Semantic Web xi ‘What dows ean want? | an = (a) t om Seer t serexe or piscoraneamaoaroumaments L == ence _. See 4 Ae perengman wars 85 _S, peeee een tence Lami ta gen rt =o ee tte ned O87) ne Ro pepe Pantin a oem ot t ee ee eeeteaeeert Der ge rarer ee oe menses erent at rea apa gcee an ptm | — secre nate apt 9 Gog ee — eae ae ee Rare i faae peer opener _. rea (b) Fig. 42. Debategraph for CNN’s Amanpour TV shown in (a) Debate Explore view; (b) text view. ——— NCTE LT re nae Pent stmt Eat View ary | ea (sean) mee | Don't touch my junk? Grow up, America. = Th ot epi nem, ameris) Ma a egisce Source Toe nseryen Pe Leliaal Date Novenber 242010 nerenron oe Beene tne ore Serer tm saat tt 4 a — one A wr ogee et ha wwe Aor ese data ee Tae es lone (a) Page Ocum Flead anwar form East View ristory = Con} iteme New START ‘Tha a opie, wth slat date ApeT BL 2010 re urhrowm ert ie the Peres} gee lens date Poations ata shown ruvoner-cteonclegicaly. ty Per eartest sini tar Newgate : ; ieee Summary of positions for this topic amt = Treay outers (tema ve vasa aay Full listing om, (es ton nae vie gto om a ie en lecuuin Treaty should be ratified einen tarveslender For ‘Tha Mlepwtlican sen tr rating Mew START by Herey Knsirger, Corerps Eva, Jeena Baer, Lame a for nay (cot city START by fan Franckco Ceenicle ediovl board (Sin Francice Cronies, owner 1, 2080) coreete ae ‘Add an epmon tee Se Eee ener Againet | Now STAT, Ol Whine by Investors Business Daly estorial Soar (investors Business Day, November 24, Saracens 2010) Wve As ape ‘A Poor START by Rich Lowry (Aasonal Review, Novericor 25, 2010) (view o) Hy boven Reems Daly moral tare [inendor's Mera Daly, ‘Tes eed Onarra’strnw-warn focus on te hiew START vaty ty Gorge F.Wil(The Wiaatingean Post Oecember 2 Wai ee Fees ‘B00 (view Peat -ma-pay wot Seeew pape All terns that address: this: topic (8 total) Prreatin verses fianaa a fs Tite FS Auther Hy Seome i Bate tema cepetes ‘Atom START Tasty? Ne Das Inverters fusinaes Daay ecaeral invericry unas 17 Rioweroar beara Day amo an Francteco Chvontle ogterial Gan Franciece (b) Dont delay STAAT November Fig. 44. In DiscourseDB, (a) users catalog opinion pieces; (b) this generates an overview of the positions for, against, and mixed on a topic. and position, along with whether the author’s argument is for, against, or mixed on the position. DiscourseDB uses Semantic MediaWiki [95] with the SemanticForms!!? extension. This makes it possi- ble to list all commentary written by particular person, published in a particular venue, and so forth. Further, since items indicate the position they take on a topic, DiscourseDB can list all commentary for or against a given position as shown in Figure 44(b). Bhttp://www.mediawiki.org/wiki/Extension: Semantic_Forms When a topic has multiple positions (e.g. Darfur!'*), DiscourseDB is especially helpful in summarizing the discussion. A.25. Dispute Finder Dispute Finder!!5 [53,54] is a browser extension that alerts users when information they read is dis- puted, based on a database of disputed claims. This M4hetp://discoursedb.org/wiki/Darfur_ conflict 'Shttp://ennals.org/rob/disputefinder.html xii Schneider et al. / A Review of Argumentation for the Social Semantic Web database was created by asking activists (who are in- terested in informing or convincing others) to indi- cate disputed claims manually, and then extended algo- rithmically. While the Dispute Finder plugin remains available!'®, it notes that the project has ended; unfor- tunately, the plugin no longer highlights phrases such as the “abortion reduces crime" phrase used in paper examples. A.26. Hypernews Hypernews!’ [21] is a general purpose Web forum, inspired by Usenet news. Its use of message types dis- tinguishes HyperNews from other forums. Users are asked to indicate what kind of message they are post- ing (None, Question, Note, Warning, Feedback, Idea, More, News, Ok, Sad, Angry, Agree, Disagree) as shown in Figure 45(a) on page xiii; the message type is then displayed as an icon in the forum’s thread view (Figure 45(b) on page xiii). A.27. Living Vote At Living Vote!!