1. Introduction
It is well-known that in many a legal dispute the question arises what the exact extension of a predicate is. The difference of opinion in such cases almost always concerns the question as to whether an incident comes under the reach of a concept that is expressed by a particular word or phrase in a legal text in which the rights and obligations of the persons holding legal rights are established (for example a law or agreement). In such cases of difference of opinion the lawyers are forced to declare what a certain word or group of words means in their opinion. And in the discussions that may be carried out they often also give definitions of the words or phrases concerned and will, in principle, have to justify the acceptability of such definitions.
The question now is: how do lawyers – and more particularly judges – deal with this kind of language controversy; what kind of definitions do they give and how do they present and justify them? I attempt in this article to give an interim answer – an interim answer due among other things to the insufficiency of the systematic research I have done into the judgements of judges in The Netherlands.
The article is set up as follows. In paragraph 2 a case is given in rough outline and in paragraph 3 there is the development of part of the legal discussion as a result of that case. In paragraph 4 I go into the question of which types of definition can be distinguished and how the plausibility of each of these different types of definition can be argued. In paragraph 5 I reconstruct part of the legal discussion in the light of the typology of definitions dealt with in paragraph 5. Paragraph 6 constitutes the conclusion of this article.

2. A case: fire in a building[i]
Mr. Matthes owned a house of nine rooms. In 1979 the house was inhabited by Matthes with his wife and four children and also by a tenant and her son. All the rooms were in use by Matthes and the members of his family, except for one room on the first floor which was used by the tenant.
Matthes wanted to take out fire insurance with the Noordhollandse insurance company and submitted an application form for this purpose for an ‘index/extended insurance for private house’. On the reverse of the form it stated:
1. the applicant declares: a. that the private house on which or in which insurance is requested, is of brick/concrete with a hard roofing, with no business or storage and without increased danger to adjoining properties’.

From 17 July 1979 the Noordhollandse insurance company insured the house for the period until 17 July 1989 including fire risk. The policy for extended building insurance dated 2 August 1979 referred to the house with the addition:
2.  ‘serving solely as private house’.

On Monday 3 December 1984 at about eight-thirty p.m. fire broke out in the house resulting in considerable damage. At that time the house was inhabited by Matthes and his wife and a total of five rooms were rented out to three different single gentlemen. Naturally Matthes claimed on the insurance company for the damage which amounted to some 500,000 Dutch guilders. However the company refused payment on the grounds of the insurance since in its opinion the premises insured no longer served as a private house but was used as a room rental business for which during the insured period the use of the insured object was altered, whereas Matthes had not informed the insurance company of the fact. The Noordhollandse appealed to article 293 of the Commercial Code of The Netherlands:
3.  ‘If an insured building is given a different use and is thereby exposed to increased danger, so that the insurer, if such had been in existence before the insurance was given, would not have insured the same at all or not on the same conditions, this obligation is terminated.’

Naturally Mattes did not agree with this and went to court. However the lawcourt, the court of justice and the Supreme Court successively declared him to be in the wrong.

3. The course of the legal discussion in this case
The legal discussion for the various authorities concerns to a large degree the question of what meaning should be given to the word ‘private house’ on the application form for the fire insurance and the phrase ‘acting solely as private house’ in the insurance policy. The lawcourt was of the opinion that the word ‘private house’ should have the following meaning:
4.  ‘a house that serves as a general rule for the permanent accommodation of several persons who are partially dependent of each other economically and furthermore have an emotional bond with each other.’

This means according to the lawcourt in general:
5. ‘that such persons have a greater concern for each other and each others interests than random otherwise respectable citizens may be expected to have and that the social control of their doings is greater than that normally found among the same citizens.’ The situation on 3 December 1984 was, according to the lawcourt, other than that in 1979, since:
6.  ‘the private house was occupied on 3 December 1984 for the greater part by tenants who would require more privacy and whose behaviour was subject to less social control.’

According to the lawcourt this meant that a change of use in the meaning of art. 293 of the commercial code of The Netherlands took place whereby the house was subject to ‘increased fire risk’. Matthes was thus put in the wrong and appealed.
He declared among other things to the court of justice that:
7.  ‘a building destined as ‘private house’ should retain this designation irrespective of whether it is occupied by the insured and his family or by the insured with a number of tenants.’

In short Matthes employed another definition of ‘private house’, namely:
8. ‘a building destined mainly for residential purposes.’

In view of this definition of ‘private house’ there is no question of a difference in destination in the light of the policy, since at the time of the fire Matthes lived in the house with his wife and three house-mates/tenants who did not form part of the family. Matthes contended further that:
9.  ‘the manner in which the term ‘private house’ was interpreted by me was perfectly in keeping with the normal use of language, in view of the fact that the description ‘private house’ is the most obvious and was employed for the insured object as it was used during the fire.’

The court of justice refuted the plea of Matthes, supporting the rejection by yet another definition of ‘private house’. It stated:
10.  ‘the term “private house” on the application form and the words “serving solely as private house” in the relevant policy are to be understood as “private house serving mainly as private dwelling for the insured whether or not with his family”.’

The court of justice then considered that:
11. ‘now that the insured building was inhabited by the Matthes family on taking out the insurance, consisting of husband, wife and four children, together with a tenant with one child, and that when the fire broke out it was occupied by Mr. and Mrs. Matthes with three tenants, there was a question of an actual alteration of usage.

This was all the more convincing now that according to Matthes’ own declaration the rooms concerned were rented out so that the revenue could contribute to the university expenses of his children, the which implied that rental of the accommodation could not be said to lack a certain business nature.’ And further:
12.  ‘that private house as understood by the court should not be taken to mean a building of which, as in the present case, more that half the rooms are let to third parties, and that such building rather had the nature of an accommodation business for the insurance of which a different premium or conditions applied than to the insurance of a dwelling, the which was not contested by Matthes.’ And:
13.  ‘the court of justice regarded as obvious the fact that a building of which the owner-occupier had at his disposal three rooms and a guest-room and of which the other five rooms had been let to third parties which in principle were independent of each other and had no reason to occupy themselves with the affairs of their fellow residents, even if they referred to themselves as a community, was exposed to a greater danger of fire than when this building was occupied by a family with children and a single tenant.’

At the court of justice Matthes thus was again said to be in the wrong and determined to appeal to the supreme court.

As plaintiff in appeal he declared essentially the same as before the court of justice, namely that based on the most usual definition of the term ‘private house’ there was no question of a change of use. In his summing up Solicitor General Asser also explored the definition of private house as given by the court and stated the following:
14.  ‘The meaning given by the court to the concept “private house” seems to me, where there is talk of “private occupation by
the insured whether or not together with his family” hardly obvious in the first instance in the light of the proposition of the parties. I have not come across this very narrow interpretation of the concept “private house” anywhere, more particularly not in the propositions of the Noordhollandse. On the contrary, the Noordhollandse has stated in the memorandum of reply in appeal that in general speech a private house is considered to be a house occupied by a family, it being of no consequence whether the house is owner-occupied or rented by the occupiers. There should thus not be in the policy any clause stating a home “solely serving for own occupation”, according to the Noordhollandse. The Noordhollandse did state that the situation was different when there was a question of more independent tenants and more particularly an accommodation business, of which according to the Noordhollandse there was a question in this case. In this connection I would also wish to assume that what the court considered should be read thus that “private house” is taken to mean occupied mainly by a person alone or as a family, whereby the intention is other than occupation by tenants. The explanation of the court thus amounted to what the lawcourt considered in somewhat elaborate terms.’

Finally the Solicitor General advised the rejection of the appeal made by Matthes. The Supreme Court took this advice, considering more particularly the following:
15  ‘Against this background judicial consideration 4.4 is apparently to be so understood that Matthes, in the opinion of the court, could reasonably have understood from the term “private house”, or the words “serving solely as private house” – and that the Noordhollandse could reasonably expect that it should be clear to Matthes – that the use thus described included the situation of the insured who occupied the largest part of the building himself (whether or not together with his family), “a single tenant” was present in the building, but not the situation in which as in the present case, the larger part of the building, namely more than half the rooms, was let to third parties, in which case the building, as the court stated “had rather the nature of an accommodation business”.’

Law professor Van der Grinten in his note following the judgement criticises this pronouncement:
16.  ‘Has the court rightly assumed that the words “serving solely as private house” are to be interpreted as “dwelling serving for the private accommodation of the insured”? (…) I would be inclined to judge this differently than the court. The words “as private house” could be interpreted as “accommodation”. The circumstance that an important part of the house was later – after taking out insurance – used by the tenants as residence does not involve any alteration in the use.’

At first sight this discussion is rather unsatisfactory. More particularly it is not clear on what the lawcourt and the court of justice each based their own definition of the term ‘private house’ and neither do either of the bodies go into the argument of Matthes that his definition of ‘private house’ fits in most closely with normal speech. Due to this fact the discussion has all the characteristics of a yes-no discussion but nevertheless one with considerable financial consequences. This naturally gives rise to the theoretical legal question of how free the judge is in giving meaning to non-legal terms in the explanation of written agreements and to what extent he can be required to motivate his definitions.

In short, this discussion – and more particularly the judgement of the court of justice – demands rational reconstruction. But this is only possible when we evolve a theory about definitions.

4. A pragmatic-dialectic approach to defining
The theory about definition and the theory about argumentation are closely related, as Viskil showed so convincingly (see Viskil 1994a, 1994b, and 1995). Definition is regarded as an important instrument in interpretation, assessment and formulation of points of view and arguments. According to the classic view, a definition is a statement concerning the essence of a thing. In modern theories with a perspective of dialogue on argumentation, a definition is considered in the first instance to be an instrument to clarify discussions. It is necessary for partners in discussion to clarify their terms, since not only the soundness of arguments but also the acceptability of, for example, standpoints are only to be realised if the meaning of the terms is clear.

