Justificatory and Explanatory Argumentation for Committing Agents

04/13/2013 ∙ by Ioan Alfred Letia, et al. ∙ UTCluj 0

In the interaction between agents we can have an explicative discourse, when communicating preferences or intentions, and a normative discourse, when considering normative knowledge. For justifying their actions our agents are endowed with a Justification and Explanation Logic (JEL), capable to cover both the justification for their commitments and explanations why they had to act in that way, due to the current situation in the environment. Social commitments are used to formalise justificatory and explanatory patterns. The combination of ex- planation, justification, and commitments



There are no comments yet.


page 1

page 2

page 3

page 4

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.

1 Introduction

The institutional economics demands a clear account of a fuzzy world and is therefore in accordance [17] with the dictum attributed to Keynes: "it is better to be roughly right than precisely wrong". As many multi-agent systems are organized in some kind of institution, judging the behavior of agents in such contexts requires sometimes understanding of how their actions stand vis-a-vis to the goals of the overall system.

Habermas [17] proposed different types of discourses dealing with different validity claims. In the interaction between agents we can thus have an explicative discourse, when communicating knowledge, and a normative discourse, when considering normative knowledge. For justifying their actions our agents are endowed with a Justification and Explanation Logic (), capable to cover both the justification for their commitments and explanations why they had to act in that way, due to the current situation in the environment. The distinction between factual and normative knowledge is important in argumentative agents since facts of various kinds might not provide agents with any choice, while norms are possible to be bent and eventually changed.

The difference between explanation and justification has not been clearly delimited in computational models of arguments. From this perspective, there is a gap between argumentation in the philosophy of science and computational-based argumentation. Given the idea that argumentation is a means to justify claims or to persuade other agents of these claims [10], there are two approaches when defining what "good argumentation" represents: i) argumentation able to justify its target claims and ii) argumentation able to convince an audience. These two lines form the basis for investigating the contrast between justificatory arguments and explanatory arguments.

The philosophy of science distinguishes between explanatory reasons and normative or justificatory reasons [12]. While explanations are reasons why events occur, justifying reasons are "considerations which count in favor" or "explanations of ought facts". Differently, in the computation models of arguments there is a blurred distinction between explanation-based argumentation and justification-based argumentation, as in [16], where explanations and evidence are used to construct justifications.

In many practical domains the distinction between explanation and justification has pragmatic force. Many examples come from the legal domain, where an illegal action can be explained, but in many cases the action is not justified. For instance, a theft can be explained by the loss of money because the person lost his job, but it does not have normative justificatory power. Another common example is based on the well known lack of time explanatory pattern: "I could not finalize the task due to lack of time", with its many instantiations: "I could not review the paper because I was at a conference during the deadline" or "I could not finalize the article, because input data arrived too late". In most situations, this explanation pattern does not have justificatory power at all. In other words, something can explain a behavior but it cannot always justify it.

In this paper we investigate the inter-living of explanations and justifications in business-oriented situations. Current business interactions are affected by postmodern ideas like post-structuralism and heterogeneity at different levels. In line with postructuralism, business commitments act as a flexible framework for guiding business interactions, agents preferring soft law against the hard law governance, whilst the idea of heterogeneity of individuals is reflected at the business policy level through the concept of heterogeneity of customers [20]. Consider the contractual clause in which the debtor promises to deliver an item to the creditor within a pre-agreed deadline. Not meeting the deadline, the agent provides an explanation: "I could not deliver the product because my supplier has not delivered the parts yet". This explanation provides the creditor with some insights on the current situation, but it does not have enough justificatory power in order to justify the behavior of agent for not delivering the item. It may be a case in which the supplier has no normative obligation to deliver the components of the product: either a commitment for delivering the parts does not exist between and the supplier , or the agent did not pay the components in due time. Valid justifications should be normative, like: "emergency situation forces me not to deliver the item", where emergency is a normative concept. Observe that the justifier does not help the debtor to understand the situation. The agent can ask for further explanations for the conveyed justifier, but also for further normative justifications for it.

The paper advances the state-of-the-art in logic of argumentation in three ways: i) proposing for handling both justificatory and explanatory arguments; ii) introducing commitments as proof terms; and iii) formalising several justificatory and explanatory patterns. The remaining of the paper is organised as follows: Section 2 extends justification logic [2] with explanatory capabilities and introduces commitments as proof terms. The expressivity of social commitments is exploited in section 3 to represent both justificatory and explanatory patterns. Section 4 describes types of argumentative agents based on the combination of justificatory and explanatory attitudes, whilst section 5 illustrates the developed instrumentation through an illustrative scenario. We end the paper with related work and conclusions.