®, the general public can discuss pro and con arguments of issues, creating argument maps, as shown in Figure 46 on page xiii. A tree view provides a coherent view of the argument, which can be drilled down, where arguments and their counter- arguments are presented side-by-side. Users can add arguments, and voting colors the nodes according to whether you agree (green), disagree (red), or haven’t voted (white). Living Vote is unique in the way that it handles and uses votes. To vote, users must answer questions de- signed to test whether they’ve read the arguments. Liv- ing Vote also prunes unhelpful arguments and aims to provide a “complete, persistent, constantly changing and up-to-date record" of everyone’s opinions and the most convincing arguments. A.28. Opinion Space Opinion Space is software developed by UC Berke- ley’s Center for New Media “designed to collect and visualize user opinions" on a variety of topics [55]. The "6http://addons.mozilla.org/en-US/firefox/ addon/11712/ http://www. hypernews.org/HyperNews/get/ hypernews/reading.html '8http://www.LivingVote.org/ U.S. Department of State is using Opinion Space!’ to aggregate opinions about foreign policy and create a “virtual town hall" as shown in Figure 47. Fig. 47. Opinion Space maps comments in a constellation view. Opinion Space is unique in its use of deliberative polling and visualization. With deliberative polling, participants are polled both before and after deliber- ation, to better understand how public opinion can change based on increased understanding of the issues. Users move sliders to express their opinions on five is- sues. The system then maps the user’s opinion, using principal component analysis, to show the user where they stand. Each point in the visualization represents a perspective; larger points represent more popular per- spectives. Users can also view and rate others’ com- ments (Figure 48). Ratings can be used to choose the most informative comments for display. Ue esl aul ieicM (1s) Sam lm tee ieee ery J ene eg eRe eae] Oar Fig. 48. Opinion Space uses sliders to collect and display users’ opinions on five issues. http: //www.state.gov/opinionspace/ Schneider et al. / A Review of Argumentation for the Social Semantic Web xiii Kind of message: 19 @) No "Next message" in response header by Nils Davis, 1995, Aug 30 1 @ Lnoticed that too by ben@wiliki.eng hawaii.edu, 1995, Oct 06 (If this node is a Message.) © th None O P Idea OQ Question © & More O @ Angry O 2 Note O # News O g} Agree : 2 @Ok © & Disagree 381 Next and Previous will be rejoining us soon by diberte@hypernews.org, 1995, Oct 09 20 P It would be nice to have "TOP" option , 1995, Nov 17 1 4) Lost without a "Top" option by jaf@iyrell.net, 1996, Jun 15 1 & Implemented a HOME/Top option by haroon@wwwnoet.attmail.com, 1996, Jul 10 2 & A better implementation for Top/Home by haroon@wwwanoet attmail.com, 1996, Jul 10 21 Reversing Threads? by Randy Cosby, 1995, Dec 30 1 & Other solutions possible too by liberte@hypernews.org, 1995, Dec 30 1 =4 Another alternative... by jap@te.cornell.edu, 1996, Jan 02 (a) (b) Fig. 45. (a) Users are asked to specify their message type, using this Hypernews taxonomy; (b) Part of a Hypernews discussion thread. How It works Username or Email; California Prop 8, which states “anly Livingvote,org; Same Sex Marriage (CA Prop 8) (1212 vs 647 Password: Login beats Re Share/Bookmark/Linl# The Whole Tree California Prop &, which states “only marriage between a man and a woman is valid or recognized in California” should stand. Disagree Total vate Add an Argument Arguments Against Gay couples deserve the dignity of marriage as well, 74% 65% Agree Add an Ar Arguments in Favor Prop 8 protects the free expression of retigion 65% ara ran end Children are best served when reared in a home with a married mother and father. 70% Lngend: (C) You agree Ow Disagree You Haven't Veted Fig. 46. At Living Vote, the weight given to a user’s votes increases as they read and vote on more arguments. A.29. Online Visualisation of Arguments (OVA) Online Visualisation of Arguments!?° (OVA) is an online argument analysis and mapping environment [169] which exports AIF. In OVA, web pages can be displayed adjacent to an argument mapping canvas, helping analysts create a graphical representation of the arguments in online forums or news stories. The re- sulting argument maps can show the relationships be- tween premises (supporting or attacking) as well as the participants responsible for each point of view. In ad- dition to AIF, users can export JPEG and SVG images of the argument. OVA is part of a pipeline of argumentation tools [168] which starts to bridge the gap between human- Mnttp://ova.computing.dundee.ac.uk oriented argumentation tools and calculation-based agent argumentation. Mixed initiative discussions are enabled by the argument maps created by OVA or any other AIF-based tool. Thus, instead of represent- ing one’s point of view countless times in a forum or FAQ, it would be possible to delegate these conversa- tions to a machine agent using an underlying argument map, as prototypes like MAgtALO'! [145,202] and the Google Wave discussion bot Arvina [169] show. A.30. Parmenides Parmenides!”