Viskil proposes considering definition as a speech act and in view of that fact he arrives at the typology of defining speech acts and thus corresponding definitions, to which the following three also belong:
17.
a. Stipulative definition
b. Lexical definition
c. Stipulative lexical definition.
The act of stipulative definining is a section of the class of language usage declaratives, a subclass of declaratives. Stipulative defining is bound to felicity conditions (18) and (19) (see Viskil 1994a: 144 et seq.).

18.  Essential condition for stipulative defining
Performing speech act T counts as establishing the meaning of a word (or phrase) in order to clarify this meaning for the listner or reader.

19.  Propositional content condition of stipulative defining
Each proposition which is expressed in a sentence of which the subject term is formed by a quoted (group of) word(s) and the predicate exists (1) of a verb that indicates that the remaining portion of the predicate is the meaning of the subject term and (2) one or more words or groups of words with or without modifier.

Examples which meet the propositional content condition are the following.

20.
a. The word bungalow means ‘a house where all the rooms are on the same level’ (= connotative stipulative defining).
b. Inventiveness means ‘resourcefulness’ (= stipulative defining by means of the giving of a synonym).
c. I take breaker’s yard to mean: junkyard, centre for used car parts, wrecker’s yard and car damage businesses (denotative stipulative defining).

The act of lexical defining is a section of the class of language usage assertives, a subclass of assertives. This speech act is bound to the essential felicity condition (21) (see Viskil 1994a, 153 et seq.)

21. Essential condition of lexical defining
Performing speech act T counts as a description of the meaning in which language users use a word (or phrase) in order to clarify this meaning to the listner or reader.

The propositional content condition of lexical defining is identical to that of stipulative defining. The speech acts are thus identical with respect to content, but they differ in the illoctutionary purpose, which is also noticeable in the essential condition (but also of course in the preparatory condition and the sincerity condition). For that reason the examples given in (20) could also be examples of lexical definitions.

Some definitions are not purely stipulative or purely lexical, but partly stipulative and partly lexical. In the simplest mixture of these two speech acts the speaker or writer attempts to clarify a word by a description of the meaning of such word which is valid as an establishment of the meaning. This speech act can be indicated by the term ‘stipulative-lexical defining’. There are at least two sub-types. First there is the case where the speaker or writer defines a term in the conventional way while declaring at the same time that in using that term he will also keep to that meaning, see example (22).

22.  The word chair usually means a seating unit for one person and I shall be using it further in that sense.

In the second place there is the case in which the speaker or writer gives a specification of the lexical definition and declares that he will use the term in the meaning of the specification given, see example (23).

23. The word chair usually means a seating unit for one person, but I use this term in the sense of a seating unit for one person and provided with four legs.

In both cases the speaker or writer commits himself to a conventional meaning (the lexical aspect of the definition) but at the same time calls up a situation within which the defined term is used in conformity with the meaning, whether or not specified (the stipulative aspect). The class to which the speech act of ‘stipulative lexical defining’ is to be reckoned is that of the language usage declaratives. But otherwise than in the case of stipulative defining, stipulative-lexical defining is no ordinary language usage declarative, but a combination of a language usage declarative and an assertive. The conditions of success of the speech act ‘stipulative-lexical defining’” then combines the felicity conditions of stipulative defining with that of lexical defining (see Viskil 1994a: 156 et seq.)

24. Essential condition of stipulative-lexical defining
Performing speech act T counts as a description of the meaning in which language users use a word (or phrase) which has the force of establishing this meaning for the language usage of the speaker or writer, in order to clarify the meaning for the listner or reader.

Naturally the propositional content condition for stipulative-lexical defining is also equal to those of stipulative definition.[ii] That is to say that the sentences under (20) may also count as examples of stipulative-lexical defining.
Viskil also pays attention in his approach to the question of how definitions can be justified, for which purpose he makes use of the pragmatic-dialectic argumentation theory.[iii] The justification of a definition is based on the fact that the definition should solve the problems for which it is drawn up and is acceptable to the definer as well as to the persons for whom it is intended. The definer justifies his definition to convince the listener or reader of the acceptability of his definition and thus obtains inter-subjective agreement regarding the definition.

A stipulative definition should be an adequate attempt at clarification and be functional. A lexical definition should be an adequate attempt at clarification and contain a true proposition: the meaning that is described in a lexical definition should concur with the meaning in which the language users in question use the defined word. In order to be acceptable a stipulative-lexical definition should answer to three demands: the definition has to be an acceptable attempt at clarification, be functional and give a description of the meaning that agrees with the facts. The standard argumentation structure for the defence of the acceptability of a stipulative-lexical definition, when seen as described, appears to be as follows (see Viskil 1994a: 253).

5. A rational reconstruction of part of the legal discussion of the case
A rational reconstruction of an argumentative discussion or a part thereof is a reformulation of that discussion or of such part of it with a view to the testing of its rationality. Such a reconstruction always assumes of course a theoretical perspective from where is reconstructed. Let us now look at our legal discussion through the spectacles of the theory sketched above regarding definition. We are now able to pose the following two questions: (a) of what type are the definitions which play a part in this discussion and (b) are the definitions given – dependent on their type – adequately justified?
Question (a) is of course not solely to be answered by regarding the form of the sentences in which the definitions are formulated.  After all, we have seen that the three types of defining speech acts should be distinguished based on their illocutionary force, expressed also in the different essential conditions. If we base ourselves on the illocutionary purport we have to refer for the reply to question (a) to the difference between the legal bodies and the other participants of this discussion. The following legal rule is here important:
26. Should there be a difference of opinion between the parties concerning the explanation of a term in a written agreement, the judge of the facts of the case is at liberty to explain the term concerned independently, quite apart from what the parties advance in this connection.[iv]

In other words: in the matter of the case dealt with here the lawcourt and the court of justice were at liberty to give an independent meaning to the term ‘private house’, without having to take into account what Matthes and the Noordhollandse had advanced in that case. This explains, in my opinion, why neither the lawcourt nor the court of justice went into the argument advanced by Matthes that his definition of the term ‘private house’ linked up more closely with the normal use of language.

Rule (26) indicates further that defining speech acts which are advanced by judges in the context of the explanation of agreements, should be regarded in any case as being of a stipulative nature. After all, the definition by the judges of the term ‘private house’ cannot be regarded as other than an establishment of the meaning which is aimed at making matters clear(er) to the listener or reader. The question is however whether there can be any question of a purely stipulative definition. This amounts to the question of  whether the judge is also at liberty to explain terms in an agreement – and certainly non-legal terms – entirely free of normal use. In my opinion the judge does not enjoy such liberty. After all, if we assume that for the explanation of agreements it is a directive what the parties should have understood by it and what they were to expect of each other, this cannot be taken apart from the conventional meaning of terms which are used in a linguistic community. This leads to the fact that definitions that are given by judges in similar circumstances, bear the nature of stipulative-lexical definitions.

If we assume that the judge of the facts advances stipulative-lexical definitions in this context, we can also ask ourselves the nature of the sub-type of the given definition of ‘private house’. It seems to me that we are here confronted with a specifying stipulative-lexical definition in the sense that the judge gives a specification of the daily term ‘private house’, as found, for instance, in Van Dale (the Dutch authoritative dictionary).

27.  Van Dale – Groot woordenboek der Nederlandse taal, (‘Van Dale – Large dictionary of the Dutch language’), 11th edition
private house (n), house, arranged as dwelling or where a person lives, as against office, shop (…);

The parties in the trial took a different position in this discussion. They will more particularly have to make clear to the judge what they were to expect of each other in the context of the agreement. It is therefore clear that they would make a claim in particular on the conventional meaning and thus advance definitions that were especially lexical. After all, as far as they are concerned it means especially giving a description of the meaning which language users within a certain language community give to a particular word or group of words. In our discussion this applies both to Matthes (see the verdicts (7), (8) and (9) above) as for the Noordhollandse, as far as this can be concluded from what Solicitor-General Asser said about it (see (14) above).
Once we have ascertained with what kind of definitions we are confronted in the discussion, we can also check whether the definitions are justified adequately (question (b)). If we assume, for example, that the court of justice gave a stipulative lexical definition of the term ‘private house’, then for the reconstruction of the account of this definition structure (25) should be taken as basis. It is now striking that in the plea of the court of justice no attention at all was paid to two of the three coordinative primary arguments in this standard structure: no single word is addressed either to (1a) nor to (1c). Attention is paid on the other hand to the question of whether the definition is functional. This part of the plea may be (partially)
reconstructed as follows.
In the scheme of this article it naturally does not concern the question of whether the definition given by the court of justice was adequate and whether the argumentation advanced was sound. The above is to illustrate more than anything that for a critical judgement of this type of discussion and argument a rational reconstruction is necessary in terms of a theory regarding definitions.

6. Conclusion
I assume for the time being that the discussion which I have given here is representative of those cases in which judges have to make a judgement on the meaning of non-legal words and groups of words when explaining written agreements.
It may be concluded that in this context judges give other types of definitions than the parties. Judges advance stipulative-lexical definitions whereas the parties make use of lexical definitions. In addition is can be stated that judges on the justification of the plausibility of the definitions they give do not pay any attention to arguments which are concerned with the question of whether the given definition of a word or group of words makes the meaning clear or clearer, neither do they answer the question whether the description of the meaning agrees with the facts, but merely go into the question of whether the definitions they provide are functional. Further research should indicate to what extent this picture is right and, if it is, to what extent this development has its origins in the specific nature of this kind of legal discussion.[v]

NOTES
i. See Supreme Court of The Netherlands 10 August 1988, NJ 1989, 238.
ii. See Van Haaften (1996) for treatment of the question of which types of definition generally arise in the context of legislation and judicial pronouncements.
iii. See F.H. van Eemeren & R. Grootendorst (1992) regarding the basic assumptions and approach of the pragmatic-dialectic argumentation theory.
iv. See also the final pleading of the Public Prosecutor for the Supreme Court of The Netherlands dated 6 February 1987, NJ 1987, 438, under 3.2 with further references.
v. It is perhaps good to notice that what I have said about definitions is by no means in contradiction with the now rather generally accepted idea of – as H.L.A. Hart calls it – ‘open texture’ of legal concepts and concepts in general, which means that it is in principle impossible to frame rules of language which are ready for all imaginable possibilities. That is to say that however complex our definitions may be, we cannot render them so precise as for them to be delimited in alle possible directions. It is thus not possible for any given case to say definitely that the concept either does or does not apply to it. As Hart (1983:275) puts it: ‘We can only redefine and refine our concepts to meet the new situations when they arise’. But of course all this does not mean – as sometimes people seem to conclude – that definitions are of no use at all.