2 Commitment-Based Justification and Explanation

The first part of this section extends the justification logic with explanatory capabilities. The second part illustrates how social commitments are used to represent proof terms in justification logic.

2.1 Enhancing Justification Logic with Explanation

In order to model justificatory and explanatory arguments we propose an extended version of the Justification Logic (JL). Justification Logic combines ideas from epistemology and the mathematical theory of proofs. It provides an evidence-based foundation for the logic of knowledge, according to which "F is known" is replaced by "F has an adequate justification".

Simply, instead of "X is known" () consider , that is, "X is known for the explicit reason t" [2]. Justification logic lacks this component of interpretation. It also lacks the capability to express explanation or partial justification. This section extends the justification logic with explanatory capabilities, by introducing the explanatory operator , where is an explanation for .

Definition 1

The language of Justification and Explanation Logic contains proof terms and formulas

Proof terms are abstract objects that have structure. They are built up from axiom constants , proof variables , and operators on justifications and explanations , , ,,. The application operator takes two proof terms and constructs a new justification based on them. The sum operator concatenates the given proofs, whilst the unary operators and are used to request for positive and negative proof terms for a formula in . The operator precedence decreases as follows: , and are left associative, and right associative. To express that is not probative justification for supporting one uses . Parentheses are needed to express that is a justification for : . Note that justification is used to support negated sentences too, as in . Similar semantics applies for the explanation operator .

classical propositional axioms
(proof checker)
(explanation checker)
(negative proof checker)
(negative explanation checker)
Figure 1: Axioms of .

The axioms of are shown in figure 1, where axiom forces all formulas to have a justification or an explanation . The compounds or represent a formula, which should have their own justification. This corresponds to the principle of inferential justification: for sentence to be justified on the basis of one must justify and justify that makes plausible. Constants are used to stop the ad infinitum justification chain by representing a kind of justification that does not depend on other justifiers.

The application axiom takes a justifier of an implication and a justifier of its antecedent , and produces a justification of the consequent . If at least one of the terms , represents an explanation, the formula is considered only explained, but not justified (axioms , , and ).

The j-sum axiom says that if a formula is justified by the justifier , than for a new justifier

, the formula is still justified. Thus, justification reasoning is monotonic, new justification not defeating the existing one. The corresponding axiom for explanation is missing, leaving space for contradictory explanations and non-monotonic explanatory reasoning. The rationality behind this is that norms are considered static at a given moment and assumed apriori known by the participants, whilst explanations are constructed dynamically during a dialog.

Justifications and explanations are assumed to be verified. Based on axiom , a justifier for formula can be further justified by the term , but it can also be explained. The axiom limits the possibility to justify an explanation, an explanation can only be further explained. The negative proof checker forces agents to provide justifications or explanations why they are not able to justify a particular formula . Justifiers cannot be used to justify why the formula is not explained by the explanandum , as axiom states that only explanans can be used. A proof term can be stronger than another one, given by the operator .

2.2 Commitment Based Proof Terms

One of the issues regards what sort of things can be a justifier. In a normative framework regulated only by social commitments, a justifier can be represented by such commitments. In the proposed approach, by restricting the acceptable justifications and explanations to commitments, it means that the proof terms in the language represent commitments.

Expression Informal Semantics
is probative evidence for for agent .
is not probative justification for for agent .
The absence of is a justification for for agent .
is evidence for for the agent .
Table 1: Commitment based justification in a multi-agent system.

The classical definition of a conditional commitment states that the debtor promises to creditor to bring about a particular formula under the condition , encapsulated as . In multi-agent systems, a justification accepted as probative evidence for an agent may not meet the standard of proof for another agent, which rejects it. To model this, we link the justification and explanatory operators to the agent accepting the evidence. Thus, the construction says that commitment is a probative justification for agent regarding the sentence (see table 1). A commitment may be preferred to another one by the agent when choosing an explanation, formalized as . If one of the terms in the commitment is not constrained in any aspect, “do not care” sign “” is used. Examples 1, 2, and 3 illustrate how the formalism is enacted in a commitment-based multi-agent setting.

Example 1 (Distributed Application)

Assume the commitment is agent’s justification for and the conditional commitment represents a justification for agent regarding the formula . According to axiom , the application operator builds a justification for of the form . Notice that the commitment aggregation is based on the view of agent .