? [32,33,34] is a structured survey tool for gathering public opinion on a proposal. Based on inttp://www.arg.dundee.ac.uk/?page_id=61 http: //cegi.csc.liv.ac.uk/~parmenides/ xiv Schneider et al. / A Review of Argumentation for the Social Semantic Web argument schemes and critical questions from argu- mentation theory, Parmenides can pinpoint the source of the disagreement, by having participants respond to a series of questions. In a Parmenides debates, partic- ipants are first asked to agree or disagree with a posi- tion on a question such as “Should laptops be banned in lecture theatres?" (Figure 49(a) on page xv). Those who disagree are stepped through several screens (such as Figure 49(b) on page xv) of yes/no questions to de- termine the source of the disagreement. Limited free text boxes allow users to add further information. At the end of the survey, users are offered the choice of submitting an alternative proposal, and are shown the answers they chose. Administrators can then analyze the group’s responses, which are displayed in graphi- cal argumentation frameworks [51]. A greater under- standing of the most popular reasons for disagreement could support further discussion and debate about the key issues. A.31. PDOnline SWAN/SIOC is itself used in PDOnline!”?, an on- line community for scientists, funders, and medical professionals working in Parkinson’s disease science, which is funded by the Michael J. Fox Foundation [48]. Figure 50 shows a PDOnline discussion about a recently-published paper and indicates how the topic fits into the “PD Guide" taxonomy of research and communication topics. The discussion links both for- ward to responses and related contributions and back to a thread on Papers of the Week (itself contained within a Research Question board). Members’ full names, credentials, and institutional affiliations are listed, with links to user profiles and institutions. Mem- bers’ profiles link to their publications, and throughout the site explicit references to the literature are given. It is unique in that it uses scientific argumentation. A.32. REASON REASON —Rapid Evidence Aggregation Support- ing Optimal Negotiation [80,81] — is a Java applet for group deliberation, used to arrive at a consensus de- cision. Drawing from decision theory, group-decision support systems, and argumentation, REASON is in- tended to improve information pooling. An argument map is used to organize group evidence shared dur- Welcome Guest | Be & petof PO Onine Research pd ONLINE RESEARCH — — Log in | recov password ee oe ee to ea ee Re ee ad Is GIGYF2 a PD gene or not? By Bran K. she, PD, Po Guide: iit Response te Research Question: POWs: Papers of the Weck View all 17 resporses to this Research Question DIRECTLY RESPONDING TO: POWs: February 23-29, 2010 OPSSPNEST Other opinions on the GIGYF2 [ink to: vn, Ls oe By: Be eke, ME ‘The report from Wang et al identitying variants of GIGYF2 (PARKI4) in Chinese PD patients seems to bring back a gene that I had assumed was ne longer considered a genetic factor involved in D (see Bras et al, 2009). Is there any consensus in the field an this gene? If its ota facto, what is the best way to shaw that once and for all? REFERENCE: fang L, Guo IF Zhang WW MU Q, Za0 x, SCH, et al. Revel GIGYF2 gana variants patents wh Parkinsie’s Gsase n Chinese population. Neuroscience tars. 2010. Ernote XML BIOTEX Stncher J, Feder , Margadinhe JanuarieG, Ribaka Mt al Lack of relation yen CIGYF2 varlans end Parkin disease: Hum Mol Ganet. 2009; 18(2):3415, RESPONSES No By: Mark Cookson, NHL {Mar 2049 94:28 8M EST Fig. 50. Part of an argumentative discussion at PDOnline ing the decision-making process; further, in an adap- tive version of REASON, aggregate weights express- ing the group’s view of each alternative are displayed. Uniquely, arguments start as threaded discussions in REASON, and are colored based on whether they agree (blue) or disagree (yellow) with their parent in the thread. A.33. Riled Up! Riled Up!!