REFERENCES
Eemeren, F.H. van & Grootendorst, R. (1992). Argumentation, Communication and Fallacies. A Pragma-Dialectical Perspective. Hillsdale: LEA Publishers.
Haaften, T. van (1996). Typen definities in de taal van het recht. Taalbeheersing 18, 255-269.
Hart, H.L.A. (1983). Jhering’s Heaven of Concepts and Modern Analytical Jurisprudence. In: H.L.A. Hart, Essays in Jurisprudence and Philosophy (pp. 265-277). Oxford: Clarendon Press.
Viskil, E. (1994a). Definiëren. Een bijdrage tot de theorievorming over het opstellen van definities. Amsterdam: IFOTT.
Viskil, E. (1994b). Definitions in argumentative texts. In: F.H. van Eemeren & R. Grootendorst (eds.), Studies in Pragma-Dialectics (pp. 79-86). Amsterdam: Sic Sat.
Viskil, E. (1995). Defending definitions, A pragma-dialectical approach. In: F.H. van Eemeren et al. (eds.), Perspectives and Approaches, volume 1 (pp. 428-438). Amsterdam: ISCA.

1. Introduction
Problem-solving discussions, conducted in all situations where people jointly have to solve problems and reach decisions, are an important part of public as well as private life. Since considerable interests are often at stake, it is important that these discussions be carried out in such a way as to ensure that the best possible decision is reached. In view of the importance of safeguarding the quality of problem-solving discussions, it is relevant to develop instruments for analyzing and evaluating such discussions. These instruments should make it possible to establish whether participants act in a fashion that is conducive to the goals of problemsolving discussions, and, if not, in what respects, at what points in the discussion, and in what ways. Such an analysis of the ways in which discussions can go wrong will yield a basis for teaching participants how to avoid these counterproductive practices in future.

In this paper, I will show that the ideal model of critical discussion, which is central to the pragma-dialectical approach to argumentative discourse developed by Van Eemeren and Grootendorst (1984, 1992), provides a diagnostic instrument which may be used in carrying out such an analysis. The model specifies the stages of critical discussion through which rational resolution of a difference of opinion is attained, and the speech acts which have to be performed in each of these stages. So far, the model has been applied mainly as an heuristic and analytical instrument for the dialectical reconstruction of discursive texts (Van Eemeren et al. 1993) and as a framework for systematizing the various fallacies which may hinder the rational resolution of differences of opinion (Van Eemeren and Grootendorst 1984, 1992). I will demonstrate that the model can be used also for determining the quality of problem-solving discussions qua discussion, that is, as the medium through which the resolution of differences of opinion is accomplished. In a pragma-dialectic perspective, a discussion qua discussion is good if it provides optimal opportunity for the systematic critical testing of ideas. What this comes down to is that a good discussion is one which optimally enables the execution of the stages of critical discussion. The quality of a discussion qua discussion, then, may be determined by examining how well it enables the execution of the stages of critical discussion. In this paper, I will examine a real-life problem-solving discussion in this fashion, showing that an analysis along these lines enables the analyst to gain a rather precise insight into what went wrong in the discussion, in what respects, and why.[i]

2. The context of the discussion
The discussion took place during the staff meeting of an organization which initiates and manages co-counseling groups. Three of the participants, A, B, and D, are paid staff members of the organization: A and B full-time administrators, D a part-time group coordinator. The fourth one, C, is a volunteer, a representative of the group leaders. C and D are members of the training program committee; A and B regularly meet with the board of directors of the organization. The topic of the discussion is the organization of additional training for group leaders, after the one year of basic training which they receive. A has opened the discussion with the question “where does it belong”.

That something did go wrong in this particular discussion is certainly the opinion of at least two of the participants. After more than one hour of discussion without a decision having been reached, A, in line 1659, queries:
(1)
A: that would have to be something for that kind of committee.
1655C: but they’d have to have something to start from
A: but they’d [have] to have something to start from
D: [yes: hm]

(2)
A: and do we have any ideas on that, then
1660
(2)
because that’s one thing I’m worried about

After a fifteen (!) seconds pause, C gives the following answer:[ii]

(2)
C: well, so what they’d have to start from is the inventarization we’re going around in circles
1665A: ye::s [no but that’s what the problem]
C: [and we’ve been doing that] for the past hour or so,
C and A obviously are of the opinion that the discussion has got stuck.

As a first step towards uncovering what occasioned A and C’s complaint, I will briefly relate what points of view are brought forward in the discussion and how the linear process of trying to resolve the differences of opinion evolves.

After A’s introduction of the question, B briefly sketches the past situation and then argues for the view that the organization of additional training belongs to the domain of the training program committee: a standpoint which A, later in the discussion, also will advance. B’s arguments elicit no reaction; instead, D argues for his own point of view: he questions the need for additional training. A and B attack one of the two arguments which D adduces, but the one which he himself declares most important – the group leaders have never asked for additional training – remains undiscussed. C then brings up another point: who is supposed to pay for the training. During the ensuing discussion of this point, D repeatedly questions the need for additional training, but his questions receive no answer. C replies with practical proposals for finding out what possible topics for training might be and for integrating additional and basic training. The discussion ends in general banter about the financial state of the organization.
After this intermezzo B once again brings up for discussion the standpoint that the program committee should organize the training. D objects by pointing out that nobody on the committee can take on additional work. When B rejects this line of argument as merely practical, D brings in another argument: others may do the job just as well; he then once again poses the question what need there really is for additional training. B says she would like to discuss this question at another occasion, but C `answers’ it by bringing forward a standpoint of her own: before anything else, an inventory of the topics on which training is required must be taken; that is the only sensible basis for any policy at all. A counters that a committee charged with organizing the training could do this; C maintains that it should be done before appointing a committee, repeating her policy argument. A then changes tack. He points out that an agreement has already been made to organize additional training and that it is high time something were done about it. D denies the binding force of this agreement and claims that it is not at all clear what urgency there is for such training. C brings forward doubts of her own against the status of the agreement. The discussion bogs down in an exchange of reproaches.

A manages to soothe the parties and re-initiates the discussion about the question where the organization of additional training belongs. C responds by naming the sources for the inventory which she once again proposes. A doesn’t react to this, but argues for his own proposal to charge the program committee with the organization. C asks for a response to her proposal. A then repeats his own proposal and says it amounts to the same thing. B supports A’s proposal. C once more repeats her proposal. Asked for his opinion, D says he agrees, but only because it will show there is no need for additional training. When B reacts to this with the statement that added training always is necessary, C reiterates that an inventory of the topics on which training is needed must be taken first, A repeats his proposal to charge a committee with this task, and C repeats that first there needs to be an inventory. The discussion closes with both C and A lamenting the fact that the discussion is moving in circles, after which C unilaterally puts an end to the impasse by implementing her own proposal through distributing the tasks for inventarization among those present.

C and A’s lament, we can see now, is justified: the discussion has got stuck in a repetition of standpoints without any progress being made. C forces a breakthrough, but none of the differences of opinion have been resolved. In fact, the various standpoints have hardly been discussed at all.
A and B’s standpoint, that the organization of additional training belongs to the domain of the training program committee, receives direct discussion at only one point, when D argues against it by saying that it is not feasible and that there are other people who can be charged with the task. The first of these counterarguments is rejected as merely practical, the second receives no response at all. For the rest, C and D’s reactions concern the standpoint only indirectly; they address presuppositions of the question to which it is presented as an answer.
D’s standpoint, that there is no need for additional training, is only responded to with regard to a subordinate issue; his main point remains undiscussed. D’s questions regarding this need are reacted to by C with practical proposals for conducting an inventory and for integrating initial and additional training. A and B, implicitly or explicitly, declare these questions out of order.
C’s standpoint, that an inventory of the topics on which additional training is required must be taken first, is not discussed at all; A, who is C’s main opponent, does not respond to her arguments, but invariably replaces her proposal with his own one.
By investigating how the successive stages of critical discussion have been executed in this particular discussion, I think we can reach a diagnosis of how this unfortunate course of events could develop. I will deal with the stages in their order.

3. The confrontation stage
In the confrontation stage of a critical discussion, the differences of opinion which the discussion addresses must be externalized. Our discussion pertains to a multiple mixed difference of opinion: involved are three main standpoints and three contra- standpoints against these, and all of these standpoints meet with doubt. The three main standpoints are: additional training belongs to the domain of the program committee (A and B); it is unclear what the need for additional training would be (D); before anything else, an inventory of the topics on which additional training is required must be taken (C).
The three main standpoints are expressed, but this is not the case for the doubt against them, the contra-standpoints, and the doubt against these. That this doubt exists and that these contra-standpoints are being maintained can only be inferred from the fact that the participants repeatedly respond to the expressed standpoints by bringing forward a different standpoint of their own.
In itself, the fact that doubt and contra-standpoints are not expressed explicitly is not unusual, nor does it necessarily form an impediment to a proper execution of the procedure for resolution of a difference of opinion. But the fact that the various positions which the participants take have not been clarified, almost undoubtedly is one of the causes for the defective execution of the subsequent stages which we shall encounter below.
At another level, a more serious defect can be observed. Behind the differences of opinion which get talked about in the discussion, the existence of another difference of opinion may be divined; this one, however, is not talked about.