Not being omniscient, agents are not aware of all the consequences from their commitments. When used by the same agent, the proof checker operator acts as a positive introspection function, helping agents to be aware of the justification of their own commitments. Given a commitment justifying the formula , the agent can inspect its own commitment store to further justify the formula by enacting the application of the proof checker operator on itself: . Consequently, given the commitment an agent can construct a justification for it based on other commitments.

Example 2 (Positive Introspection)

Because agent has promised agent to deliver the , given by , this is the justification of agent regarding the commitment with agent , in which agent has to pay the components provided by agent . Given the right associativity of the justificatory operator , parenthesis are introduced only for clarity:


Example 3 (Explaining a Justification)

Although a hotel specifies in the commitment that check-in is at time 14, usually there is no problem if the guest arrives at 12. The situation in which the guest is not served, is often not justified by personnel with the commitment itself (the active contractual relation) but by other valid justifiers such as: "no clean rooms at the moment". The customer can explain the situation to himself by the reduced number of staff, which can be further explained by the economical policy of the hotel, with the following chain of reasoning.

Justifying the impossibility to check-in by the existing contractual commitment leads to a different chain of reasoning.

Given the recurrent nature of business interactions, in common situations business agents convey the same justificatory and explanatory schemes. The following section formalizes two justificatory and two explanatory patterns.

3 Justificatory and Explanatory Business Patterns

An explanation involves discovering the meaning of an event in a particular context, such that an is explained by a coherent set of . The explanation aims to understanding the explanandum by indicating what causes it. An explanation answers to questions or to contrastive why P rather than X questions when the opponent requests an explanation for proponent preferences.

The theory of justification advocates the idea that justification is a normative activity, where a concept is defined as normative if it depends on norms. The thing that justifies a proposition is called . Justifiers act as a vehicle between beliefs and knowledge, as the definition of knowledge as justified true belief suggests. Supplementary to explanans, a justifier should be legitimate by objective factors such as social structures, normative frameworks, or abstract rationality.

To synthesize the differences, a justification acts in a more normative framework, whilst an explanation works in a social context. An explanation implies the existence of an audience which understands the claim and its explanans, whilst justification is in relation only with an objective world of true beliefs. The above observations are used to differentiate between two classes of business dialogs: justification patterns and explanatory patterns.

3.1 Justification Patterns

Commitment-based justification patterns are constructed in terms of the intention of the agent which provides the justification, the context in which they are usually used, the formal representation of the pattern, within specific cases. An example is illustrated by informal text and its representation in . All commitment-based justification patterns of the general form may be questioned by a common set of critical questions.

  • : If , which is the relationship between the agent and ?

  • : If , which is the link between the formulas and ?

  • : Is there a stronger commitment which does not justify ?

Specific attacking options for each justification pattern are added at the end of the pattern.

3.1.1 Gratuitous Promise.

In a gratuitous promise the debtor promises the creditor to bring about , without requesting anything: . This may serve as enough justification for agent for the achievement of , given by (figure 2). The debtor is the one affected by the violation of the commitment. In case the agent enacting the gratuitous promise is not the creditor, one continuation path regards the relationship between the proponent of the justification and the creditor. In the example from figure 2, in case is risen, the relation between father and son can be legally justified.

Justification from gratuitous promise≑GP : Stressing out that no one has forced the debtor to commit. : The creditor or other agent perform actions based on the promise. : : b=c, a=c, F=P : Grandfather promises to his nephew to pay for a trip. Based on it, the father buys a new bag for his son. :

Figure 2: Commitment-based justificatory promise.

For the attack, the justification stands only if there is an evident connection between the promise and the action . Note that the particular variant avoids the above two attacks. Assuming that in sequent step, the nephew informed that he has no bag for the trip, the grandfather reacts by assuring his nephew that he will pay both for the trip and buy a new bag. From the father’s perspective, the newly stronger commitment acts as a justification for the opposite conclusion :

The promise made by the creditor can justify future actions of the creditor itself. For instance the grandfather starts saving money due to its gratuitous promise: .

Justification from request ≑R Justifying actions based on the directive conveyed by a normative empowered agent. The creditor or other agent act on request from other agent. b=c, a=c, F=P, During driving lessons, the instructor requests the student to stop the car. Consequently, the next student’s action is to signal right. : Is the request legitimate?

Figure 3: Commitment-based justificatory patterns.

3.1.2 Directives.

In a fact request the debtor does not promise anything, it only requests the precondition to be satisfied, given by (figure 3). The additional critical question regards the normative rightness of the directive. For instance, there is no obligation for the agent to meet the request or, if there is indeed a power relationship between the creditor and the debtor, it may not have jurisdiction in the context of the sentence requested. In the example of figure 3, the requested action of stopping the car is legitimate by the relationship between the instructor and the student. If the requested act is negated, the commitment represents a or an interdiction.

