«Trust in Electronic Commerce - a Language Action Perspective Hans Weigand, Willem-Jan van den Heuvel Infolab, Tilburg University P.O.Box 90153, 5000 ...»
2.4 Communicative and Strategic Action In this section, we investigate a taxonomy that has been developed by Habermas in his theory of communicative action Habermas, 1984.
Communicative action is de ned by Habermas as "action toward understanding Verstandigung as a way of coordinating the actions of the participants". Coordination is achieved because the participants agree on certain beliefs and norms. Because of this agreement, a request for action for example is e ective when it is validated by the shared knowledge. Or, to put it informally, the hearer accepts the request because he agrees with the speaker that this is the right thing to do.
The result of a communicative action is dependent on the acceptance or rejection of a validity claim. Habermas presents three di erent kinds of validity claims: truth, justice and sincerity. These claims refer to subsequently the objective, social and intersubjective world as we have seen these in section 2.2. The claim of truth refers to facts in the object world. The claim of justice regards the adequacy of interpersonal relations, and the claim of sincerity entails the intentions of the speaker. These claims can be challenged; the agreement is the theoretical endpoint of the discussions that can arise when a party raises a claim.
Besides communicative action, Habermas distinguishes strategic action. Communicative action is oriented towards mutual agreement, whereas strategic action is aimed at attaining individual goals. When acting strategically, parties interact because they think this serves their own interests. So a hearer accepts a certain request because he thinks this will help him in achieving some goal of his own this can be a sel sh but also an altruistic goal.
According to Habermas, communicative action and strategic action cannot be reduced to one another. Moreover, he claims that it is not strategic action but communicative action that achieves coordination in our lifeworld.
That opportunistic action, as presupposed in neoclassical economics, can lead to ine cient or even absurd results is nicely illustrated in a famous story told by the Nobel prize winner Amartya Sen. Two strangers meet. "Where is the railway station? he asks me. "There", I say, pointing at the post o ce, "and would you please post this letter for me on the way?". "Yes", he says, determined to open the envelope and check whether it contains something valuable.
If we looker deeper in the taxonomy g. 1, we see that Habermas makes a di erence between two types of strategic actions: concealed strategic action and open strategic action.
Habermas speaks of open strategic action whenever the strategic character of the interaction between the speaker and hearer is clear. In the case of concealed strategic action, the speaker makes the hearer believe that all the presumptions for communicative action are satis ed.
If the speaker is deceiving himself, Habermas speaks of unconscious deception. Otherwise, the speaker is willingly oriented towards his own success, and tries to manipulate the hearer.
Instead of opposing strategic action and communicative action, we propose to view them as two di erent levels of social behaviour. Strategic action refers then to the economic level at which agents are viewed as individuals maximizing some utility function, or, put di erently, as pursuing certain goals, sel sh or not. They pursue these goals by nding subgoals and means that contribute to the satisfaction of the goals. In commerce, these goals are typically related to pro ts of the company. Communicative action resides at the social level at which subjects interact by means of language section 2.1. In the same way as communicative illocutionary acts are performed by means of utterance acts signs, strategic goals are pursued by means of communicative acts. Deception occurs when the strategic goal expressed is not the real one. Assuming that goals are part of the subject world of the communicative agent, deception is a case where the sincerity conditions are not met. On the other hand, open strategic action occurs when the strategic goal is explicit. In that case, there is no deception and no insincerity.
It may be objected that our proposal does not do justice to Habermas' extensive argument that coordination by means of strategic action is fundamentally di erent from coordination by means of communicative action. However, we still folllow Habermas in the claim that communicative action cannot be reduced to strategic action and, more precisely, that the coordination is e ectuated by Verstandigung rather than by the two actors strategic behaviour. Although the need for coordination may arise from the individual goals but not necessarily so, the coordination itself is achieved by communicative acts that have a di erent orientation.
Habermas also makes a theoretical argument that ontologically speaking, strategic action is grounded in communicative action. The thrust of his work is not that system worlds focused on strategic action such as "the market" should be abolished, but that these system worlds should not colonize the lifeworlds. We interpret this in the present context that electronic commerce should not be approached from a strategic action point of view only, but its concrete applications should support ways of achieving mutual understanding Weigand and Dignum, 1997. These channels may be used as back-up facility only when the system breaks down, but can also serve to support trust-building in the long run. An example is the possibility of discussing and agreeing on certain rules of conduct. We will Social Actions
come back on some concrete examples in section 4.2. Interestingly, it seems to be the case in current paper-based international trade that most arrangements are made informally by telephone and fax, and the papers only follow afterwards. If this is true, we should be careful when replacing the paper-based procedure by electronic means not to ignore this informal part.