* (Figure 51(a) on page xv) is a debate- centered site which motivates participation with a point-based authority system. Aimed at people who enjoy debate, Riled Up!’s tagline is “Like Raising Cain? So Do We." Users can add debates, arguments, and comments, and vote for others’ arguments, as well as add friends and enemies. Riled Up! is unique in its comment system—users can respond with positive (green), neutral (grey), or negative (red) comments, as shown in Figure 51(b) on page xv. In addition to a standard layout, a contributor view gives an overview of the arguments but not the comments. IBhttp://www.pdonlineresearch.org/ P4énttp://riledup.com/ Schneider et al. / A Review of Argumentation for the Social Semantic Web mA efor a a a a a Wa betere fat Blan lagtope in tecture theatres in ig bcos nine curent amaaton eter eae, Univer a techie Peres ‘sucercr unre ices Guects nce, Unreg lpsios rey acre inh Daretoa tr among. Ung apnea ny hcnmes Gare Oter manoat of he aumence (Our poate are Make Pe rune consistent Alow fo lecirw in concanéatr on piving fee amseniason ncreaee auéence concentrator ne fechum, Reduce diaracton of cher materce mame This wil achiene Fecire Garmin wanes grace menee's nrveses Cee’ pecpie Jearang nemae msence conceresean on cm jrumeaes Feraina seareng [Anos Pa cue’ nce on veg he Dasenatin pomcas Ratgett ake he Wen cnie' omOE COME ery ‘lo you agree with the minal of our position? (a) DU egy eee Red aera ensen Sne XV The use of laptops in lecture theatres Dis pou Beek Bea tslowrng enc valves ww wanth pring? Ik (tw becle soem ® Peas arg . mee . onaitcy ® ‘Are Bee ary oft valoen.sctemartionsclin fa ti above, hich ys fink aru wer romana? es. smne ree nen (b) Fig. 49. (a) In Parmenides, participants are asked to agree or disagree with a starting position. (b) Next Parmenides steps participants through a series of yes/no questions to pinpoint the source of their disagreement. Lek be People are aimays preaching “eat heattiy ‘expenaive than eating unhealthy. Fresh vou get the wheat kind ee Is it posible to eat healthy AND cheap? don't realize is that esting healthy can often be mare itand vegetables are expensivel Pasta ia wcess is bad fi for you unless iS | ce } Eating healthy can be cheap, Eating unhealthy food i: B Lo] you just need to plan ahead. Its unfortunatly cheaper tha a . more expensive to take care of healthy food. Only the rich ca y v a hear eat healthy. ie: Lo] ES Ls aa a =n : t a eens Sensi w a mB fava De di (a) (b) Fig. 51. RiledUp (a) debates allow structured discussion on a topic; and (b) readers can respond with positive (green), neutral (grey), or negative (red) comments. A.34. SEAS SRI International’s SEAS! [105,104] is a template- based structured argumentation tool originally de- signed for collaborative intelligence analysis. It has since been tested in other domains such as by IRS tax auditors and in a simulated public health emer- gency. SEAS’s most unique feature is its emphasis on templating; users can author templates which provide transferrable notions of how to make an argument, and specify authorized coeditors. Figure 52 on page xvi shows one question from such a template. These tem- plates, which are in essence domain-specific argument MShttp://www.ai.sri.com/~seas/ schemes, allow non-experts to make sound reasoning. SEAS automatically answers some questions based on earlier responses. The developers report that a threat- assessment template originally developed by U.S. in- telligence analysts was successfully applied by non- experts in their laboratory. SEAS visualization fea- tures are also considerable: to visualize multiple di- mensions, SEAS uses starburst, constellation, and ta- ble views. SRI International runs a SEAS server with paid accounts and SEAS server software is available. © ° o 4 4 1.3.1 UNUSUAL TRANSACTION: Is there a large, unusual, or questionable Jransaction? Schneider et al. / A Review of Argumentation for the Social Semantic Web added rich relationships suchas is elaborated by, is supported by, is summarized by, and stands though contradicted by, which the system stored in XML, RDF, and DAML+OIL. In contrast to the detailed argumentation of Rich Trellis, Tree Trellis uses only pro and con, and col- John Lowrance 15 Jan 2007 16:45:30 @ Yes, almost certainly Likely, more likely than not Even, about as likely as nat Uniiicely, more unlikely than not No, almost cartainiy net suspicion. The bank transaction is sufficiently large to warrant strong laborative discussion is supported, while Table Trellis allows feature and value pairs to be arranged in a ma- trix, allowing the user to compare and evaluate alter- natives according to their own criteria. |A.36. TruthMapping e Bank Transaction Record This contains @ lang ction that is cut of Ihe ondinary lor a business of this size. Deleted re: ds on his computer inciuded references to a known problematic company. e TruthMapping!’ is an online deliberation tool which seeks to structure the conversation to focus around the “aha!" moment, avoiding digressions and soapboxes, and making hidden assumptions explicit. TruthMap facilitates structured conversations which use argu- ment maps, critiques and rebuttals (Figure 53(a) on Fig. 52. SEAS uses a series of questions to structure the argument [105]. A.35. Trellis software The argument analysis system Trellis!26 [61,40,41] was built on Semantic Web technologies, including the Semantic Annotation Vocabulary Section 7.7, page 20. Trellis, inspired by intelligence analysis, began as a credibility and analysis system to help structure de- cisions, for example to construct a family geneology based on contradictory information [61]. Originally, Trellis was designed to help capture ar- gumentation, grounded in documents, whose reliabil- ity the user rated, and from which the user extracted statements; although users did not work directly with the underlying ontology, arguments could be exported into XML, RDF, DAML, and OWL. In addition to the original version, now called Rich Trells, two other modes, Tree and Table Trellis, described in [41], are now supported, for incremental formalization. In Rich Trellis, statements are given likelihood- qualifiers such ‘surprise’ (indicating the analyst’s sub- jective reaction); reliability-qualifiers such as ‘com- pletely reliable’; and credibility-qualifiers such as ‘possibly true’. Statements may also be associated with a document providing evidence. The source for each document, including creator, publisher, date, and format, is recorded. Originally, in Rich Trellis, users 26nttp://www.isi.edu/ikcap/trellis/ page xvii). Users can vote on and rate topics, and watch particular conversations Only one user, the original ar- guer, modifies the map; feedback comes in critiques attached to each premise and conclusion (Figure 53(b) on page xvii), which can be rebutted. One unique as- pect of TruthMapping is that users can continually modify each contribution, but can only post one cri- tique on each node. This is designed to make it easier to contribute a persistent comment to the discussion, which can not be drowned out by a single opponent. The system indicates when comments are out of sync, and a wiki-style history is available. Another unique aspect is the use of votes to color the map: as shown in Figure 53(a) on page xvii, each node is colored based on the percentage of votes agreeing (green) and dis- agreeing (red). A.37. Videolyzer Videolyzer!? [50] allows the general public to have sensemaking and argumentative discussions about the quality of online videos. It builds on gamelike-creation of video transcripts and on machine tagging of areas of interest in either the transcript (claim verbs, peo- ple, money, and comparison) or the video itself (faces) (Figure 54(a) on page xvii), to provide an integrated discussion forum for annotating and challenging the claims a video makes (Figure 54(b) on page xvii). Videolyzer is unique in its focus on integrating argu- mentative discussion into a video platform. Qhetp://www.truthmapping.com/ P8http: //videolyzer.com/ Schneider et al. / A Review of Argumentation for the Social Semantic Web xvii Statenents Statement me i atingn 1B) select a wae (a) Hide Statements Critiques: ema Genetic mutations do eccur in biological organisms that possess a genetic code. Further, these mutations can be beneficial, harmful, or irrelevant to the ability of a mutated organism to reproduce and pass on Its genes. ~~~ select ~~ |i CRITIQUE by rkm (cl), Mar 18, 2006, 6:55pm GMT EI xX 4 Mutations of genes, as I understand It, are not a significant factor in evolution. Most mutations are dysfunctional, often leading to the death of the individual. The significant factor in evolution Is the recombination of existing genes. Changing conditions can change the relative survival rate of different characteristics, leading to a gradual shift in the species. 7 | Agree | Disagree -- specific disagreement — generalization misplaced authority red herring straw man begs the question false analogy personal attack [report as inappropriate] (b) 2) FROM 1ITFOLLOWS THAT: @ E14 The genetic mutations of biological organism differ from those that existed in the past. Oy Fig. 53. Truthmapping (a) allows users to construct an argument by laying out premises and conclusions. Each node is colored based on the percentage of agreement (green) and disagreement (red). (b) Each premise and conclusion can be critiqued in comments, and critiques can be responded to with rebuttals. end en (a) Tr Highlight comparisons vj Claim Verbs |_| People |v Money , You save each year for important things like retirement and college. But you probably didn't know that lawsuits are forcing your family to pay thirty five hundred dollars more each year for everyday goods and services. Some trial lawyers are exploiting our courts using frivolous lawsuits to make millions. Every family pays the price. Some would, say it's