=Paper=
{{Paper
|id=Vol-3090/paper01
|storemode=property
|title=Software for Modeling Deliberative Argumentation: Requirements and Criteria
|pdfUrl=https://ceur-ws.org/Vol-3090/paper01.pdf
|volume=Vol-3090
|authors=Elena Lisanyuk,Dmitry Prokudin
|dblpUrl=https://dblp.org/rec/conf/ims2/LisanyukP21
}}
==Software for Modeling Deliberative Argumentation: Requirements and Criteria==
https://ceur-ws.org/Vol-3090/paper01.pdf
Software for Modeling Deliberative Argumentation:
Requirements and Criteria
Elena N. Lisanyuk a,b, Dmitry E. Prokudin a,c
a
St Petersburg State University, 9 Universitetskaya Emb., Saint-Petersburg, 195299, Russia
b
University Higher School of Economics, Address, City, Index, Russia
c
ITMO University, 49 Kronverksky Ave., St. Petersburg, 197101, Russia
Abstract
Methods of deliberative argumentation are widely employed for solving applied tasks in
various fields of practical activities, where choosing of a line of behavior in a certain situation
or making decisions is at stake. These methods enjoy permanent attention in the contemporary
education with respect to teaching argumentation and training the critical thinking skills. In the
last three decades, the progress in the information and communication technologies has led to
the development of software designed for visualization and modeling of deliberative
intellectual activity for solving various kinds of practical tasks and for supporting the relevant
education. We propose the five (groups of) criteria for developing the software designed to
model and represent deliberative argumentation, which have to be observed both in the
development software and in its classification. We suggest four ontologies for such software,
which will enhance implementing functions for evaluating arguments and finding solutions in
such software.
Keywords1
deliberative reasoning, conceptual bases, software, modeling, representation
1. Introduction
In contemporary society, deliberative argumentation is widely used in various areas of human
activity, where the results are achieved in the process or with the help of substantiating actions and
justifying decisions. Such areas include law and jurisprudence, politics, public administration, social
interaction, science, etc. The deliberative, or practical, argumentation, is distinct from the theoretical,
or discursive, argumentation. The former focuses on justifying claims about the line of behavior in
different circumstances – how to act in certain situation or what should we do with respect to certain
goals and intentions. The latter pursues the justification of claims’ truthfulness, and the discursive
arguments are put forward to support or criticize the claims. The discursive arguments as well as the
claims themselves are descriptive propositions which can be true or false. The deliberative arguments
consist of descriptive and non-descriptive sentences expressing norms, values or intentions playing key
role in justifying or refuting their conclusions expressing intentions to act [1]. These formal and
semantic differences of discursive and deliberative arguments is connected to the properties of
intellectual agents participating in the argumentation of those two kinds and entail differences in how
the arguments are evaluated. On one hand, the deductive arguments, mostly regarded the strongest in
the discursive argumentation, are seldom applicable in the deliberative argumentation. On the other
hand, the non-deductive plausible arguments, the most persuasive in the deliberative argumentation,
which include such widely used schemes of reasoning as appeals to expert opinion, to consequences,
negative or positive, to popular opinion or behavior, etc., are often considered fallacious in the
discursive argumentation.
IMS 2021 - International Conference "Internet and Modern Society", June 24-26, 2021, St. Petersburg, Russia
EMAIL: e.lisanuk@spbu.ru (A. 1); d.prokudin@spbu.ru (A. 2)
ORCID: 0000-0003-0135-4583 (A. 1); 0000-0002-9464-8371 (A. 2)
© 2021 Copyright for this paper by its authors.
Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR Workshop Proceedings (CEUR-WS.org)
12 PART 1: Information Systems for Science and Education
The structure of intellectual agents in the discursive argumentation presupposes exclusively
descriptive elements, such as knowledge and opinions (beliefs), and in the deliberative argumentation
it also includes their opinions about norms, values, as well as desires, goals and intentions. Accordingly,
there are special requirements for the agentive properties that are imposed on the agents with respect to
evaluation of deliberative arguments, for example, whether an agent is a reliable source of information
when the argument at question appeals to his or her authority, whether he or she is an expert in the issue
under consideration when it appeals to his or her expert opinion, whether he or she is it trustworthy in
assessing the consequences, etc. Deliberations often involve many people, and therefore it is necessary
to take into account many individual and group parameters in justifying decisions by means of the
arguments [2].
To enhance intellectual activity, many approaches based on the implementation of elements of
argumentation and deliberation in software have been proposed, developed and applied. They aim at
supporting the deliberation in decision-making in various areas of human activity, for example,
medicine [3], public policy and e-democracy [4, 5], law [6, 7], scientific argumentation [8, 9, 10],
business and other areas.
Our present study is one of the stages of a comprehensive research project conceived to assess the
adequacy of modeling of argumentation by means of appropriate software and information systems.
The project aims to bridge the theoretical gap between the concepts of argumentation, implemented in
the software, and the concepts of argumentation, yielded by academic studies of argumentation. At the
previous stages of our research project, we 1) studied the capabilities of the software for modeling
argumentation [11], 2) identified the key characteristics of the software designed for modeling
argumentation, deliberative reasoning and mind mapping [12], 3) formulated the conceptual
foundations, or criteria, for assesing the software, by which we divided it into two groups - on the basis
of its descriptiveness / normativity and on the modifiability of reasoning [12, 13].