Example 4 (Justification from Interdiction)

"I can not sell you cigarettes because the law interdicts to sell them to the minors" will be formalized in as

Here the law is personalized by the normative agent . According to contract law, when exposing an item for selling, an open offer is created. The seller commits to sell the item in case of acceptance, which usually occurs by payment: . The exact representation captures this when the agent justifies his refuse to sell, because he is not committed to do so, which is further justified by the normative interdiction to sell cigarettes to the minors.

The two justificatory patterns introduced above are not exhaustive in terms of basic justificatory schemes. However, by combining such basic patterns one can increase the expressivity of the justifiers. Firstly, by composing a gratuitous promise with a fact treated as a condition, a unilateral contract is obtained: . The same pattern can also be obtained by composing a request with a fact treated as a promise: .

Secondly, when the term used for composition is a commitment itself a higher order commitment [13] is constructed. In this line, justification from bilateral contract is formalized as . Here, both parties make promises: the agent commits to deliver the item if the agent promises to pay. The pattern is a composition between a unilateral contract and a justification from request pattern. The composition is applied on the requested action which represents a commitment itself: . For the given example, the justifier is constructed as follows: . The justification from promise to commit pattern is a composition of two gratuitous promises, applied on the fourth term of the commitment: , where agent promises agent that he will commit to agent to deliver the item.

3.2 Explanation Patterns

Explanation from cognitive_consistency ≑CC Explaining actions based on the goals that an agent is following. The debtor commits itself to achieve a particular sentence. a=c, Tom cannot join the party because he wants to learn for the exam. : : Is the debtor aware of his commitments?

Figure 4: Explanation for preferred commitment.

Explanation patterns are not rooted in an objective normative frame, having a subjective component. Explanatory schemes can be viewed as providing subjective reasons, with a more flexible relation between explainers and what is explained.

3.2.1 Public goals.

Consider the situation in which the agent commits itself to bring about : . It is on the edge between social semantics and mentalistic semantics of communicative agents (figure 4). The commitment belongs to the social semantics because it is public, and points toward the mentalistic approach because it represents a goal of the agent. Assuming sincere agents, the attack option questions the possibility that the agent may not be aware of all its commitments.

Explanation from preferred_commitment ≑PC Explaining choice between two commitments. The debtor is committed with different strengths to creditors. a=d, b=d, c=d The agent promised his boss to attend at a late meeting. He also promised his wife to take the child from the school if he has time. Aware of the constraint, the wife decides to go to the school herself. : Is it not possible to achieve the both commitments? Is the preference relation explained by the cognitive consistency property of the agent ?

Figure 5: Commitment-based explanatory preference.

3.2.2 Explanation from preference.

The preference relation usually has a strong subjective component, making it a candidate for explanatory arguments rather than justificatory ones. In our example (figure 5), is a gratuitous commitment, whilst is a unilateral contract. The gratuitous promise representing a stronger promise compared to a unilateral contract, the preference relation can be deduced based on the strength of each type of commitment.

4 Argumentation Framework

4.1 Arguing with Justification and Explanation

Definition 2

An argument is a pair where is the chain of justifiers or explanans and the conclusion such that justifies or explains and is minimal.

We distinguish between explanatory arguments and justificatary arguments.

Definition 3

A justificatary argument is supported by justifiers only. An explicatary argument contains at least one supporting explanandum.

A fact can be supported at the same time by several explanatory or justificatory arguments.

Definition 4 (Conflict among arguments)

In the case of an undercutting argument, its conclusion attacks one justifier or explanation in the support of another argument. In the case of rebuttals, the justified formulas contradict each other directly.

There are two types of interrogative requests: request for explanation and request for justification.

Definition 5 (Request for justification)

Agent requests agent to provide its justification why t is agent its justification for : .

The general case reflects the situation when the agents , and are different.

Example 5 (Request for justification)

Consider that the agent requests the agent to legally support the belief of the victim that is a justification for , expressed as:

Note that the justification is requested to be constructed from the perspective of the lawyer agent. The expressivity of the language allows to ask the lawyer to present the victim’s justification regarding the sentence , given by .

When , the agent is requested to justify itself. When , the agent requests itself to identify or construct a justification supporting why is accepted by the agent as valid justifier for . Not being omniscient, in , the agents are not assumed to be aware of all the justifications that can be built from their knowledge base. Only an explicit constructed justification is considered, which makes sense for the agent to interrogate its own knowledge base to identify a valid justification. Similarly, when , the same agent is applying itself to the task of justifying its own sentences.