almost criminal. Lawsuit abuse it's a national problem. See where your state ranks (b) Fig. 54. Videolyzer (a) allows users to comment on the points made in a video; and (b) algorithmically determines segments of possible interest to help focus the discussion: in the transcript these are claim verbs and comparisons as well as mentions of people and money, and in the video these are peoples’ faces. B. Matrix Comparison of Tools We now present comparison charts of the tools re- viewed. Figure 55 on page xxi shows an overall com- parison, in which tools are compared according to var- ious features, which we outline shortly. For the down- loadable tools, Figure 56 on page xxii provides the li- cense, programming language(s) and data storage. In both tables, we use ‘?’ to indicate that we could not locate a piece of information. First, we record the intended purpose of the tool. Next we provide the representation style and func- tional type. As introduced in Section 9.3, page 23, rep- resentation style is drawn from linear, threaded, graph, container, and matrix (including combinations of these styles); functional type is drawn from issue network- ing, funnelling, and reputation. Then we indicate what xviii Schneider et al. / A Review of Argumentation for the Social Semantic Web ArgDF Arguehow Argument Blogging Argumentum Argumentations.com Create argumentation Distill the best points Enable argumentative Gather and use news Purpose schemes on the Web to amine ond responses on the Web pee oe ei stories Representation style Text Linear Argument Map Threaded Threaded Functional type Issue networking Reputation Issue networking Reputation Issue networking Advanced visualization (from AIF) - - - Tag spheres Perspective’ Single Single Single Single Single Distributed architecture: N N Y N N Downloadable or hosted] Hosted, downloadable Hosted Combination Hosted Hosted Registration Site-specific Site-specific No login Soogle, foc) Site-specific Twitter oo . "Addthis" plugin for Blogs and publishing Argumentum Facebook 3rd party services integration a . sharing platforms App Social networking capability a So er a Compare users De en history history Stable URLs ? Y ? Y Y Tags’ N N N Y Bookmarklet N N Y Y N Promote embedding N N N Y N Attach media N N Y N N Input methods Form-based Form-based HTML eae “orm-based, depends on type on type Consistency checking Use AIF tools N N N N Credibility metrics N Y N Y Y Export formats’ AIF None ? None None (a) AGORA: Participate- . a . . Argunet Avicenna bCisiveOnline Belvedere Deliberate Express arguments in Support secondary Purpose Visualize debates Sketch and share argument maps OWL to enable classification Real-time collaboration for decision-making school students learn critical inquiry skills Representation style Argument Map Argument Map Argument Map Argument Map Argument Map* Functional type Issue networking Issue networking Issue networking Issue networking, Issue networking funelling Advanced visualization - Map associated items - - Concept map, matrix Perspective Single Personal Single Single Single Distributed architecture N Y N N N Downloadable or hosted Combination Hosted, downloadable Hosted Hosted Downloadable Registration - Site-specific Site-specific - 3rd party services integration - a plugin ioe - Skype - sharing Social networking capability - SD ae nouser - Lists Skype usernames - profiles Stable URLs - Y ? Y ? Tags N ¥ N N N Bookmarklet N N N N N Promote embedding N Y N Y N Attach media N N N Y Y Input methods’ Visual controls Visual controls Visual controls Visual controls Visual controls Consistency checking N N Use AIF tools N N Credibility metrics N N N N N Export formats None Locally stored AIF PowerPoint ? (b) Schneider et al. / A Review of Argumentation for the Social Semantic Web Cabanac's Annotation System Competing Climate CoLab Hypotheses Cohere Consider It ‘ConviceMe Sensemaking of Understand the Collective intelligence © Connect and share Analysis and cross- Purpose arguments in . . . pros/cons behind your _ Have fun debating . on climate change ideas checking an annotations opponents’ opinions Representation style Threaded Threaded Argument Map Matrix Container Container Functional type Funelling Reputation Issue networking Funelling Issue networking Reputation Sort by the most likely Show points accordin Advanced visualization - - Map, timeline y uv E e hypotheses to who holds them Perspective’ Personal Single Personal Personal Single Single Distributed architecture N N N Y N N Downloadable or hosted Hosted Hosted Hosted, downloadable Hosted, downloadable Hosted, downloadable Hosted Registration Site-specific Facebook Site-specific Site-specific Facebook Facebook RSS, Send te Find Facebook, RSS, Send t 3rd party services integration - Facebook like en reas ° Find us on Twitter ind us on aceboo en een ° Twitter Twitter Facebook, Twitter . . as . . . Persistent chat, Attribution, but no Forum discussions, Social networking capability - Discussion, profiles Groups, profiles. . 