As part of our previous research, we have selected and assessed the software and information
systems aimed at supporting the representation of reasoning and critical thinking. The development of
such systems and their applications started in the mid-90s of the XX century; and their active
development and updating continues up to this day with the top intensity of the development in the first
decade of the XXI century.
A characteristic feature of the development of the software is that the ideas of its development are
born inside interdisciplinary academic communities, whereas the conceptual projects for its creation are
realized mainly by the representatives of the logical community including logicians and specialists in
logic programming and artificial intelligence. Here is a list of the most widely used software products
for modelling argumentation and reasoning:
OVA – developed by the Centre for Argument Technology of Dundee University (Scotland),
incorporates D. Walton’s ideas of ‘new dialectic’;
Carneades – developed by T. Gordon (Potsdam University) and D. Walton;
Rationale – initially developed by T. van Gelder’s team in Melbourne University; today is a
commercial software https://www.rationaleonline.com/;
bCisive – elaboration of Rationale for representation of argumentative support of decision-
making (https://www.bcisiveonline.com);
Belvedere – initially developed by A. Lesgold and D. Suthers team in the University of
Pittsburgh, later elaborated by D. Suthers’s team in Hawaii University.
The existing software is used mostly in teaching critical thinking and argumentation skills, for
example, Belvedere [10], LARGO [7], ARGUNAUT. Some systems are initially designed to teach
critical thinking and argumentation skills in jurisprudence - Carneades, ArguMed, LARGO, QuestMap,
others - in research or all-purpose argumentation in general, for example, Belvedere [14], SenseMaker,
Convince Me [15]. Some software products have been developed to implement the IBIS (Issue-Based
Information System) methodology [16] for joint planning and design in various subject areas. The
earliest implementation of this methodology is gIBIS [17], followed by QuestMap and Compendium
[18]. Some software products are used independently of specific subject areas for training general skills
related to critical thinking and practical argumentation, for example, Rationale and bCisive [19],
Hermes [20].
IMS-2021. International Conference “Internet and Modern Society” 13
Carneades, OVA and some other software abstract from the distinction between defeasible
argumentation, which is based on plausible arguments mostly used in deliberative argumentation, and
indefeasible argumentation, which includes deductive and inductive arguments [21, 22]. The
abstraction allows modelling both the discursive and the deliberative argumentation, but at the cost of
a vague mechanisms of its assessment.
With respect to its practical purpose and regardless of its subject area, the software can be divided
into the following groups:
for modeling of argumentation;
for visualization of the discursive and deliberative reasoning;
for mind mapping.
This division is arbitrary as some software systems fall into more than one group. Nevertheless, its
criteria put as the groups’ titles provides us with a preliminary clue for sorting the software.
The diversity of available software is rooted in the manifold approaches to its creation. However,
most of the software systems have some common characteristic features which have been observed in
recent review papers appeared as a output of its comprehensive comparative studies. One of such studies
is the LASAD project carried in 2008-2013 [23], in the framework of which its team examined 45
systems available to the time and designed for supporting the representation of argumentation and
critical thinking. The project team compared the software in relation to the goal of using these systems
for teaching reasoning and critical thinking skills and identified the key functional characteristics
implemented in them.
2. Implements of elements and functions of the deliberative argumentation
in the software
We limit our study of the software to the products designed for modeling argumentative dialogues
(disputes) and represents the argumentation in the form of graphs and protocols. The software designed
to visualize argumentative dialogues offer no tools for scoring assessments of arguments and
establishing solutions to disputes, which means that with respect to the analysis of argumentation, it has
descriptive character even in those cases where it implements the concepts regarded normative by their
developers, as in cases of Rationale and bCisive which are said to imply the code of critical discussion
in pragma dialectics [19], or OVA and Carneades, which involve evaluation of arguments by means of
the critical questions [24]. The developers of the software do not explicitly suggest using it for
intellectual support of deliberative reasoning, but it is applicable for visualizing some aspects of public
deliberations.
Deliberative public opinion plays an essential role in political decision-making and formulating of
the political and social agendas in the deliberative democracy with its evolving contemporary feature
of disagreement and polarization about many issues. Special software systems and platforms are
developed (DemocracyOS, Democracy 2.1, Loomio, OpaVote, Delib, Decidim and others) for
supporting of the deliberative democracy. Most of them are social platforms for polls, exchange of
views, debates and discussions, they aim at supporting decision-making in state and municipal
management, which remain human-oriented. These systems implement technologies for collecting and
processing Big Data by statistical methods and imply no function of solving the discussed problems.
There are several levels of implementation of deliberation elements in the software:
multi-user synchronous (on-line) and asynchronous (off-line) mode for collective
argumentation mapping in teaching argumentation skills - Belvedere, OVA, Hermes;
dialogue modes through feedback toolkits for controlling students’ activities and progress
(Digalo, ARGUNAUT) or for playing dialogues in teaching critical thinking skills (AcademicTalk,
InterLoc), which can be used for group deliberations, too;
web-oriented systems for wide disputes, which allow an unlimited number of participants to
interact in the debates (DebateGraph (http://www.debategraph.org) or Collaboratorium [5]);
constructing arguments, in which users can themselves pick and assemble argument
components (Digalo, Athena), which allow modeling their deliberations, too;
14 PART 1: Information Systems for Science and Education
evaluating justification of statements by weighing single pro and contra arguments with the
help of special assignments (Carneades, ArguMed), which support determining the solutions [25].