As A makes clear when he refers to the earlier agreement (in lines 850-880), the issue of additional training has been around for quite some time, without anything being done about it. A mentions that he even had to account for this to the board of directors:
(3)
A: That’s sort of the way it is the expectations of uh
D: yes
A: the board
D: yes [but]
940 A: [and and that]’s e- because of because I‘ve, yes, because I’m involved because of course I’ve mentioned that the other time I said well hh uh (.), the additional training, that was on the staff agenda, that was last time then we didn’t get to it ((…)), well, then there was a big hullabaloo right away, gee what a shame ((…))
955 you see, so that’s the expectation there

Later, A attributes this failure to execute the agreement to the training program committee (of which C and D are members):
(4)
A: I’m also to blame for this myself I think, but I think, like, the program committee
1065 as well as far as that is con- if there would have been time for that so to speak, huh, or space at least that is my estimation, I don’t know whether that is the case, then that could’ve been worked out (.) or faster. right? but now

This opinion doesn’t surface until almost three-quarters of the discussion has gone by and it is at no point explicitly made into an issue for discussion. Earlier in the discussion, it is mirrored only indirectly in the content of A and B’s standpoints: the organization of additional training is the province of the program committee.

D, in turn, feels that he cannot be expected to take this task on in the context of the part-time job which he holds. That comes out most clearly in the part of the discussion in which the participants engage in reciprocal reproaches:
(5)
B: yes well I think you as a member of the program committee, that it’s up to you
1130 to fill in the details on that. how is a board supposed to know, hh
D: make it into a full-time job then, then I’ll do it

D, too, fails to make this opinion of his into an explicit issue for discussion. It only indirectly surfaces in the fact that, whenever A and B try to assign the committee of which he is a member the task of organizing added training, D puts the need for this training into question.

4. The opening stage
In the opening stage the roles of protagonist and antagonist must be distributed and the shared starting points for the discussion must be established. In our discussion, neither of these tasks gets performed properly.
All participants have the role of protagonist for their own standpoints. In addition, they all have the role of antagonist against the other two standpoints and that of protagonist for the contra-standpoints against the same. In our discussion, however, the latter two roles do not get performed adequately. The participants hardly address each other’s arguments and points of view. They argue almost exclusively in favor of their own standpoints. They thus simply replace one standpoint by another, without subjecting the replaced standpoints to any criticism. They don’t seem to realize that taking a different point of view implies doubt and a contra-position, which carries a burden of proof, against the original one. This may very well be a consequence of the fact that in the confrontation stage the various positions of the participants were not clarified.
As to the shared starting points: one of these is certainly that at some point and by someone an inventory must be made of the topics on which additional training is required. This idea is a presupposition of a number of contributions of various participants, and it is challenged by no one. But the fact that it is a common starting point is not established by any one. In itself, that is not strange – common starting points typically remain implicit -, nor is it particularly wrong, but the discussion could have been simplified considerably if it had been. The discussion could then have been reduced to the questions of when and by whom the inventory should be taken.

More serious is the fact that on other issues there exists a profound but unacknowledged difference of opinion as to what belongs to the common ground. On the one hand, according to D, before the question of where additional training belongs can be discussed, there must be agreement about the need for such training, and according to C, data must be available about the topics for which this training is required. Neither agreement nor data exist. So, with neither whether nor what established, A and B demand an answer to where. A and B, on the other hand, take it for granted that there exists a long-standing agreement to organize additional training, and that it is merely a question of who is going to do it. Whether and what are no longer relevant issues, according to them.
The result of this implicit difference of opinion as to what does and does not belong to the common ground, is that the discussion cannot progress. Every time A and B pose the question where, D and C return to the questions whether and what. And those questions cannot be answered in the discussion because A and B consider them no longer relevant.

Figure 1 – Discussion sequence

5. The argumentation stage
In the argumentation stage, the protagonist brings forward argumentation for his standpoint, to which the antagonist critically responds. In our discussion, the execution of this stage is flawed in several respects. Partly, this is the direct result of the inadequate division of dialectical roles mentioned above: the participants hardly react to the standpoints and arguments of the other party. A crass example of this is A, who does not at all respond to C’s proposal, but instead presents one of his own, and when B and C protest and demand a reaction, repeats his own proposal and claims it boils down to the same thing.
But in other respects as well, the connection between the various contributions is rather loose. This applies, for one thing, to the local relevance of these contributions. Many of them relate only superficially to the preceding utterances of the co-participants. Examples are the passages where D asks whether there is any need for additional training, and C replies with practical proposals for finding out what possible topics for training might be and for integrating additional and basic training. The recurrent absence of local relevance results in conceptual confusion, talking at cross purposes, false agreement and, in the end, a fragmentary discussion of the standpoints.
Overall relevance, as well, is less than ideal. The participants hardly seem aware of the main thread of the dispute. Digressions abound. As a result, the discussion takes a meandering course (see Figure 1: discussion sequence). A topic or proposal will get discussed for a shorter or longer while, but every time, before the discussion is brought to a close, another topic emerges, which in turn is not dealt with decisively, after which earlier topics once again come into focus, are again not dealt with decisively, etcetera, without, and that is the point, any progress being made.[iii]

6. The closing stage
In the closing stage, the results of the defence of the standpoints, which has been undertaken in the argumentation stage, are determined. If a standpoint has been defended successfully, the antagonist must withdraw his doubt; if the standpoint has not been defended successfully, the protagonist must withdraw it. In our discussion, this stage, too, is only partially performed.
Apparently, since every one in the end cooperates in implementing C’s proposal, that is the proposal which all participants accept. In itself, that is not surprising, since no one has objected to the idea of inventarization. But the other proposals have not been refuted, nor have they been retracted. A keeps on defending his proposal to the very last, even when B voices agreement with C’s. D, too, maintains his own standpoint; he combines it with C’s. In addition, the `acceptance’ of C’s standpoint is not the result of a weighing of the different standpoints. Such an assessment simply has not taken place. In pragma-dialectical terms, then, the difference of opinion has been settled, not resolved.
In large part, the inadequate execution of the closing stage can be traced back to the deficiencies in the preceding stages which I have pointed out. Because the different positions of the participants with regard to each other’s standpoints have not been clearly explicated, making up the balance becomes more difficult. Because the participants mainly take on the role of protagonist for their own standpoints, other standpoints and arguments have not been scrutinized critically and therefore cannot be rejected or accepted on the basis of a critical assessment. And, finally, such assessment is hindered by the fact that the participants hardly have any awareness of the main thread of the dispute: they lack an overview of what has been adduced pro and contra the different standpoints.

7. Conclusions
In this paper, I have examined a problem-solving discussion which the participants themselves declared unsatisfactory. I outlined the development of the discussion and pointed out what went wrong. The participants turned out hardly to have responded to one another’s standpoints and arguments. As a result, with regard to none of the three main standpoints could the differences of opinion be resolved. By looking at the way the stages of critical discussion were executed in this discussion, I then was able to establish how exactly this had come about. None of these stages turned out to have been performed fully. In the confrontation stage, the various positions of the participants were not clearly explicated, nor was the underlying difference of opinion brought out and put up for discussion. In the opening stage, the positions of antagonist and of protagonist of the contra-standpoint were not taken on, nor was there full agreement about the starting points for the discussion. In the argumentation stage, contributions often were only loosely connected, and in the closing stage no assessment was made of the various positions. Through this analysis, then, the sources of the unfortunate development of the discussion could be established.
To be sure, the analysis carried out in this paper only revealed whether the discussion process did enable the procedure for resolution of a difference of opinion. I did not establish how well this procedure itself was carried out. That would imply evaluating the substance of the moves which were made: whether contradictions and inconsistencies were present, whether any fallacies were committed, what the quality of the arguments was, and whether the assessment of these arguments was appropriate. My purpose in this paper has been solely to demonstrate that the model of critical discussion can be used fruitfully as a diagnostic instrument in the evaluation of problem-solving discussions qua discussion, that is as a process creating the conditions for rational resolution of a difference of opinion. I might as well mention here that in view of this purpose something else was not done, either: I did not present a detailed account of my reconstruction of the positions of the participants and of the moves they made in the discussion.[iv] Obviously, in a full analysis and evaluation all of these tasks must be performed.

The process-oriented diagnostic use of the model of a critical discussion which I have demonstrated in this paper has several advantages. In the first place, because it focusses on the interactional processes between participants, it gives perspective on some of the deeper, social causes of the derailment of discussions. In this discussion, for instance, it turns out that there is a conflict of interests, connected with the different institutional positions of the participants, which hinders the progression of the discussion. A and B, who try to obtain a decision as to where the organization of additional training should be placed, are policy-making staff members who regularly meet with the board of directors of the organization and who have to set things in motion. C and D, who launch concrete questions and objections regarding the need for and the content of additional training, stand, as volunteer group leader and group coordinator, respectively, and as members of the training program committee, with both feet in the arena of practical action. They are the ones who have to put the proposals of the policy-makers into effect. Obviously, the interests and responsibilities of these two parties differ. This difference is at the root of the different positions which they take in the discussion and of their persistence in maintaining these positions.
In the second place, applying the model of critical discussion makes it possible to enumerate the tasks which, if performed, create the conditions for a discussion to issue in as good a decision as possible. These tasks would include: making sure that the different standpoints which are at stake are explicitized, encouraging participants to react critically to standpoints and arguments, stimulating participants to take stock of their common ground, keeping an eye on the main thread of the discussion, providing summaries of arguments pro and con, guarding against digressions, making relevant distinctions, ensuring critical final assessment of all positions, etcetera. A list like this, derived from the steps which should be taken in the different stages of critical discussion, may help participants in problem-solving discussions to improve the quality of their participation: it may thus provide an instrument for safeguarding the quality of problem-solving discussions.[v]

NOTES
[i] That it is justified to analyze problem-solving discussions as critical discussions, is argued in Van Rees (1991).
[ii] Most pauses last no longer than one second (Jefferson 1989).
[iii] In itself, such a meandering course is not unusual, but ordinarily, contrary to what happens here, it produces progress towards consensus (see Fisher 1980).
[iv] How such an account can be given, is demonstrated in Van Rees (1995, 1996).
[v] There is a point here which may be so self-evident as to escape notice. The concept of critical discussion makes it possible to develop a workable conception of quality. So far, quality of problem-solving discussions has been an extremely unmanageable notion (Hirokawa et al. 1996). In a pragma-dialectical framework, a precise elaboration of this concept becomes possible: the quality of a discussion is directly linked to the degree to which it enables the rational solution of a conflict of opinion.