Definition 6 (Request for explanation)

Agent requests agent to explain why is the justification of for : . Similarly, agent requests agent to further explain why is the explanation of agent for : .

Notice that one cannot justify an explanation, but an explanation can be requested both for a justifier and for another explanation. The combination of the above explicatory/justificatory patterns in dialogs helps the parties to understand normative-based decisions.

4.1.1 Strength of Justification.

The strength of a justification depends on the commitment used as evidence and the formula needed to be justified. The stronger the commitment used as a justification term, the stronger the justification.

Proposition 1

The commitment is stronger than if the debtor promises more and requests less.

Example 6 (Strength of justification)

"The supplier commits to deliver more items faster to the retailer " is stronger than "The supplier commits to deliver the items earlier if he receives the payment earlier":

In the particular case of no promised action (), the same semantics works for defining stronger requests. Thus, the request pattern is stronger or more specific than . This can be used during negotiation dialogs, where the less stronger request allows more options for the requester and introduces higher flexibility in the system.

Example 7

The request to pay either by credit card or by wire transfer is stronger to the wire transfer option, formalized as: .

Proposition 2

A fact is stronger than the commitment created to bring about that , given by .

Example 8

"I commit to deliver the item after you confirm the order" is stronger then "I commit to deliver the item if you have promised me to confirm the order": .

A preference relation on commitments can be defined based on the strength of the commitment.

Proposition 3

The creditor prefers stronger commitments, whilst the debtor prefers weaker commitments when explaining the formula . If then we have the following.

4.1.2 Conflict Resolution.

In all arguments are considered to have justifications or explanations. When deciding between two conflicting formula, the stronger justified formula will be preferred by the debtor, because the stronger commitment provides more guarantees for the promised action. If and , the formula would be accepted. In case the justifications have the same strength, forces the agents to provide further justifications on request.

Definition 7 (Preference-based conflict resolution)

If attacks and attacks and and , is preferred to if there is a chain of explanans so that and .

4.2 Argumentative Agents

Combining commitments with justification and explanation logic provides flexibility for defining several types of argumentative agents.

Firstly, regarding the logical framework, in the simplest approach, three types of agents can be formalized: The less demanding agent requires only explanations in order to accept a formula, given by the axiom . A rigorous agent will accept only normative justified formulas: . The most demanding one requests both explanation and justification before accepting a sentence: . Aggregation of these components can lead for instance to a agent, which accepts a formula if it has a valid justification and no valid explanation supporting the opposite conclusion , formalized as:

In a more elaborate agent system, the existing theories of justification (foundationalism, infinitism, internalism, externalism [12]) and explanation (causal, teleological) [15]) can be exploited to define the corresponding agent type. Each theory requires different amount and type of evidence before a formula can be considered justified or explained. For a agent, the existence of a basic justificatory pattern would be enough for accepting the supported formula. Starting from the infinitism theory, a n-type credulous agent accepts a formula if its justification chain has at least length . An agent should be able to justify a sentence only through its own commitments, whilst for an agent, third party commitments can be used as justifiers.

Secondly, regarding the commitment patterns, agents can convey or accept as valid justifiers only commitments meeting a strength threshold. By composing the basic gratuitous promise and request justificatory patterns, one obtains higher order patterns having different degrees of strength. Table 2 illustrates possible composed patterns that may justify that the payment will be made.

By promising to deliver an item and only requesting a promise for payment, the pattern is the weakest one, when justifying the formula . Only an agent having a low justification standard as scintilla of justification would accept this pattern as valid justification. A little bit stronger justification is given by the pattern where a promise for delivering and a request for paying do exist. The agent should have the reasonable justification proof standard to accept this justifier. In the bilateral contract the promise to pay representing a precondition for delivering, gives preponderence of justification for the agent to consider it as a valid justifier. The pattern meets the convincing justification standard of proof of the agent , whilst the most attractive pattern for the agent is , where it promises to deliver the item only after the payment is finalized. This justifier should meet the most skeptical standard of proof of an agent, which is behind any reasonable doubt.

Pattern Justifier Meaning
commits to deliver the item and
requests to commit to pay for it.
commits to deliver the item
and requests to pay for it.
commits to deliver the item
if commits to pay for it.
commits to deliver the item
in case pays for it.
will commit to deliver the item if pays.
Table 2: Composed patterns with varying degrees of strength.