Recently the Critical Thinking Skills BV, the developers of Rationale, have proposed a new software
for modeling decision making bСisive (https://www.bcisiveonline.com), which is based on the concept
of deliberative protocol [26]. They suggest bCisive for visualization of deliberative reasoning and
decision support and consciously avoid differentiating between those two otherwise distinct modes of
practical argumentation.
In other approaches some developers propose to supplement the ontology of argumentation with
"means that allow modeling the audience to which the arguments are directed, and means that allow
representing the content of the statements included in the arguments" [8], which open a possibility of
taking into account the parameters relevant for the tasks of discovering arguments with special focus
on deliberative argumentation.
As regards the modelling of the deliberative argumentation, most of these developments towards
creating the software are capable for modelling it either as a side result of their modelling of
argumentation and reasoning, in general, or are adjustable for that with subsequent reservations. At the
present, there is no software comprehensively aimed at supporting the deliberative argumentation with
functions of evaluating arguments and finding solutions.
3. Guidelines for the software for modeling and representation of deliberative
argumentation with a resolution function
There are diverse approaches and methods to the development of the software designed to model
and represent argumentation. The developers seldom clearly indicate the requirements and criteria by
which they were guided when creating their software. We examined the software toolkits [23, 27, 28]
along with the conceptual approaches to their design [8, 29, 30] and found a number of problems that,
on the one hand, restrict the comprehensive use of the software for modeling argumentation and
deliberative reasoning, and, on the other hand, resist development of a unified general approach to
designing of the software for both representation of argumentation and deliberative reasoning and
implementing algorithms for searching solutions:
- unavailability of systems’ technical documentation, which prevents implementation of the
successful solutions in further developments and creating of the integrative solutions based on using
the advantages found in different systems. The documentation for the system installation as well as in
the user manuals, which is available in many cases, is of little help for solving those tasks;
- low flexibility in the system settings, which prevents configuring it for specific use. For example,
the preset argumentation schemes or types of visualization presuppose no modifications;
- implementation of specific conceptual foundations restricts application of the software for solving
a wide range of tasks in modeling argumentation.
There are two other obstacles to exploring and approbation of the software: some products are no
longer supported by their developers; others are described only in research papers (ProGraph, ConArg2)
which contain no links to the software itself. In general, most projects in the field explore just some of
the special aspects of the software design, and very few of them comprehensively focus on its design
and development. The special properties of the software for modelling of the deliberative argumentation
are left outside the research scope of those projects.
One of the notable achievements in the examination of the software is the LASAD (Learning to
Argue - Generalized Support Across Domains) software platform [3, 30, 31, 32] developed with the
support from the German Research Foundation (DFG) (https: // www.
dfki.de/en/web/research/projects-and-publications/projects-overview/projekt/lasad/) by the German
Research Center for Artificial Intelligence in cooperation with Clausthal University of Technology in
2008-2010. The LASAD team explored the existing software and approaches its creation [23],
compared them to the platform developed by themselves and proposed a concept for the creating of a
software platform which would consider the challenges and shortcomings in existing systems identified
by the team. One of the LASAD goals was to simplify the creation of the formal argumentation systems
by means of a flexible configuration mechanism [27], for which the team formulated the special
requirements and implemented them in developing of their platform:
IMS-2021. International Conference “Internet and Modern Society” 15
1) general properties – special conditions for installation, maintenance and use;
2) cooperation (joint work) – toolkits supporting joint work;
3) analysis and feedback implementing machine learning in the libraries of samples and templates;
4) ontology, based on definite conceptual foundations (Tulmin [33] or Wigmore [34]) and providing
the possibility of employing the system for solving various tasks belonging to diverse subject areas;
5) diversification of the options for visualization and representation in the data sets including
argument maps;
6) journaling for the discovering, modelling and restoration of the argumentation processes and
output in full-fledge explicit forms for spotting fallacies. This requirement ensures the entry of new
participants into the already running joint activities including the argument mapping.
These groups of requirements clearly aim at creating of a software system that can be effectively
used in education for training of practical argumentation and critical thinking skills relevant in many
subject areas. Modular approach of the software designed according to the LASAD requirements
presupposes flexibility and extensibility, which allows creating, updating, and applying of the special
modules with additional functional potential for solving specific tasks. The architecture of the designed
platform reflects the modular approach.
The LASAD system of requirements includes no special guidelines for modelling of the deliberative
argumentation, although it contains some elements adjustable to support the deliberative reasoning.
Another restriction is that it lacks explicit criteria which would allow implementation of the function
for identifying the solutions. Yet another restriction is that the platform is available only in the form of
source codes (https://sourceforge.net/projects/lasad/) and is impossible to properly testify its work, as
it is available in its beta-version, and its demo version is blocked by an empty link to (http://lasad-
demo.cses.informatik.hu-berlin.de).