REFERENCES
Eemeren, F.H. van & Grootendorst, R. (1984). Speech Acts in Argumentative Discussions. A Theoretical Model for the Analysis of Discussions Directed towards Solving Conflicts of Opinion. Dordrecht: Foris/Berlin: Mouton de Gruyter.
Eemeren, F.H. van & Grootendorst, R. (1992). Argumentation, Communication, and Fallacies. A Pragma-dialectical Perspective. Hillsdale, New Jersey: Lawrence Erlbaum Ass.
Eemeren, F.H. van, R. Grootendorst, S. Jackson & S. Jacobs (1993). Reconstructing Argumentative Discourse. Tuscaloosa: The University of Alabama Press.
Fisher, B. A.: 1980. Small Group Decision Making: Communication and the Group Process. New York: McGraw-Hill.
Hirokawa, R.Y., L. Erbert & A. Hurst (1996). Communication and group decision-making effectiveness. In: R.Y. Hirokawa & M.S. Poole (Eds.), Communication and Group Decision Making (pp. 269-301), Thousand Oaks: Sage.
Jefferson, G. (1989). Notes on a possible metric which provides for a “standard maximum silence” of approximately one second in conversation. In: D. Roger & P. Bull (Eds.), Conversation: an Interdisciplinary Perspective (pp. 166-196), Clevedon: Multilingual Matters.
Rees, M.A. van (1991). Problem solving and critical discussion. In: F.H.van Eemeren e.a. (Eds.), Argumentation Illuminated (pp. 281-292), Amsterdam: Sicsat.
Rees, M.A. van (1995). Analyzing and Evaluating Problem-Solving Discussions. Argumentation, 9, 343-362.
Rees, M.A. van (1996). Accounting for transformations in the dialectical reconstruction of argumentative discourse. In: J. van Benthem, F.H. van Eemeren, R. Grootendorst & F. Veltman (eds.), Logic and argumentation. KNAW Verhandelingen, Nieuwe Reeks, 170 (pp. 89-100), Amsterdam: North-Holland.

1. Introduction
I am told by my informants that in Dutch the term argument has intrinsically positive connotations, that positive approval is built into the use of the term. Arguments and arguing are good things. That may be the case for Dutch, but in American English the term argument is starting to become a bad name. People accuse argument of being a force for social exclusion, a means of enforcing hierarchies of power and privilege. Others see in it adversaries, antagonists, contestants, winners and losers, conflict, competition, criticism, and social alienation. It is found in the trickery and stratagems of lawyers and spin doctors whose doubletalk can make anything seem reasonable. Argument appears to others as just one more instrument in the arsenal of slick Madison Avenue admen and selfserving Washington politicians who can justify anything, promote anything, excuse anything, and get away with anything. There is even disenchantment with its seeming use as a forum in which experts and authorities may dither and debate any issue until the public finally loses interest or it is too late to do anything meaningful. We live in a world in which O.J. Simpson walks, Bill Clinton smirks, greenhouse gases still spew into the atmosphere, the tobacco companies continue to sell cigarettes to children, and the lawyers all get rich. If argumentation isn’t part of the problem, it isn’t much of a solution either. At least, that’s how it seems to many people these days.

Now I happen to think this is all mistaken. I happen to think argumentation has a lot to offer in the way of solutions to these kinds of problems. And I think most of you will agree with me. But I also think that this suspicion and this mistrust of argumentation has little to do with the kind of concerns we have traditionally emphasized as a field of study. I think that is why there is suspicion and mistrust, mistaken as it may be. I think these people see something about argumentation that we academics tend to overlook and need to address. What ordinary people see are problems in the pragmatics of argument.

2. Traditional Pictures
For most of its contemporary history, argumentation theory has been dominated by a particular picture of what an argument is. The picture is a visual model that looks like this:
(1)
All Greeks are men.
All Athenians are Greeks.
– – – – – – – – – – – – – – – –
Therefore all Athenians are men.
(Copi, 1953: 163)

or sometimes it looks like this:
(2)
Harry was born – – : – – Harry is a British subject in Bermuda
A man born in Bermuda will be a British subject.
(Toulmin, 1958: 99)

or other times it looks like this:
(3)
P1 P2 P3
C

P1. Frances is very successful in her career.
P2. Frances has a secure and supportive marriage.
P3. Frances had a stable and secure childhood.
C. Therefore, Frances is a happy person.
(Hughes, 1992: 82)

Whatever the details, the general character of the picture remains pretty much the same. What we see is a picture of arguments as semantic structures, as assemblies of propositions. It is an essentially geometric and logical conception of argument (Toulmin, 1976). In order to highlight the structural form of inference, we have come to treat arguments as very abstract entities. In fact, we think of them so abstractly that we easily slip into talking about arguments simply as ideas, as virtual entities that exist independently of any medium of expression, without any time or any place of occurrence. This picture invites us to think of arguments deprived of their functioning, stripped of their context, divorced from the social engagements in which they actually occur, and even isolated from the issues and concerns that motivated their production in the first place.

I think many if not most of us at this conference are not altogether happy with this picture. But we are comfortable with it. It is a picture that has insinuated itself into most of our theoretical puzzles, and I am afraid that it has instilled in us a kind of occupational blindness, a trained incapacity to work with aspects of the actual phenomena that ultimately we are really concerned with. Even when we remind ourselves that these models are only that – models of arguments and not the actual arguments themselves – we still tend to narrowly restrict our selection of real-life cases. We still tend to work with those cases that most directly correspond to the model form. We still tend to present sanitized cases that are already standardized, unitized, explicated, and otherwise neatened up for easy application of the models. That’s fine, if you are concerned only with the properties of arguments that these models were designed to highlight in the first place – properties like premise acceptability, argument strength, or inferential form. But if you are concerned with other properties of argument, properties having to do with interpretive meaning, functional design, procedural organization, situational adaptation, and the like, we need something else.

3. Normative Pragmatics
I think that something else is normative pragmatics. I want to suggest that normative pragmatics provides a useful corrective and a helpful complement to the kind of modelling we ordinarily undertake when we analyze and evaluate arguments. I like the term “normative pragmatics” – which was first coined by Frans van Eemeren and Rob Grootendorst – because it cuts across the old distinctions between rhetoric and dialectic and because it insists on attention to the uses of argument in ordinary language (van Eemeren, Grootendorst, Jackson & Jacobs, 1993). I like it because the terms points to analytic practices that are empirical in much the same sense that the broader field of discourse studies is empirical: Our theories and principles ought to be accountable to the actual practices and intuitions of natural language users (van Dijk, 1997). I believe that argumentation is first and foremost a linguistically explicable phenomenon, and as analysts we must hold ourselves
accountable to the details of actual messages. Simply put: in normative pragmatics, messages become our object of study. That’s an idea that I like to think echoes J. L. Austin’s (1962: 148) injunction summing up his 1955 William James lectures at Harvard. He concluded:
(4)
“The total speech act in the total speech situation is the only actual phenomenon which, in the last resort, we are engaged in elucidating.”

Austin was suggesting a unit of analysis for analytic philosophy in contrast to the traditional attention to propositions, and he was suggesting standards for assessing utterances far more varied than simply that of truth and falsity. And that’s one of the things that I find very appealing about normative pragmatics.

But the study of argumentation is not just pragmatics; it is normative pragmatics. So, it is not simply empirical; it is also critical. And it is the complex interplay of empirical and critical attitudes which truly animates the normative pragmatic study of argumentative messages. One thing this means is that the scope of argumentation theory extends beyond clearcut instances where arguments obviously occur. As argumentation theorists we should be concerned with discourse where arguments should be used, whether they are used in any obvious way or not. The observation that some discourse is not an argument (and so it’s not our problem) doesn’t necessarily carry much weight. It might, but the real question to be asked is whether or not it is useful to examine some discourse with respect to how we think argument should work in this context. The real question is whether or not the perspective of argumentation theory provides a useful frame of reference for analyzing and assessing what is going on in the discourse. So, as students of argumentation from the perspective of normative pragmatics, we must be concerned with a wide range of discourse, messages, and interactions whose properties can be explicated with an interest in their argumentative functions and structures despite their overt appearances.
Now, I don’t intend to hawk in my talk today any particular version of normative pragmatics. Normative pragmatics is a broad genre that encompasses many particular theories and research paradigms. I want to simply argue today for two ideas that I think are fundamental to any particular approach to normative pragmatics. Those two ideas are this: First, normative pragmatics calls for a return to the study of the communicative properties of actual argumentative messages. Second, normative pragmatics makes central the analysis and assessment of the functional properties of those argumentative messages. I am convinced that if argumentation theory is going to have anything important to say about the kinds of misgivings so many people have about contemporary argumentive practice it must address those two properties. And it’s the importance of those two properties to which I now turn.