Thirdly, with respect to the explanatory component, the cognitive consistent pattern is used by agents. Such an agent is committed to itself to bring about the items that he promised to perform. It comes with several flavors: An agent would try to bring about if he promised to do so, no matter who its partner is and no matter if its partner has already performed or not the requested action: . Some agents can manifest sincerity only to some partner : . Here, if the agent is committed to the particular agent than he is committed to itself to bring about . Re-assurance can be provided to its partner: “If I have promised you to bring about I will.”, given by or to a supervisor: “I am committed to my boss to keep my promises to all agents which do not have to perform something in exchange.”, formalized as .

Quite differently, for a agent , even he has promised to bring about this is not conclusive for itself that will hold, given by . Depending on the justification provided for this sentence, it can be: i) the agent simply does not trust its capabilities to accomplish the task; ii) the agent is aware of some stronger commitment that may block the execution of ; or iii) he is not committed to itself to satisfy its own promises and it is aware of this.

5 Trip Booking Scenario

Commitment Store:
Figure 6: Commitment store in the running scenario.

Consider the scenario in which the agent wants to book a trip to Valencia from the tourism agency . The company has contracts for flights with the air company and for accommodation with the business entity (figure 6). The agent promises agent to deliver the trip after he pays for it, represented by the pattern . The capabilities of the agent are represented by the explanatory pattern in which, if he manages to book the flight and accommodation, he can deliver the requested trip for his client. The tourism agency is aware of the unilateral contract from the accommodation company , given by , and also of the bilateral contract issued by : . Here, the agent will provide plane tickets if the partner agent promises to pay the amount .

Given the location of the agent within the European Community, it is doing business under the jurisdiction of the corresponding normative agent . Following the contract law, the normative agent requests to all its business entities under its umbrella that an open offer accepted by a client should be honored, otherwise the breaching agent should return the money and also pay penalties. The composition between a unilateral contract and a gratuitous promise is used to model this:

Here, if the client pays the amount and he does not receive the trip, the tourism company is committed to pay the amount to the client.

Assume that the tourism agency is an internalist 1-type credulous agent with a convincing justification standard of proof. The client is externalist, cautious agent having the preponderance of evidence justification standard.

The agent commences the dialog by requesting justification guarantying the trip in case he pays, given by the pattern:

Being an internalist agent, the tourism company can guarantee the trip based only on its contractual clauses with the suppliers and :

We note the compound justifier above with , such that . The first term in is a UC, whilst the second commitment represents a BC. Both terms satisfying the proof standard of agent , the compound justifier will meet the justification standard of agent . Being an externalist agent, the client can accept the justification of the agent as valid:

Note that in this line of justification, the agent releases some private information like the values and .

Being a cautious agent, the client should check if there are explanations supporting the opposite conclusion . Firstly, it requests explanans why offers enough justification for the formula :

Using the explanation pattern, the tourism agency explains based on its business practice that plane tickets and accommodation are enough to provide the requested trip:

Consider that the agent is aware of the regulation requesting visa for non EU citizens, formalized as . This is the main concern of the agent for rebutting the formula.

We note the above justification chain with such that . At this moment the tourism agency may request explanations to agent regarding his concern:

The agent can provide the explanation that he is not an EU citizen. This helps agent to figure out the case, but being a 1-type cautious agent it needs a justification to accept the formula . Such a justification is requested through:

One option would be to object the request by activating the critical question of the pattern, considering that the agent is not legitimate to ask for such justifications. The other option is to provide an ID card or passport acting as a justification: . A fact being stronger that a commitment according to proposition 2, it satisfies the preponderance of justification standard of the agent. Note that it may represent a constant in the justificatory logic framework.

Knowing that Swiss citizens do not need a visa for traveling in Europe, the commitment undercuts the justifier supporting formula:

We note the above justifier with such that . commitment refers to the agent too, which is under the normative framework of the agent . Therefore, it can be used by the internalist agent to constructs justifications. Being cautious, the agent needs a supplementary explanation, requested with:

The EU preferring to encourage traveling across Europe from safe countries, rather then imposing unnecessary security constraints, explains why a visa is not required:

Having both a justification and an explanation, the agent accepts that his visa concerns are defeated. Consequently, given the accepted valid justification for the trip, and no valid explanation for , the agent accepts the formula.

6 Discussion and Related Work

Justification Logic and Argumentation.

In the proposed approach we apply the justification logic mechanism to an argumentation context. The connection between explanation, justification, and argumentation is best evidenced in scenarios where understanding is the focus (eg. learning). One reason is that the constructivist approach needed in understanding is employed by justification logic. Justification Logic treats justification as the primary object, whilst the claim is secondary. Quite the opposite, nonjustificational criticism works towards attacking claims themselves, which is closer to classical approach in the argumentation theory. The proof checker operator of JL promotes reflective communication or deliberation on justificatory and explanatory knowledge.