The developers of another kind of the software suggest employing of an ontological approach with
an extensible ontology [8]. The proposed extension is justified by the tasks of modelling of
argumentation in popular scientific discourse, where it is necessary to consider the reliability of the
sources of scientific information or the characteristic properties of the audience. They rely on the AIF
ontology (Argument Interchange Format) [35] which represents arguments as graphs. The software has
the following functions [29]:
- storage of argumentative markup of texts, as well as of the information about the source of
argumentation (storage of annotated text corpora);
- genre-, subject area- and linguistic-sensitiveness to the style of the discourse, where the
argumentation at question is found;
- a comprehensive analysis of the created argumentation graphs (argumentation maps).
The software can verify the argumentation graphs as an option of the general assessment of the
argumentation. The automatic verification algorithms of the software can search for the cycles, analyze
the connections, consider the textual indicators of argumentation, compare the obtained maps. For the
automated analysis of argumentation, the software proposes the following functions: search in the
corpus of experts' output in the system; preparatory processing of texts with marking out the indicators
of argumentation; assessment of the arguments’ persuasiveness.
The developers certified their software and registered it according to the legal rules of the Russian
Federation [36]. Although the software is thoroughly described and screenshotted in the academic
papers, nevertheless its unavailability for regular testifying and use limits its assessment to purely
theoretical. According to the papers, the key advantages of the software include the possibility of
extending the ontology with deliberation elements (value attitudes, weights of arguments, etc.), as well
as a special algorithm that "calculates the weights of conclusions by carrying out calculations along a
chain, in which the conclusion inferred out of an argument serves as a premise for the next argument,
including the pieces of reasoning in which the chain mapped in one and the same graph involves not
only supporting claims but the conflicting claims as well as [29] ". For the calculations, the system is
operated by a truth values algebra based on fuzzy logic. Alternatively, it contains the algorithm for
weighing of premises and conclusions by user manual assignments. Judging by these properties, the
software can be classified as proposing a mechanism for solving argumentative tasks and can be applied
for automated decision-making in the deliberative reasoning.
Its key restrictions amount to the risks of subjectiveness in the manual assessment of the premises
and in its non-flexibility of varying the modes of evaluating arguments in relation to different types of
16 PART 1: Information Systems for Science and Education
dialogues. Plausible arguments can be acceptable in deliberations as well as in other types of dialogues,
in which they can be assigned with positive weights. However, such arguments can be fallacious in the
discursive argumentation, for example in formal or critical discussions which instantiate what we call
scientific discussions, and in those dialogues the same plausible arguments have to be assigned with
negative weights. Other restrictions of the software include the following:
- visualization is limited to graph representation;
- the system is limited to the analysis of argumentation in popular science discourse, although the
developers promise to further elaborate the software for making it applicable in broader subject areas;
- there are no functions of joint activity, feedback and restoration of argumentation.
In general, the descriptions of the key functions of the software and its general functional properties
can be taken as requirements for the design and development of that kind of software.
The above considered approaches to designing and creating of the software for modeling and
representation of argumentation point to two essential shortcomings. Either there are no requirements
or criteria that are explicitly put as those that should be or are taken into account in its development
with respect to solving broad or specific tasks related to the deliberative argumentation, or the software
or approach to creating it exhibit sensitive functional limitations for its use, which are generated by
overly broad or narrow criterial toolkit.
We propose our approach to the development of a body of criteria (requirements) that have to be
considered in the development of the software for modeling the deliberative argumentation. The
proposed criteria include the guidelines for implementation of a function of arguments’ evaluating and
finding solutions, and can be taken into account both in the applied and the conceptual agendas of
designing of the software. Our proposal is based on the three following issues:
- exploration in the research approaches and publications relevant to designing and development of
that kind of the software;
- the results of our own research;
- our experience of using the special software in research and teaching.
We propose the following five (groups of) criteria which take into account definite special properties
of the deliberative argumentation as well as presuppose necessary functional options for arguments’
evaluation and search for solutions (Table 1).
In the technical documentation of the software, it is preferable to explicitly reflect the cases when
the developers consider some (group of) criteria relevant or irrelevant for the software they create.
The development and use of ontologies belong to the key logical and conceptual criteria determining
the possibility of modeling of the deliberative argumentation. We propose to use four kinds of
ontologies and to implement them as the corresponding libraries: arguments, relations (functions),
dialogues (disputes), and agents. As a foundation for their construction, we suggest employing the
Argument Interchange Format (AIF-Argument Interchange Format) proposed by an international team
of argumentation researchers [35]. AIF covers the first three libraries, but includes no elements for
agent profiling. At the present stage, AIF is a common platform for the following three different trends
in the development of the software products for modelling argumentation:
Argumentation protocols, for example, ASPIC+ with molecular arguments, [37],
Software for visualization of argumentation, such as Rationale [38] or OVA [39],
Descriptive logic matching tools of mathematical logic and IT-representation of knowledge
[40, 41].
The AIF is a template for building ontologies, and it is a result of the collective efforts of the
scientists in their development of those three directions and in creating it as a lingua franca of formal,
or computational, argumentation analysis. Similar to how gadget users are divided into those who prefer
either iPhones or android smartphones, AIF divided the software products and the formalisms for the
analysis, modeling and visualization of argumentation into two groups, into those which employ that
format as a basic ontology or those which are based on the specially constructed formats. This allows
classifying the software products with respect to the ontology employed. Thus, the LACAD project
employs not AIF, but a different specially created ontology. The developments of Russian scientists [1]
and [8] are based on AIF.