2.5 Trust, fraud and deception In this subsection, we will take a closer look at the relation between trust, fraud and deception, in order to come to a language action-based de nition of trust.
We say that an action is fraudulous whenever a subject tries to deceive another group of subjects in order to gain an wrongful pro t out of it, be witty at somebody else. Hence, the major element of fraud is deception. In practice, deception often occurs by manipulating documents in such a way that somebody believes she has to do something Bons, 1997.
Deception and fraud involve communication. Parties can cause various kinds of harm, but deception can only be achieved by means of signs. Hence prevention of deception and fraud is closely related with control measures at the level of communication Trust has been de ned as `a particular level of the subjective probability with which an agent assesses that another agent or group of agents will perform a particular action, both before he can monitor such action independently of his capacity ever to be able to monitor it and in a context in which it a ects his own action" Gambetta, 1985. Kini and Choobineh have de ned the concept of trust in the context of Electronic Commerce systems as `an individual belief in the competence, dependability, and security of the system under conditions of risk' Kini and Choobineh, 1998. In terms of risk analysis, the hazard is the existence of mala de parties, the peril is the occurrence of the deception and the exposure are the objects or funds being lost to such parties. All three components can be taken as a basis for prevention. The hazard can be reduced by dealing with trusted parties only which of course prompts the question how trust can be enabled; the peril can be reduced by preventing deception e.g. by using secure communication means, and the exposure can be reduced by e.g. insurances, outsourcing, but also by the back-up availability of legal court.
This analysis also gives an indication of the relationship between trust and deception.
We have two additional remarks on the above de nitions of trust. The rst is that they describe trust as a cognitive disposition without taking into account where the trust is built on. In the context of Electronic Commerce, the existence of trust or distrust is less interesting than the question how trust is enabled. This is what we would call the objective dimension of trust.
The second remark is that the de nitions ignore the social dimension. Trust is also a relationship. People put trust in the other party, or in the agent that they hire, and in the case of nonconformance, the trust is violated, and the relationship is jeopardized. This is what we would call the social dimension of trust. An example in Electronic Commerce relating to this dimension of trust is the value of brand names, but also the whole issue of virtual organizations see below.
We would like to summarize this discussion in a de nition of trust from an L A perspective. Let us rst de ne a communicative agent as an agent that takes responsibility for the communicative actions he is engaged in. As a Speaker, he takes responsibility for what he says in terms of truth, sincerity and justice. This includes that he keeps up his promises.
As an Addressee, he takes responsibility for what he hears. This includes con dentiality and an interpretation of the contents of the message in accordance with the shared beliefs and values. We can now de ne trust to be the level of grounded expectation that the other agent is acting communicatively. Note that this relationship may be symmetrical, but not necessarily so.
This de nition recaptures several important elements from the de nitions given above in an integrated form. For instance, it takes over the social dimension because of its link to communicative actions between agents, the objective dimension by the word "grounded" and the subjective dimension "expectation". In some sense, it is wider that the de nition of Gambetta since it is not necessarily oriented towards the future, and in some sense it is more strict in that it does not apply in situations in which another agent may a ect my behavior without there having been any communication between us.
To account for the derivative use of the word "trust" to inanimate objects, we can take over the de nition of Kini and Choobineh. This notion of trust is supposedly metaphorically derived from the primary one, but they cannot be merged: objects are not communicative agents, nor can agents be reduced to objects. Note that in the context of Electronic Commerce, both notions of trust play and role and should be sharply distinguished: trust in the other party and trust in the system the Internet, the rewall etc. For Electronic Commerce to be adopted, both forms of trust must be present.
3 Communication and Security In the preceding sections, we have sketched a Language Action based framework of communication. In this section, we show what this means for securing trustworthy electronic commerce.
In our framework, electronic commerce is doing business by means of speech acts. Security measures can and should therefore apply both to the means and the ends, that is, both the utterance acts the physical level and the illocutionary acts the essential level. These two levels correspond with what Bons calls the communication level and the business level, respectively Bons, 1997. These two levels of security are implementation-independent.
The communication security draws on the lower level, the channel level. This level serves to secure the transport medium for the messages. The channel level is implementation-speci c, and can be realized by rewalls and encryption techniques. We talk about levels rather than aspects because each level is enabling for the next-higher level.
Before going on to describe the possible measures at these levels, it is important to keep in mind what the purpose of these measures is. Parties want to execute communicative actions, that is, achieve e ects in the social worlds, but they can only do that by making another act in which they represent the action. The utterance act makes use of a medium, either the medium of oral speech or the medium of paper-based text, or some digital medium.
There will always be a medium, but the quality of di erent media can di er.
In addition to the levels distinguished by Bons, we include one highest level, the strategic level, corresponding to the strategic goals that subjects achieve by communicative acts.