4. Expressive Design
First, let me talk about the need to attend to the communicative properties of actual argumentative messages. Too often the problem of reconstructing arguments has been a problem of refashioning stated propositions, filling in missing premises, drawing out implied conclusions, but without any real sensitivity to the total message that is being conveyed. Oftentimes it seems that argumentation theorists treat the vagaries and complexities of communication as though this were an analytic predicament, as something to be solved through methods that render what is said into the “actual” argumentative form. Another way to think about these features, one which flows naturally from a pragmatic understanding of messages, is to see the interpretive problems of communication as an analytic puzzle – not as a barrier to analysis, not as a predicament, but as a thing to be analyzed, as a fact to be explained. The traditional response treats communication as a curtain drawn over the underlying argumentative structure, as something to be brushed aside if possible. Normative pragmatics invites us to treat communication as a tapestry into which the argument itself has been woven (Jacobs & Jackson, 1992). I am here reminded of Manfred Keinpointner’s (1998) observation in his address Wednesday that figures of speech are fundamental to language, and not just ornamental.
Information conveyed in a message is not limited to what can be extracted from sentences by rules of syntax, semantics, and logic. And the information constructed by means other than these rules should not be discounted or dissolved. When people interpret a message, they construct a context of assumptions and inferences that make sense of what was said and of what was not said but could have been said, and that make sense of how and when all of it is said. The words are not the message. The words and sentences are simply part of an assembly of cues that people use to construct the message. It is the context of interpretive assumptions and inferences that is the message. And it is the message that has argumentative functions.
To see what I mean, consider Senator Edward Kennedy’s nationally televised account of what happened the night in which, following a party at a summer cottage on Chappaquiddick Island, he apparently drove off a bridge; his passenger, Mary Jo Kopechne, drowned; and then the Senator waited all night until the following morning to report the accident. This speech, given in July of 1969, marks a turning point in American political history. It occurred at a time when the overwhelming majority of Americans and political commentators expected that Ted Kennedy would not only one day run for President of the United States of America, but would become President of the United States of America. Short of a bullet, few people believed anything would stop his ascension. Of course, no one saw the bridge. And, apparently, neither did Kennedy. Here are two excerpts from his speech. The first excerpt refers to the time immediately following the accident after the Senator had failed in his own efforts to swim down to the submerged car and find Miss Kopechne and get her out. The second excerpt reports what the Senator did after waking in his hotel room the following morning.

(5a)
Instead of looking directly for a telephone after lying exhausted in the grass for an undetermined time, I walked back to the cottage where the party was being held and requested the help of two friends, my cousin, Joseph Gargan, and Phil Markham, and directed them to return immediately to the scene with me – this was some time after midnight – in order to undertake a new effort to dive down and locate Miss Kopechne. . . . In the morning, with my mind somewhat more lucid, I made an effort to call a family legal advisor, Burk Marshall, from a public telephone on the Chappaquiddick side of the ferry and then belatedly reported the accident to the Martha’s Vineyard police. [Underlining has been added – ed.] (Senator Edward Kennedy’s Address to the People of Massachusetts July 25, 1969)
The speech as a whole is clearly an exercise in political apologia. This is a speech of self-defense, and the details of the story told in that speech convey an argument to the effect that the Senator was not culpable of any wrongdoing in the events preceding or following Miss Kopechne’s death. Both of these passages help to convey information that supports this claim. The passages suggest the impression of a distraught and disoriented young man searching for help from his friends. The Senator does not overtly argue that his actions were not motivated by some scheme to cover-up his involvement in the accident. Nor does anything he says logically imply that. But the impression given is clearly a contrast to such scheming, and that is the argument these passages are no doubt intended to convey.
To see that this is part of the message, simply consider the underlined passages (“two friends, my cousin, Joseph Gargan, and Phil Markham,” “a family legal advisor, Burk Marshall”). Both characterizations are true, and both characterizations are no doubt relevant to explaining Kennedy’s conduct. But the truth and relevance of the descriptions per se are secondary to the commonsense knowledge these labels invoke. Harvey Sacks (1972) called such labels “membership categorization devices.” Sacks claimed that labels like “friend,” “cousin,” or “family legal advisor” give particular meaning and motive to associated activities like, in this case, “requesting help” or “making a call.” They also imply their adequacy relative to other possible labels. That is, people assume that these labels are not merely descriptively sufficient; people assume that these labels are the most sufficient descriptions relative to other possible descriptions. And that pragmatic assumption is where the real argumentative impact of these labels is to be found. The role of this pragmatic assumption can be seen by considering an alternate possible description. what if Kennedy had said this?

(5b)
Instead of looking directly for a telephone after lying exhausted in the grass for an undetermined time, I walked back to the cottage where the party was being held and requested the help of two attorneys, my long-time political aide, Joseph Gargan, and former U.S. Attorney for Massachusetts, Phil Markham , and directed them to return immediately to the scene with me . . . . . . In the morning, with my mind somewhat more lucid, I made an effort to call long-time advisor for the Kennedy political machine and a man Bobby Kennedy considered the sharpest lawyer he ever met, former Assistant Attorney General Burk Marshall.
Now, these descriptions are equally true, and perhaps equally relevant to explaining Kennedy’s conduct. But these descriptions suggest quite different motives and activities, and in no way do they communicate the impression that Kennedy was not involved in a cover-up that night or was not capable of hatching some scheme to try to save his career from catastrophic political scandal, to say nothing of charges of reckless driving, driving under the influence, and involuntary manslaughter.
And in addition to comparing what was said to what could have been said, consider the related matter of things left unaddressed – what was not said and was omitted as an issue altogether. Again we can see that people make a kind of pragmatic assumption in interpreting discourse: The assumption goes that if something was not mentioned, it must not be important and what was mentioned must be informationally sufficient for the purposes of the message. So, consider the following alternate story, again based on previously excluded but true and presumably relevant information. What if Kennedy had added this passage to his first excerpt?

(5c)
Instead of looking directly for a telephone after lying exhausted in the grass for an undetermined time, I walked back to the cottage where the party was being held and requested the help of two attorneys, my long-time political aide, Joseph Gargan, and former U.S. Attorney for Massachusetts, Phil Markham , and directed them to return immediately to the scene with me – this was some time after midnight – in order to undertake a new effort to dive down and locate Miss Kopechne. [I did not alert any of the other five women and three men at the party, including Raymond LaRosa, a fireman trained in scuba-diving rescue.]
Withholding the information about who was not alerted can be seen to have a pretty clear argumentative impact once the information is provided. And I think most people would think that omitting that information from the story is deceptive in some way. But what kind of assumptions are constructed for Kennedy’s story that are falsified by this new information? I’m not exactly sure what kind of propositions we should reconstruct here – or even whether explicating substantive assumptions is what is really called for here. The assumption of some very general pragmatic principles of communication may be all that is needed. But the point to see is that whatever those assumptions are, they create an impression of sincere  and honorable intentions, and those assumptions are not the kinds of assumptions that we ordinarily “explicate” when reconstructing an argument. But they ought to be explicated – at least they ought to be explicated if we want to explain why people consider political speeches like this one to be so sleazy and why people think politicians can get away with anything these days. And whatever the pragmatic principles of interpretation are that people are using to make sense of Kennedy’s story, we should see that they are principles that have a real impact on the argumentative reasoning encouraged in the message.

Now, my point about Kennedy’s argument is not to show simply that it is defective in some important way. Rather, the point is to see that the kind of information that I have just provided, exposes the message as defective and so this kind of critical comparison tells us something important about what kind of a message is being communicated. People would not have these intuitions of defectiveness given this information if they did not also have certain intuitions about the argumentative message design in Kennedy’s story. Consider another example of message design. This one appears somewhat simpler, and that is part of what is tricky about it. It’s the product claim for Tylenol:
(6)
Tylenol. The pain reliever hospitals use most.

That product claim is repeatedly presented in ad copy as a compelling reason to conclude you should choose Tylenol over other pain relief products. But again, what assumptions do people make in constructing the message conveyed by these words? How, exactly, do people see the product claim, “Tylenol is the pain reliever hospitals use most,” as somehow supporting the tacit main claim, “You should choose Tylenol for pain relief”? Presumably, there is some kind of sign reasoning that depends on the reliability and authority of hospital choice as an indication of the reasonableness of one’s own personal choice of Tylenol. But how much deeper do we go? Deeper, I would say, than we are ordinarily used to going as argumentation
analysts.

One of the complexities here is that in almost all their ads, Tylenol offers additional product claims to superiority. For example, one ad features this header in large bold print in the page center: “There are more pain relievers than ever. But there’s only one that hospitals use most. TYLENOL.” Then in the bottom righthand corner, beside a picture holding a bottle of Tylenol capsules, appears the following ad copy:
(7)
Nothing’s more effective. Nothing’s safer.
TYLENOL products give unsurpassed pain relief without the stomach irritation you can get with aspirin or other kinds of pain relievers.
For you and your family, doesn’t it make sense to choose the pain reliever hospitals use most? There’s only one.
TYLENOL.
The pain reliever hospitals use most.

Another ad appears over a picture Extra-Strength Tylenol geltabs placed across from a row of three boxes of pain relievers containing aspirin, naproxen sodium, and ibuprofen. The header reads: “Your stomach knows the difference between these pain relievers… And this one.” The ad copy in the bottom righthand corner explains:
(8)

The pain relievers doctors call NSAIDs – aspirin, the latest drug with naproxen sodium, and even ibuprofen – have a number of similarities.
An important one has to do with your stomach. To varying degrees, every NSAID brand can sometimes irritate your stomach.
That’s because NSAIDs may reduce your stomach’s natural ability to protect itself.
But TYLENOL is different. It won’t irritate your stomach. You know how well TYLENOL works. And now you know it’s definitely gentler to your stomach.
The choice is clear. The choice is yours.
Tylenol. The pain reliever hospitals use most.