The constants in JL can be mapped to prima facie arguments, which do not require any sort of justification, whilst standard arguments should be supported by justification chains. The framework of assures that all arguments are supported through a chain starting from prima facie arguments only.

The proposed framework applies the JL formalism to a multi-agent system. Differently from existing approaches [21, 14], we introduce the explanatory operator and force the justifiers to be social commitments. Particular to the classical approaches in JL [7], the explanatory operator introduces non-monotonic reasoning in the JL framework. One of the advantages that commitments bring in this justification logic landscape, is that there are more possibilities to define communicative acts between agents: like requests or promises between conveying agent and receiving agent .

Argumentation Dialogs.

In the dialog typology of Walton and Krabbe [19], information seeking and inquiry dialogs implies a search for a true answer to some factual question [5]. In this respect, justification and explanation are subtypes of such dialogs. In the Habermas discourse theory [9], the classification includes explicative, theoretical, and moral discourses, where explicative discourses aim at increasing comprehensibility, whilst theoretical discourses try to discover the truth.

A first step towards integrated argumentation reasoning with explanatory reasoning is made in [4], where the distinction between argumentation and explanation does not come from the statical structure or mode of reasoning (often abductive for explanation, mostly deductive in argumentation), but rather from broader dialogical context. Justifying reasons appear in two classes: epistemic reasons and practical reasons, treated as arguments [1]. Epistemic justifications (or theoretical reasons) are based only on beliefs and are used to justify beliefs. Practical justifications are constructed from beliefs and preferences and used to justify options or actions.

Compared to argumentation schemes, which aim to capture domain independent argumentation patterns, our proposal is more business oriented. We try to construct basic argumentation blocks used in negotiation and business process monitoring. This represents a step towards developing an argumentation pattern language directly applied on business process modeling languages.

Based on the content of explanation, rule-based systems exploit four types of explanations: (i)

, which shows the line of reasoning supporting the conclusion, (ii) , describing the rationale behind each reasoning step, (iii) or explanations, which bears out the problem solving strategy or control behavior of the system, and (iv) , providing definitional information [23]. Justification-type explanations seems to give rise to more positive user perceptions of a knowledge-based system than trace and strategic explanations [22]. Many knowledge-based systems address the issue of presenting user-adapted explanations [23].

Explanation types depend on domain knowledge. Within the social domain, folk psychological explanations are conveyed, in the physical domain both naive and sophisticated theories, whilst in the religious domain explanatory standards prevail. In our case, the main goal of seeking explanation in the business domain is to increase the ability to make predictions and to understand normative-based decisions.

Social Commitments.

Explanation and justification are treated here as social constructs. Commitments have been used within an argumentation framework in [3], where argumentation relations as defend, attack, or justify are defined between two commitments. In our case, the strength relation is inferred based on the commitment pattern, whilst the preference relation depends on the agents role in the commitment, as creditors or debtors. In both approaches, the agents should be able to justify their commitments.

The expressivity arising from combining commitments have been exploited for representing business patterns [6, 18] or legal contracts [13]. The large amount of work on commitments deals with modeling business patterns, which has been proved satisfactory for real life business protocols. Often, a dialog occurs in which the creditor requests for justifiers or performance, whilst the debtor provides the requested justification or explanation in order to re-activate the commitment or to find an alternative solution.

Defining commitment-based protocols is easier due to the closed word assumptions when designing a protocol. This is not the case in argumentation where agents may convey in an open world different types of justifications and explanations. That is why an argumentation process is needed on top of the commitments in order to assure the flexibility encountered in real life business interactions.

7 Conclusions

The main goal of our study has been to propose a technical instrumentation for handling both justificatory and explanatory arguments. The first contribution regards the setting of some basis of exploiting in computational models of arguments the differences between justification and explanation, as already stressed out in the philosophy of science.

As a second contribution, the is developed to cover both justifiers and explanans, in the line of using logic in argumentation as envisaged by Gabbay [8]. Introducing commitments enhances the capability of agents to reason over justifications and explanations of the other agents.

A third contribution regards the formalization of justificatory and explanatory commitment-based patterns. The individual actions are taking place within a framework of interdependent social commitments. We consider agents as placed in various networks of commitment relations with other agents, where social influences formulate the justification behind agents decisions [11]. The main benefit here regards the flexibility to construct a large variety of higher order patterns.