7 message board user profiles user profiles Stable URLs N Y Y Y Y ¥ Tags N N Y N N Y Bookmarklet N N Y N N N Promote embedding N N Y N N Y Not in argument ma Attach media’ N . E . D URLs only N N N discussions . Form-based, some . Input methods Visual controls - Form-based Form-based Visual controls Form-based HTML and Wiki Consistency checking Y N N Y N N Credibility metrics Y Y N N N ¥ Export formats ? None None None None None (c) CoPE_IT CreateDebate Debate.org Debategraph Debatepedia Debatewise Help groups e-Learning, K-12 education, Meet people through DET Clarify public debate, | Help make informed Purpose’ collaborate on Collaboration Debating debate . engage citizens decisions complex issues Argument Map, . . Argument Map, + , Representation style 8 P Container Linear 8 P Container Container Threaded Threaded Functional type: Funelling Reputation Reputation Issue networking Issue networking Reputation . is Automatically Summary graphs and Automatically change Advanced visualization . . a - . - - formalize the view statistics the view Perspective’ Personal Single Single Single Single Single Distributed architecture N N N N N N Downloadable or hosted] Hosted, downloadable Hosted Hosted Hosted Hosted Hosted Registration OpenID Site-specific Site-specific Site-specific Site-specific OpenID Import phpNuke API, pull in external oo. / port pip! uP Facebook like, Post to . RSS, poston Facebook, "Addthis" plugin for 3rd party services integration webforums, RSS feeds, Facebook RSS, email Twitter, Send email Twitter, Delicious, Digg sharing, Facebook like Compendium maps fanpage Social networking capability Groups oS ene) ete) Extensive user profiles User profiles ea Cea od - enemies history Stable URLs N Y ¥ Y Y Y Tags Y Y - N N N Bookmarklet N Y N N N N Promote embedding N Y N Y N N Attach media u See embedded nT ae embedded 7 A videos only videos only Input methods]Form-based, WYSIWYG Form-based, WYSIWYG Text with HTML Form-based Wiki formatting Form-based, Wiki-style Consistency checking N N N N N N Credibility metrics N Y ¥ N N Y Export formats’ None None None None None None oy) XX Schneider et al. / A Review of Argumentation for the Social Semantic Web DiscourseDB Collect opinions of Dispute Finder Discover what's widely HyperNews LivingVoice Opinion Space Argumentative Web "An up-to-date record Online Visualization or Arguments Analyze and diagram Purpose] commentators and disputed when forum with an email of what we believe and Exchange perspectives journalists browsing the Web gateway why we believe it" arguments Representation style Container - Threaded Argument Map - Argument Map Functional type Reputation Reputation SS Orne) Issue networking Reputation Issue networking discussion Advanced visualization - De eee - - Cees pineal . (from AIF) sentences component analysis Perspective’ Single Personal Single Single Personal Personal Distributed architecture N N Y N N N Downloadable or hosted Hosted Combination Hosted, downloadable Hosted Hosted Hosted Registration Site-specific No login Site-specific Site-specific Site-specific No login 3rd party services integration - Facebook - AddtoAny plugin - - Social networking capability nae ood Profiles Discussion-based Ne onsdnae Rate comments - history opinions. Stable URLs Y N Y Y N Y Tags| N N N N N N Bookmarklet N N N N N Y Promote embedding N N N N N N Attach media’ Y N N N N N Roll-over text in user Input methods Wiki formatting mode, Form-based for Form-based Form-based Form-based Visual controls activist Consistency checking N N N N N Use AIF tools Credibility metrics N Y N Y ¥ N Export formats’ RDF None None None ArgDB, AIF (e) Parmenides PDOnline REASON Riled Up! SEAS Trellis Get feedback on oe . . Debate & discuss; Purpose proposals for e- ppeed aan SE Dieematon show that you're an Intelligence analysis Analysis oo communication pooling . . participation authority on a topic Representation style’ Argument Map Threaded epee Container Multiple Linear Threaded Functional type Funelling Reputation Funelling Reputation Funelling Funelling . oo. Analysis toolset uses Starburst, Advanced visualization - - - . - Value-based Constellation Perspective’ Personal Single Personal Single Personal Single Distributed architecture N N Y N ¥ N Downloadable or hosted Hosted Hosted