IMS-2021. International Conference “Internet and Modern Society” 17
Table 1
Necessary criteria for developing of the software for modeling the deliberative argumentation,
evaluating arguments and finding solutions
Groups of Criteria Explanation
criteria
logical Syntactical and semantic The criteria consider the qualitative structure
aspects of arguments of arguments, requirements for ontologies
Dialogue graph representation and argumentation schemes. For example,
Modifiable ontologies argumentative marking involve examination
of the semantic and syntactic aspects of the
structural elements of the created schemes
and diagrams, the compositional relations
between atomic and molecular elements,
etc.
Pragma- Rhetorical text mapping Considering and profiling of the speech
linguistic Coding and decoding of actions by which arguments are put forward
messages
Communicative Multi-use options for joint work These criteria ensure the possibility of using
Support of collaboration in the software for deliberation both in the
deliberation professional activity for collaboration and
joint work of individual participants and
groups, and in teaching and training of the
corresponding skills.
Methodological Modifiable argumentation Reflect the goals of the software and special
Defeasible arguments features of its application
Journaling deliberations
(protocols)
Digital- Modular architecture Relate the aspects of the software application
technological Options for extending or to its design and creation
modifying of the software
Support of the user-friendly
configuration by web- interface
Support of cross-platform
adaptability
Exportation of the
argumentative maps (schemes,
diagrams) in the formats
supported by other widely used
software
Journaling and profiling of the
software design and work
The basic AIF ontology contains two key groups of elements which can be viewed as conceptual
and formal. To express them, AIF provides two ontologies, an ontology (conceptual) of forms and a
top-level ontology, respectively. The formal elements represented by the top-level ontology are a kind
of syntax for representing arguments by means of graphs which consist of nodes and edges. The
ontology of forms reflects the substantive elements of arguments, such as premises, conclusions,
assumptions, exceptions, schemes of argumentation, criticisms, etc., which are designed for making the
top-level ontology meaningful by representing individual arguments, for example, the deductive or the
plausible, or representing the types of disputes. AIF and the visualization of arguments with the help of
ontologies based on this format can be compared to Wigmore's argumentation and Toulmin's
argumentation models, respectively.
18 PART 1: Information Systems for Science and Education
In the top-level ontology, there are two types of nodes, information nodes (I-nodes) containing
information about the elements of the molecular arguments - premises, conclusion, exclusions, etc., and
circuits, or schematic nodes (S-nodes), representing the types of atomic arguments by their structure
and forming the three following groups:
RA (Rules of Arguments) nodes of inference rules,
PA (Preferred Argument) preference nodes,
CA (Conflict Argument) nodes (types) of conflicts of opinions.
S-nodes act as nonspecific structural or functional schemes for I-nodes.
The nodes RA, CA and PA express the properties of argumentation at its three levels, respectively,
on the level of individual arguments, of the relations between arguments in the framework of the sets
of arguments presented by the agents of the dispute, and of the assessments of individual arguments
relative to each other. In the three directions of the analysis of argumentation, in their formalisms, the
nodes RA, CA and PA are used with different degrees of detailing.
At the present, RA nodes are the most developed, they imply two types of inference rules and divide
arguments by the method of demonstration, the connection between premises and conclusions, into the
deductive and defeasible arguments. We consider this division confusing and below propose a different
one.
CA nodes are designed to express schemes of criticism and differentiate between its two types,
symmetric, when in a pair of arguments one attacks the other and vice versa, and asymmetric, when in
the pair one attacks the other, but not vice versa. With respect to the elements of argumentation, between
which the relation of criticism is established, the CA nodes mark two of its structural types, between
the points of view of the parties and between the arguments the parties put forward for their defense or
refutation. In relation to criticism and refutation, the CA-nodes contribute to distinguishing between the
kinds of disputes depending on the type of disagreement in opinions and imply two types of disputes:
asymmetric dispute-disagreement, when one agent defends his or her point of view from doubts or
criticisms of another agent who have no point of view other than the opposite to the first one; and a
symmetrical dispute-conflict, when each agent defends his or her point and criticizes the opposite point
of view. The dispute-conflict can be viewed as two corresponding disputes-disagreements. The varieties
of asymmetrical CA nodes are used to express refutation, by which one argument attacks another one
in two ways: by undermine which questions the premise or undercut which doubts the demonstration.
The undermine and the undercut can be refined by considering the relevant argumentation schemes.
The least developed are PA nodes, designed to express the ratio of assessments of the acceptability
of arguments and to play an important role in the search and selection of dispute solutions.
AIF provides three types of relations between elements of the two ontologies: to be a subclass, to
fulfil, and to include. For example, CA nodes are a subclass of S-nodes, they fulfil (functions of)
criticism schemes and include two kinds of elements, the attackers and the attacked.
Ontologies generated by means of AIF model a dispute in the form of a directed graph, the nodes
and edges of which model the arguments put forward in the dispute and forming up its network of
arguments. Depending on the properties of the formalism created on the basis of AIF, the nodes express
the necessary properties of arguments, such as inferential quality, acceptability, belonging to the
position of an agent, etc., while the edges characterize the three types of connections between arguments
or relations between their internal elements. The edges of information connecting I-nodes with S-nodes
represent the structure of information at the level of individual arguments, for example, the function of
an argument premise fulfilled by a proposition. The edges of inference connecting S-nodes to I-nodes
express the kind of demonstration, or the kind of argument; and the edges of justification connecting
different S-nodes to each other represent the structure of argumentation within an agent's position or on
a (sub-) set of arguments in the dispute.