Using together with commitments provides opportunities for defining several types of argumentative agents. Activating specific axioms in , rigorous or cautios agents can be formalized. The justification standard of each agent is defined based on the strength relation between compound commitments.

As future work we will be investigating the role of critical questions in blocking justifications and explanations.


We are grateful to the anonymous reviewers for their useful comments. Adrian Groza is supported by the Sectoral Operational Programme Human Resources Development 2007-2013 of the Romanian Ministry of Labour, Family and Social Protection through the Financial Agreement POSDRU/89/1.5/S/62557.


  • [1] Amgoud, L., Prade, H.: Using arguments for making and explaining decisions. Artif. Intell. 173(3–4), 413–436 (2009)
  • [2] Artemov, S.N.: Justification logic. In: Hölldobler, S., Lutz, C., Wansing, H. (eds.) JELIA. Lecture Notes in Computer Science, vol. 5293, pp. 1–4. Springer (2008)
  • [3] Bentahar, J., Moulin, B., Meyer, J.J.C., Chaib-draa, B.: A logical model for commitment and argument network for agent communication. In: AAMAS. pp. 792–799 (2004)
  • [4] Bex, F., Budzynska, K.: Argumentation and explanation in the context of dialogue. In: CMNA. pp. 1–4 (2010)
  • [5] Black, E., Hunter, A.: An inquiry dialogue system. Autonomous Agents and Multi-Agent Systems 19(2), 173–209 (2009)
  • [6] Chopra, A.K., Singh, M.P.: Specifying and applying commitment-based business patterns. In: AAMAS. pp. 475–482 (2011)
  • [7] Fitting, M.: Reasoning with justifications. In: Makinson, D., Malinowski, J., Wansing, H. (eds.) Towards Mathematical Philosophy, Papers from the Studia Logica conference Trends in Logic IV, vol. 28, pp. 107–123. Springer (2009)
  • [8] Gabbay, D.: Sampling logic and argumentation networks: A manifesto. JICPR special issue on Logic and Philosophy Today To appear
  • [9] Habermas, J.: The Theory of Cummunicative Action. Boston, Beacon Press (1981)
  • [10] Jennings, N.R., Parsons, S., Noriega, P., Sierra, C.: On argumentation-based negotiation. In: Intern. Workshop on Multi-Agent Systems, Boston. pp. 1–7 (1998)
  • [11] Karunatillake, N.C., Jennings, N.R., Rahwan, I., Norman, T.J.: Argument-based negotiation in a social context. In: AAMAS. pp. 1331–1332 (2005)
  • [12] Lenman, J.: Reasons for action: Justification vs. explanation. In: Zalta, E.N. (ed.) The Stanford Encyclopedia of Philosophy. Winter 2011 edn. (2011)
  • [13] Letia, I.A., Groza, A.: Running contracts with defeasible commitment. In: IEA/AIE. pp. 91–100 (2006)
  • [14] Letia, I.A., Groza, A.: Arguing with justification between collaborative agents. In: ArgMAS, Taipei, Taiwan (2011)
  • [15] Little, D.: Varietis of social explanations: an introduction to the philosophy of social science. WestviewPress (1991)
  • [16] Lu, J., Chiu, M.M., Law, N.W.: Collaborative argumentation and justifications: A statistical discourse analysis of online discussions. Computers in Human Behavior 27(2), 946–955 (2011)
  • [17] Morgan, J., Olsen, W.: Conceptual issues in institutional economics: clarifying the fluidity of rules. Journal of Institutional Economics 7, 425–454 (2011)
  • [18] Telang, P.R., Singh, M.P.: Specifying and verifying cross-organizational business models: An agent-oriented approach. IEEE Transactions on Services Computing (2012), to appear
  • [19] Walton, D., Krabbe, E.C.W.: Commitment in Dialogue: Basic Concepts of Interpersonal Reasoning. State University of New York Press (1995)
  • [20] Yan, J., Ye, K., Wang, H., Hua, Z.: Ontology of collaborative manufacturing: Alignment of service-oriented framework with service-dominant logic. Expert Systems and Applications 37, 2222–2231 (2010)
  • [21]

    Yavorskaya, T.: Interacting explicit evidence systems. Theory of Computer Systems 43, 272–293 (2008)

  • [22] Ye, L., Johnson, P.: The impact of explanation facilities on user acceptance of expert systems advice. MIS Quarterly 19(2), 157–172 (1995)
  • [23] Yetim, F.: A framework for organizing justifications for strategic use in adaptive interaction contexts. In: ECIS. pp. 815–825 (2008)