The claim that Tylenol is the pain reliever hospitals use most is repeatedly placed in a slot where conclusions might be found. Now, should we conclude from this juxtaposition of ad copy that the advertisers are arguing that the preceding copy are the reasons hospitals use Tylenol most? Should we conclude, for example, that hospitals use Tylenol most because they believe nothing is more effective (as effective?) and nothing is safer (as safe?) as Tylenol? Should we conclude that hospitals use Tylenol most because it is gentler on people’s stomach than the available alternatives? No Tylenol advertisement ever explicitly makes that kind of link. And nothing logically requires such a link. However, people do seem to naturally assume that these reasons are juxtaposed in texts for just this sort of rationale. Again, I think it is fair to say that people have a tendency to make a pragmatic assumption that if a connection makes sense, and it’s an obvious connection to draw, and nothing is done to prevent that connection, then that connection should be drawn. Granted, this is a somewhat tenuous connection, but simply because it is tenuous doesn’t mean it’s not conveyed – only that it is conveyed tenuously.

Still, even if we take the product claim about hospital use in isolation, there is more being communicated than simply that product claim and some warrant about the reliability of signs or authority. To see what more there is, consider some additional information: The actual reason hospitals use Tylenol most is because Tylenol gives its product to hospitals for free. When they find this out, many people feel misled (though maybe not surprised).
What does that show us about the original message that people must be constructing from these Tylenol ads? Well, at a minimum, it should be seen that the problem with tricky ads like this one is not at the level that ordinary people often think it is. It’s not at the level of a lie, or some falsification of stated content. And it is not at the level of some vagueness in word meaning or ambiguity of phrasing. That’s all clear enough. The problem is with the pragmatic assumptions people make in constructing the message. Even if people took the hospital claim to be an independent reason for choosing Tylenol, they pretty clearly construct some substantive backing for the argument: They feel justified in assuming from this ad that the reason hospitals use Tylenol most is because hospitals think Tylenol is the best quality pain reliever. (And not, e.g., that hospitals think Tylenol is just not noticeably worse than any other pain reliever – which is really all that a statement like “Nothing’s safer” really says. See Jacobs, 1995.) That must be part of what people take to be the argument here, or else they wouldn’t think it’s a deceptive ad (as opposed to, say, just an underinformative ad) when they find out that such an assumption is not true.
So, if as argumentation theorists we are going to be able to see what is going on in an argumentative message, and if we are going to be able to properly assess the troubles in those messages, we are going to have to take into account the expressive design of those messages and the pragmatic principles of interpretation on which those designs are based.

5. Functional Design
It is not only the communicative properties of messages – their expressive design – that normative pragmatics calls attention to. Arguments also have a functional design: Their meanings are implicated in chains of social and cognitive consequences that have a bearing on the deliberative process. Understanding that functional design is key to seeing what makes something a useful or obstructive contribution to the decision-making process. Now by this I do not mean simply that argumentation theory should be concerned with persuasive effects. Instead, I mean something related to that: argumentation theory should be concerned with the way in which argumentative messages enhance or diminish the conditions for their own reception. Argumentative messages may be designed either to open up or to close down the free and fair exchange of information. Argumentative messages may be designed either to encourage or to discourage critical scrutiny of the justification for alternative positions. I think one of the real insights of normative pragmatics is that argumentation is self-regulating and self-sustaining in just this way. Now, this is a practical matter, and argumentation theorists have traditionally been loathe to address matters of the practical design and social engineering of discourse structures. But the pragmatic problems and solutions of argumentative practice exist in the form of discourse strategy – and not just discourse norms – and at the level of institutional procedures – and not just inferential schemes.

One such practical institutional context that has held considerable research interest for myself has been the procedures of third-party dispute mediation. As a system of
dispute resolution, mediation creates a context which in certain ways of arguing are reasonable and functionally constructive and in which other ways of arguing are not. Consider the following exchange between a divorcing husband and wife who have been required by the court to attend a mediation session for the purpose of trying to work out a custody and visitation arrangement for their children:
(9)
01 M: Okay. Mrs. ( ), let’s hear from you, what kind of plan do you think that we could reach
02 W: Well um I’d like for them to live a normal ( )=
03 H: =What’s normal, cocaine addict uh uh (aren’t you) a patient, outpatient [uh uh uh oh and] and uh=
04 W: [My ( ) people]
05 H: =uh trick every night? Is that, is that it, is that it?
06 W: I don’t under[stand]
07 H: [She had] a fifteen year old kid coming over and staying the day while these kids were locked up in the front yard while I was at work every day, I have a witness proof for that
08 M: Okay=
09 W: =you do, who
10 H: Ann Cray.
11 M: Let=
12 H: =she was the one who told me about it all= [cause ‘sher fifteen year old son
13 M: =Let’s [hear Let’s hear what, what your plan would be

This exchange comes early on in the session. The husband (H) has just proposed a plan in which he gets custody of the two children and the wife gets visitation privileges. The mediator (M) then turns to the wife (W) to hear what kind of plan she advocates. I want to focus on the contributions of the husband in turns 03, 05, 07, and in 10 and 12. He makes an argument that, taken in the abstract, is more or less reasonable. It might be pictured this way:
(10)
P1 P2 P3
C

P1. W is a cocaine addict
P2. W is an outpatient at a psychiatric hospital
P3. W carried on an affair with a minor while locking up the kids in the front yard.
C. W will not provide an acceptably normal environment for the kids if given custody.

If the wife is in fact a cocaine addict, an outpatient at a psychiatric hospital, and has carried on an affair with a minor while she locks up the kids in the front yard, there is strong reason to conclude that she is not going to provide an acceptably normal environment for her children if she gains custody.

That’s not a bad argument in principle. But it still should not be called a good argument – at least, not in context. The argument might be a good one for a courtroom or on radio talkshows, but not in mediation. The problems have to do with the pragmatics of the argument. Its tactical design is objectionable. For one thing, it is procedurally out of order. The husband not only interrupts the wife, he does so at a time when she hasn’t even yet described her proposal. But deeper than that, consider what the argument does by the way it is put forward: it seems more designed to censure, embarrass, and shame the wife than to convince her she should not take custody of the children. Notice the taunting (“Is that, is that it, is that it?”) and the offensive formulations (“addict” “trick every night”). The husband’s label in announcing “witness proof” amounts to a barely veiled threat that these arguments are about to come up in court if the wife resists his proposal. Either the husband is picking a fight, or he is acting in a way that will bully the wife into making concessions to avoid further public humiliation. The husband’s argument certainly can hardly be expected to enlist the wife’s cooperation in a collaborative search for a mutually agreeable resolution based on a sincere and careful weighing of the merits of the case. But that is precisely what is called for by the argumentative situation the husband is in: Mediation is an argumentative forum in which the disputants themselves must arrive at a resolution of their disagreement. The mediator only keeps procedural order, and does not make judgments about the merits of either party’s case. In other words, a rational argument here (unlike in, say, a courtroom) must be adjusted to the need to create and maintain a framework of joint problem-solving. That is a functional requirement that is just as crucial to argument quality as requirements of premise adequacy.

One of the things that normative pragmatics quickly reveals is the close connection between the expressive design of messages and their functional rationale. Much of the functional design of arguments has to do not just with what is said when, but with how the information gets conveyed. And one of the real concerns we should have about fallacies is not just what norm of good practice they violate – but how do fallacies pass without notice? How does a fallacy get away with it? One of the very general problems of contemporary argumentative discourse is that information gets conveyed in ways that let the communicator avoid commitment or accountability to the message. The framework of intersubjectivity on which communication relies becomes strained and problematic to the point that what the receiver finds cannot with any certainty be attributed to the intentions of the sender.
Phenomena like this should not be treated as methodological or analytic predicaments but as empirical facts with normative consequences. Think back to the Tylenol ads. The ads never say that their product claims are the reasons why hospitals use Tylenol most. The Federal Trade Commission would no doubt act against that claim. But then, Tylenol claims no such rationale – they only insinuate that rationale in such a way that they are not committed to defending it. And so the ads can keep coming out and readers can continue to be misled.
But we should not think that fallacies always occur by virtue of some sort of covert misdirection, some kind of camouflage or disguise. This framework of intersubjectivity can be exploited and abused in other ways as well – in ways that turn on the very obviousness of the trickery. It is a tactic that depends not on disguising the misuse of argument, but on flaunting it and even reveling in its own audacity.

Example (11)

Example (11) is an advertisement for milk from Oak Tree Farm Dairy of Long Island, New York. At the top is a picture of a three-eyed cow standing in front of the Three-Mile Island nuclear power plant. Beneath the picture is the header: “What is it about milk from Pennsylvania that gives us a bad feeling?” (11)
This is obviously a joke, and meant to be taken as such. The ad plays upon memories of the Three-Mile Island nuclear accident, knowledge that radiation can cause mutations and birth defects (e.g., three-eyed cows), and the more recent reports of the Chernobyl nuclear accident where the release of radioactive fallout actually contaminated the milk supply in nearby areas. The ad is not seriously suggesting that milk from Pennsylvania may be radioactively contaminated. The middle third eye on the cow is fake, and it is obviously fake. The joke is a kind of “hook” by verbal misdirection that is commonly used in print ads as a set-up and lead-in for the written material that follows. You see the introduction and think, they can’t really mean this. So you read on, and it turns out they don’t really mean it. The advertisers are leading into something else about Pennsylvania milk that gives them a bad feeling.

The real concern raised in the ad copy has to do with the freshness of the milk, because it must travel all the way from Pennsylvania to get to New York City (whereas Oak Tree Dairy is a “local” dairy from Long Island).

Now, we wouldn’t ordinarily call this kind of a tongue-incheek strategy of maligning a competitor deceptive. It involves no seriously claimed falsehoods. Nothing is concealed in the strategy. Nothing is disguised. It is not an effort to mislead or fool anyone. Everything is quickly cleared up. It is all above board, out in the open, and anything false is presented as such. It just looks like a pseudo-argument whose functional design really has more to do with attracting a reader than with convincing them of anything. (If there is anything misleading and deceptive about the ad in the ordinary sense, it is an implication that the milk from Oak Tree Dairy does not travel as far as milk from Pennsylvania. In fact, all of the milk processed at the Dairy comes from farms around Syracuse in upstate New York [NYTimes, 1992, Dec. 20, p.15]. Moreover, the shipping time of processed milk from Pennsylvania is only negligibly greater as far as it affects freshness.)