For modelling of the deliberative argumentation, we propose to supplement AIF (Fig. 1) to DelibAIF
(Fig. 2) by means of the following three modifications.
IMS-2021. International Conference “Internet and Modern Society” 19
Figure 1: Standard AIF
Figure 2: Modified AIF (DelibAIF) for modelling of the deliberative argumentation
First, in RA nodes of inference schemes, we propose to abandon the vague division of schemes into
the deductive and the defeasible and to replace it with a division into three classes: deductive, inductive
and plausible schemes. Then, indefeasible schemes will consist of the first two classes of the deductive
20 PART 1: Information Systems for Science and Education
and the inductive schemes; and the second and the third classes, i.e. the inductive and the plausible
schemes, together with make up the class of the non-deductive schemes. There is no need of adding
elements such as indefeasible or non- deductive schemas to DelibAIF as separate subclasses of the
Inference Schemes class.
Secondly, to the four structural elements already present in the AIF - premise, conclusion,
assumption and exclusion we propose to add the following five: generalization, cause, goal, value,
norm. The elements premise and conclusion are necessary in any argument, so they are necessary
elements of each of the three schemes. The rest of the elements are required for expressing of the
properties of the premises of the inductive or plausible arguments: generalization, cause, assumption
and exclusion - for the inductive arguments; and cause, admission and exclusion, and the rest of the
elements - for the plausible arguments. The elements goal, value, norm are necessary for modeling the
deliberative arguments which are a part of the class of plausible arguments. These elements mark out
the specific premises of the practical arguments and reflect the properties of reasoning about actions
that are not characteristic of other plausible arguments.
Thirdly, we propose to treat the two subclasses Scheme of discursive conflict and Scheme of
deliberative conflict as the subclasses of the element of the ontology of forms Scheme of Conflict and
to establish the relation to fulfil between the elements Attacker (Attacking element) and Attacked
element and those two Schemes. This allows to distinguish between the deliberative, or practical,
argumentation from the discursive, or theoretical.
The proposed modifications open the possibility of completing of the library of arguments with the
plausible arguments about actions, the library of disputes - with the disputes about actions, and the
library of relations – with the relations between special elements of the practical arguments inside the
structure of those arguments, at the level of the agent’s position in the dispute and at the level of the
whole dispute. For modelling of the agents of argumentation, be it discursive or deliberative
argumentation, the corresponding library of agents has to be generated separately, since AIF lacks
expressive abilities for providing agent profiles and reduces the cognitive diversity of agents to the
information diversity expressed by I-nodes.
4. Conclusion
We proposed a preliminary approach to the formulation of criteria that have to be considered when
developing the software for modeling and representation of the deliberative argumentation with the
function of evaluating arguments and finding solutions. However, already at the initial stage, we
propose grouping the criteria for reflecting the key properties of that kind of the software. We suggest
a modified DelibAIF scheme which allows modeling the deliberative argumentation.
Since for modelling of argumentation, in Russia we have neither domestic, nor localized software,
we propose the corpus of (the groups of) the criteria for providing the methodological support in
generating guidelines and recommendations for the creation of the software and applications for
modeling and representation of argumentation, deliberative reasoning, which will support decision-
making, teaching argumentation and training the critical thinking skills. The development of the corpus
of criteria aims at methodological support of the academic, research and educational communities and
at providing them with the effective selection tools for using the software in their research and teaching
activities related to the deliberative argumentation.
In our further research we intend to classify the properties of the software according to the five
(groups of) the criteria given in Tab.1., to testify both the criteria and their grouping against the existing
and newly developed software, and to update the body of the criteria, if needed. Its another application
will be a comprehensive classification of the software and systems for modeling and representation of
argumentation, the deliberative reasoning, support of decision-making processes and training of
argumentation and critical thinking skills. The classification will enhance the quality of users’ decisions
regarding the choice of the software and applications for solving their practical tasks.
IMS-2021. International Conference “Internet and Modern Society” 21
5. Acknowledgements
The support from Russian Foundation for Basic research, project 20-011-00485a, is kindly
recognized.
6. References
[1] E.N. Lisanyuk, Argumentacija i ubezhdenie. SPb, Nauka 2015. [In Russian]
[2] T. Davies, R. Chandler, Online deliberation design: Choices, criteria, and evidence, in: Nabatchi
T., Weiksner M., Gastil J., Leighninger M. (Ed.), Democracy in motion: Evaluating the practice
and impact of deliberative civic engagement, Oxford, Oxford univ. press., 2013, pp. 103-131.
doi:10.1093/acprof:oso/9780199899265.003.0006
[3] F. Loll, N. Pinkwart, Collaboration Support in Argumentation Systems for Education via Flexible
Architectures, in: Ninth IEEE International Conference on Advanced Learning Technologies,
2009, pp. 707-708. doi: 10.1109/ICALT.2009.55
[4] T. Nabatchi, M. Weiksner, J. Gastil, M. Leighninger (Ed.), Democracy in motion: Evaluating the
practice and impact of deliberative civic engagement, Oxford, Oxford univ. press, 2013.
doi:10.1093/acprof:oso/9780199899265.001.0001
[5] M. Klein, L. Iandoli, Supporting Collaborative Deliberation Using a Large-Scale Argumentation
System: The MIT Collaboratorium, in: Proceedings of the Eleventh Directions and Implications
of Advanced Computing Symposium and the Third International Conference on Online
Deliberation (DIAC_ 2008/OD 2008), Berkeley, California, 2008, pp. 5-12. doi:
10.2139/ssrn.1099082
[6] V. Aleven, K.D. Ashley, Teaching case-based argumentation through a model and examples:
Empirical evaluation of an intelligent learning environment, in: Proceedings of the 8th
International Conference on Artificial Intelligence in Education (AI-ED 1997). Amsterdam, IOS,
1997, pp. 87–94.