Hosted Hosted Hosted, downloadable Downloadable Registration No login Site-specific Site-specific Site-specific Site-specific Site-specific —_ / Taktolrscchosrpere! Digg, Delicious, Reddit, 3rd party services integration . - Yahoo!, Google, - - Twitter Stumble Social networking capability Profiles, comments Discussion Profiles, comments - Stable URLs Y Y N Y ? ? Tags) N N N ¥ N N Bookmarklet N N N N N N Promote embedding N N N N N N Attach media’ N Y N N N URLs only Input methods Form-based Form-based Visual controls Form-based Form-based Form-based Consistency checking Y, in admin view N N N ? N Credibility metrics N Y N Y ¥ Y Argument Markup Export formats’ ? None None None None Language, HTML, Word Schneider et al. / A Review of Argumentation for the Social Semantic Web xxi TruthMapping Videolyzer Purpose Overcome "loudest Collaboratively voice" and "last word" evaluate online videos Argument Map, Representation style Threaded Threaded Functional type Funelling Issue networking . oo Shows % agreement Advanced visualization . - and disagreement Perspective Single Single Distributed architecture N N Downloadable or hosted Hosted Hosted Registration Site-specific Site-specific 3rd party services integration = = Social networking capability Discussion Discussion Stable URLs] Y Y Tags| N ¥ Bookmarklet N N Promote embedding N N Attach media N URLs only Input methods] Form-based Form-based Consistency checking N N Credibility metrics Y Y Export formats] None None (g) Fig. 55. Overall Comparison of Tools. sort of advanced visualization is offered; ‘-’ indicates that no examples were found (i.e. that the question does not apply). The perspective row records whether an individual user has a personal perspective distinct from the group view. Next we consider whether a tool has a distributed architecture (allowing multiple copies to synch with one another). Then we distinguish downloadable and hosted sys- tems (noting that some tools are in both categories or use a combined method). To understand their current integration with the Social Web, we record whether they use a site-specific login, or allow external creden- tials (such as OpenID, Twitter, or Facebook). We further indicate whether they have any integra- tion with third party services; a single row does not do justice to the wide range of integration we found. For tools with social networking capabilities, we pro- vide an example of the interaction users can have with each other, or the information they can find out about each other. Stable URLs indicates our success in find- ing reusable bookmarks: in fact these URLs can be at multiple granularities, such as the entire argument map, issue, or conversation; each individual comment or critique; etc. We also indicate, in the tags row, whether users can provide tags for content. We also indicate which tools have a bookmarklet for use while browsing, and which promote embedding on external sites. The re- maining rows describe features related to the site’s in- teraction style, starting with whether it is possible to attach media in discussions and the input type (such as point and click visual controls or form-based editing). We also indicate which have consistency checking (i.e. avoiding obvious contradictions) and credibility met- rics (usually, but not always, voting) as well as export capabilities. Tools which export AIF can take advan- tage of an existing infrastructure. Overall, we can make certain observations regarding these tools: generally they focus either on encourag- ing discussion or having a basis in rigorous argumen- tation models. Significant amounts of innovation has occurred in the research community, but many ideas have not been propagated to the Social Web at large. There are certain common mechanisms among many systems—basic features such as upvoting, segregating pro and cons, etc. Social Web systems do not have even levels of adoption: some tools are very well-adopted while others are not. xxii Schneider et al. / A Review of Argumentation for the Social Semantic Web License Language Data storage ArgDF ? PHP Sesame . Python, Django, Argument Bloggin ? AIFDB 6 6emne Javascript, Jquery Argunet Open source Java Db40 Avicenna] Author copyright Java ae ARQ, SQL Server DB bCisiveOnline Commercial Python, Django ? Cohere LGPL PHP MySQL Competing Hypotheses GPL v3 PHPS MySQL Considerit AGPL v3 Ruby on Rails ? CoPe_IT Various C#, .NET MSS SQL Server 2005 Dispute Finder Apache Ey, ruby Scala, MS SQL Server HyperNews MIT Perl Document directory Commercial, Free to , ? ? SEAS US Gov, ? ? Trellis GPL Perl ? Fig. 56. Downloadable tools: License, language, and data storage.