[7] N. Pinkwart, V. Aleven, K. Ashley, C. Lynch, Toward legal argument instruction with graph
grammars and collaborative filtering techniques, in: M. Ikeda, K. Ashley, T.W. Chan (Ed.),
Proceedings of the 8th International Conference on Intelligent Tutoring Systems (ITS 2006),
Berlin, Springer, 2006, pp. 227–236.
[8] Yu.A. Zagorulko, N.O. Garanina, O.I. Borovikova, O.A. Domanov. Argumentation modeling in
popular science discourse using ontologies, Ontology of Designing Vol. 9 4(34) (2019) 496-509.
doi: 10.18287/2223-9537-2019-9-4-496-509. [In Russian]
[9] T. Davies, S.P. Gangadharan (Ed.), Online Deliberation: Design, Research, and Practice, Stanford,
CSLI Publications, 2009.
[10] D.D. Suthers, J. Connelly, A. Lesgold, M. Paolucci, E.E. Toth, J. Toth et al, Representational and
advisory guidance for students learning scientific inquiry, in: K. D. Forbus, P. J. Feltovich (Ed.),
Smart machines in education: The coming revolution in educational technology, Menlo Park,
AAAI/MIT, 2001, pp. 7–35.
[11] E.N. Lisanyuk, D.E. Prokudin, Modelling argumentation with OVA and Rationale (a case-study),
in: Internet i sovremennoe obshchestvo: sbornik tezisov dokladov [Elektronnyy resurs], Trudy
XXI Mezhdunarodnoy ob"edinennoy nauchnoy konferentsii « Internet i sovremennoe
obshchestvo» (IMS-2018), Sankt-Peterburg, 31 maya – 2 iyunya 2018 g., SPb, Universitet ITMO,
2018, pp. 14-17. URL: http://ojs.itmo.ru/index.php/IMS/article/view/719. [In Russian]
[12] E.N. Lisanyuk, D.E. Prokudin, Software for the representation of deliberative argumentation: the
conceptual foundations and the properties of classification and use, International Journal of Open
Information Technologies 8.11 (2020) 49-56. URL:
http://injoit.org/index.php/j1/article/view/1025. [In Russian]
[13] E.N. Lisanyuk, D.E. Prokudin, Conceptual Bases of Software Functioning for the Representation
of Deliberative Argumentation, in: Information Society: Education, Science, Culture and
Technology of Future. Issue 4 (Trudy XXIII Mezhdunarodnoy ob"edinennoy nauchnoy
konferentsii «Internet i sovremennoe obshchestvo», IMS-2020 (sbornik nauchnykh statey), SPb,
Universitet ITMO, 2020, pp. 34-41. doi: 10.17586/2587-8557-2020-4-34-41. [In Russian]
22 PART 1: Information Systems for Science and Education
[14] D.D. Suthers, Representational guidance for collaborative inquiry // Arguing to learn: Confronting
cognitions in computer-supported collaborative learning environments, in: J. Andriessen, M.J.
Baker, D.D. Suthers (Ed.), Dordrecht, Kluwer Academic, 2003, pp. 27–46.
[15] M. Ranney, P. Schank, Toward an integration of the social and the scientific: Observing, modeling,
and promoting the explanatory coherence of reasoning, in: S. Read, L. Miller (Ed.), Connectionist
models of social reasoning and social behavior, Mahwah, Erlbaum, 1998, pp. 245–274.
[16] W. Kunz, H. Rittel, Issues as elements of information systems, Working paper #131, Institut für
Grundlagen der Planung I.A. University of Stuttgart, Germany, 1970. URL:
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.134.1741&rep=rep1&type=pdf.
[17] J. Conklin, M.L. Begeman, gIBIS: A hypertext tool for exploratory policy discussion, in:
Proceedings of the ACM Conference on Computer-supported Cooperative Work (CSCW ‘88),
New York, ACM, 1988. P. 140–152.
[18] S.J. Buckingham Shum, A.M. Selvin, M. Sierhuis, J. Conklin, C.B. Haley, B. Nuseibeh,
Hypermedia support for argumentation-based rationale: 15 years on from gIBIS and QOC, in:
Rationale management in software engineering / A. H. Dutoit, R. McCall, I. Mistrik, B. Paech
(Ed.), Berlin, Springer, 2006, p. 111–132.
[19] F.H. van Eemeren, R. Grootendorst, A Systematic Theory of Argumentation, Cambridge
University Press, 2004.
[20] N. Karacapilidis, D. Papadias, Computer supported argumentation and collaborative decision
making: the Hermes system, Information Systems 26.4 (2001) 259-277.
[21] T.F. Gordon, H. Prakken, D. Walton, The Carneades model of argument and burden of proof,
Artificial Intelligence 171.10-15 (2007) 875–896.
[22] M. Janier, J. Lawrence, C. Reed, OVA+: an Argument Analysis Interface, in: Proceedings of the
Fifth International Conference on Computational Models of Argument (COMMA 2014), IOS
Press, Pitlochry, 2014, pp. 463-464, URL:
http://www.arg.dundee.ac.uk/people/chris/publications/2014/comma2014-ova.pdf.
[23] O. Scheuer, F. Loll, N. Pinkwart et al, Computer-supported argumentation: A review of the state
of the art, Computer Supported Learning 5 (2010) 43–102. doi: 10.1007/s11412-009-9080-x.
[24] D. Walton, Ch. Reed, F. Macagno, Argumentation schemes, Cambridge UP, 2008.
[25] B. Verheij, Artificial argument assistants for defeasible argumentation, Artificial Intelligence
150.1–2 (2003) 291–324.
[26] K. Atkinson, T. Bench-Capon, Practical reasoning as presumptive argumentation using action
based alternating transition systems, Artificial Intelligence 171 (2007) 855–874.
[27] F. Loll, N. Pinkwart, O. Scheuer, B.M. McLaren, How Tough should it be? Simplifying the
Development of Argumentation Systems Using a Configurable Platform, in: N. Pinkwart, B.
McLaren (Ed), Educational Technologies for Teaching Argumentation Skills, Bentham Science
Publishers, Sharjah, United Arab Emirates, 2012, pp. 169-197.
doi: 10.2174/978160805015411201010169
[28] O. Scheuer, B. McLaren, F. Loll, N. Pinkwart, Automated Analysis and Feedback Techniques to
Support and Teach Argumentation: A Survey, in: N. Pinkwart, B. McLaren (Ed), Educational
Technologies for Teaching Argumentation Skills, Bentham Science Publishers, Sharjah, United
Arab Emirates, 2012, pp. 71-124. doi: 10.2174/978160805015411201010071.
[29] E.A. Sidorova, I.R. Akhmadeeva, Yu.A. Zagorulko, A.S. Sery, V.K. Shestakov, Research platform
for the study of argumentation in popular science discourse, Ontology of designing 10.4 (2020)
489-502. doi: 10.18287/2223-9537-2020-10-4-489-502. [In Russian]
[30] F. Loll, N. Pinkwart, LASAD: Flexible representations for computer-based collaborative
argumentation, International Journal of Human-Computer Studies. 71.1 (2013) 91-109. doi:
10.1016/j.ijhcs.2012.04.002.
[31] F. Loll, O. Scheuer, B.M. McLaren, N. Pinkwart, Learning to Argue Using Computers – A View
from Teachers, Researchers, and System Developers, in: V. Aleven, J. Kay, J. Mostow (Ed),
Intelligent Tutoring Systems, ITS 2010, Lecture Notes in Computer Science, Springer, Berlin,
Heidelberg, 2010. Vol. 6095, pp. 377-379. doi: 10.1007/978-3-642-13437-1_76
[32] O. Scheuer, B.M. McLaren, F. Loll, N. Pinkwart, An Analysis and Feedback Infrastructure for
Argumentation Learning Systems, in: Proceedings of the 2009 conference on Artificial
IMS-2021. International Conference “Internet and Modern Society” 23
Intelligence in Education: Building Learning Systems that Care: From Knowledge Representation
to Affective Modelling, IOS Press, NLD, 2009. pp. 629–631.
[33] S.E. Toulmin, The Uses of Argument, Cambridge University Press, 1958.
[34] J. H. Wigmore, The Principles of Judicial Proof, 2nd Edition, Little, Brown & Co, 1931.
[35] C.I. Chesñevar, J. McGinnis, S. Modgil, I. Rahwan, C. Reed, G. Simari, M. South, G. Vreeswijk,
S. Willmott, Towards an argument interchange format, The knowledge engineering review 21.4
(2006) 293-316.
[36] Yu.A. Zagorul'ko, E.A. Sidorova, A.S. Seryy, O.I. Borovikova, O.A. Domanov, I.S. Kononenko,
V.K. Shestakov, I.R. Akhmadeeva, Programmnyy kompleks dlya modelirovaniya i analiza
argumentatsii v nauchno-populyarnykh tekstakh ArgNetBank Studio, 2020. Svidetel'stvo No.
2020665092, No. 2020663982, Filed November 9, 2020, Issued November 20, 2020. [In Russian]
[37] H. Prakken, An abstract framework for argumentation with structured arguments, Argument and
Computation 1 (2010) 93–124.
[38] T. Berg, T. van Gelder, F. Patterson, S. Teppema, Critical Thinking: Reasoning and
Communicating with Rationale, Amsterdam, Pearson Education Benelux, 2009.
[39] F.J. Bex, C.A. Reed, Schemes of Inference, Conflict and Preference in a Computational Model of
Argument, Studies in Logic, Grammar and Rhetoric, 2011.
[40] V.A. Lapshin, Ontologies in computer systems, Moscow, Nauchny mir, 2010. [In Russian]
[41] I. Rahwan, B. Banihashemi, C. Reed, D. Walton, S. Abdallah, Representing and classifying
arguments on the semantic web, The Knowledge Engineering Review 26.4 (2011) 487-511.