Nevertheless, this is a pretty sleazy tactic. It’s functional design ought to be considered fallacious. What we have here really is an argument. It only seems to be a pseudo-argument harmlessly posing as an argument. The argument only pretends to pretend. Why do I say that? Well, consider what people are going to be thinking about next time they are standing at the dairy shelf trying to decide which milk to buy. Simply raising the concern of radioactive contamination is perhaps enough to get people to think about it the next time they are buying milk, even if the concern is only raised tongue-in-cheek, and even if people know and remember that. In fact, this is an increasingly common tactic. By flaunting the fallaciousness of the argument a knowingly cynical audience is drawn in and disarmed by the very act of exposing what is going on. Thus, in another instance of this tactic, NBA superstar Grant Hill hawks Sprite soda on the television screen while a small cartoon picture of him in the corner chings up and down like a cash register tab. Each time the little picture of a grinning Hill pops up, he is covered in an even larger pile of money. The message is clear: Hill is only advocating drinking Sprite because he gets enormous sums of money to do so. And the audience knows that. And Sprite knows the audience knows that. So why not bring everyone in on the joke that Grant Hill – or any other celebrity – is a credible product sponsor? “Image is nothing. Obey your thirst” goes the Sprite ad campaign motto. But it is Hill’s celebrity image that is the only reason for his presence in the ad. And attraction to him is the cause for attraction to Sprite. And we know it. And we know they know we know it. We have the image of seeing through it all – even when seeing through it shows us that seeing through it is part of how we get sucked in. So what? That’s what makes it all so cool. And a stupid reason becomes a good reason to drink Sprite. As Bill Clinton has shown us all, it’s okay to argue disingenuously if you share the smirk.

6. Conclusion
So, I hope I have made a compelling case that normative pragmatics has a central role to play in argumentation studies. I should say as an aside that I do not see pragmatics as a substitute for traditional logical analyses – formal, informal, or otherwise. It is, I think, useful to recall that H. P. Grice’s (1975) foundational essay on the theory of conversational implicature is introduced as a way of saving the literal meaning of such logical terms as “and,” “or,” and “if…then,” and is entitled “Logic and Conversation.” As I said earlier, I see normative pragmatics as a corrective to traditional analyses and as a complement to those studies, not as a replacement of them.
But I do see normative pragmatics as an indispensable part of argumentation studies. The principles of pragmatic interpretation and practical reasoning that underlie message use are just as fundamental to argumentation as are the principles of epistemic inference. And the pragmatic demands on argumentation are just as central to argument quality as are traditional standards of argument cogency. Only when we recognize this, can we begin to really answer the misgivings and mistrust of ordinary people who must live with arguments as objects with consequences and not merely as objects for study.

REFERENCES
Austin, J. L. (1962). How to Do Things with Words (2nd ed., J. O. Urmson & M. Sbisa, eds.). Cambridge, MA: Harvard University Press.
Copi, I. M. (1953). Introduction to Logic. New York: MacMillan.
van Dijk, T. A. (1997). Discourse Studies (2 vols.). London: Sage.
van Eemeren, F. H., R. Grootendorst, S. Jackson & S. Jacobs (1993). Reconstructing Argumentative Discourse. Tuscaloosa: University of Alabama Press.
Grice, H. P. (1975). Logic and conversation. In: P. Cole & J. L. Morgan (Eds.), Syntax and Semantics, Vol. 3: Speech Acts (pp. 41-58). New York: Academic Press.
Hughes, W. (1992). Critical Thinking. Peterborough, Ontario: Broadview Press.
Jacobs, S. (1995). Implicatures and deception in the arguments of commercial advertising. In: F. H. van Eemeren, R. Grootendorst, J. A. Blair & C. A. Willard (Eds.), Proceedings of the Third ISSA Conference on Argumentation, Vol. IV: Special Fields and Cases (pp. 579-592). Amsterdam: Sic Sat.
Jacobs, S. & S. Jackson (1992). Relevance and digression in argumentative discussion: A pragmatic approach, Argumentation, 6,.
Keinpointner, M. (1998). Figures of speech: Definition, description, and critical evaluation. Plenary address, 4th International Conference on Argumentation, Amsterdam, the Netherlands.
Rabinovitz, J. (1992, 20 December). Dairy wars: An appeal to Long Island’s pride. New York Times, sec. Y, p. 15, cols. 1-4.
Sacks, H. (1972). On the analyzability of stories by children. In: J. J. Gumperz & D. Hymes (Eds.), Directions in Sociolinguistics (pp. 325-345). New York: Holt, Rinehart and Winston.
Toulmin, S. E. (1958). The Uses of Argument. Cambridge, UK: Cambridge University Press.
Toulmin, S. E. (1976). Knowing and Acting.

In the pragma-dialectical approach to argumentation, as represented by van Eemeren & Grootendorst (e.g., 1992) or Walton (1989, 1992, 1995), critical discussion provides the normative model for rational argument. But do the norms for critical discussion also apply to political debate? As rhetoricians, we insist that critical discussion and political debate are different genres with different norms. Critical discussion is dialogic, debate is trialogic (Dieckmann 1981, Klein 1991). The arguers in the discussion address each other with the cooperative goal of resolving the dispute; debaters do not argue in order to persuade each other, but to win the adherence of a third party: the audience (Jørgensen, in press).
Because of its trialogic nature, a debate must answer the needs of the audience. This means that a debate should be evaluated in relation to the functions it fulfils. This does not mean that our approach is oriented toward uses and gratifications in the traditional sense. We are interested not only in the functions of debate, but also in the specific features of debates that serve these functions; and our approach is normative.
We shall concentrate on issue-oriented debates, such as the Irish debate over the Ulster peace plan, or the Danish debate over the Amsterdam treaty. What we have to say about the rhetorical audience and the quality of public debate has particular reference to how debate is conducted on TV.

Opinion polls will tell us that the audience of such debates consists of three groups: those in favour, those against, and the undecided. Commentators typically refer to the undecided as those who have not made up their minds yet, implying that all the others have indeed made their minds up. Accordingly, it is assumed that the outcome depends on the remaining undecided voters.
But this is misleading. Both among those in favour and among those against, there are many who have not made their minds up, and who may well change sides – under the influence of events or  arguments. To document this, we may cite a poll in the French daily Libération shortly before the referendum in France on the Maastricht treaty in 1992. Here – interestingly – voters were asked whether they might change sides on the issue. No less than 37 % of those who intended to vote yes admitted they might also vote no, and conversely for 34 % of those who said they intended to vote no. It is probably true that especially in matters concerning the European Union many voters are in two minds; they feel that there are arguments on both sides of the issue, and they are constantly weighing them against each other.
What this means is that on any issue, the audience represents a spectrum of opinion, with unmoveable partisans at both ends, and with a fair number of voters near the middle of the road who lean to one side but who may be shifted. But debaters and TV programmers tend to make the undecided their primary target because they falsely believe that the static and simplistic Yes-Undecided-No model says all one needs to know about the debate audience. They forget the lesson of the Danish referendum which rejected Maastricht because many voters changed sides at a late stage, even at the polling station.
To understand how some voters can thus be in two minds, we shall propose a model of the debate audience (inspired by Tonsgaard 1992). This, in turn, will allow us to distinguish between the different functions of debate for the public audience.

In this figure, the undecided are represented by the grey area beneath the curve. The white area represents the decided voters, i.e. those who say that they are going to vote yes or no, respectively. Those near the curve are the hesitant voters. The point is that there are two variables which may explain why voters hesitate. These are represented by the two axes.

The x axis represents involvement in the issue, that is, how important the voter perceives the issue to be. The y axis represents the voter’s feeling of assurance on the issue. Those high in both assurance and involvement belong in the area marked “P” (for partisans). What they will want from debates is mainly reinforcement of their existing views. Those low in both assurance and involvement will belong in the area marked “A” (for abstainers, because these people will probably end up not voting at all). But it is also possible to have a quite fixed and assured view of the issue, either for it or against it, and yet feel that it is all quite distant and uninteresting.
These voters – high in assurance but low in involvement – will be in the “S” area (for spectators). They will probably feel little need for guidance because they know what they think – but more of a need for entertainment, and some need for reinforcement. Finally, many voters – certainly in Denmark – see the European issue as highly important, but also as complex and baffling; and that is why they are hesitant. These voters – who are high in involvement but low in assurance – belong in the “D” area (for deliberating citizens). Although they lean to one side, they feel they need to know and understand more, because they are still in two minds; hence they want the ongoing debate to give them guidance for the decision they confront.

This segmentation of the debate audience reflects the analysis of three of the audience roles defined by Gurevitch & Blumler (1977). Their account also includes roles for “media personnel” and “party spokesmen”, as seen in the following table.

In our context, we may disregard the “monitor” role, since we regard it as less relevant for members of a debate audience, and more applicable to, for example, political scientists and commentators. What the voter seeks when he appears in the partisan role is precisely “reinforcement of his existing beliefs”; as a spectator, he seeks “excitement and other affective satisfactions”; as a deliberating citizen – or, as Gurevitch and Blumler have it, “liberal citizen” – the voter seeks “guidance in deciding how to vote” (1977: 276). Our model of the debate audience explains the notion of audience roles and their underlying parameters. The model also implies that there are two basically different ways that a debater can try to increase adherence to his view, dependent on which segment of the our model he mainly appeals to.

1. The debater can prefer to appeal mainly to those who are rather high in assurance, but low in involvement. These people will basically tend to choose the spectator role. Since they are rather assured about their views, the debater must concentrate on those voters in this group who lean to his side already. Those who plan to vote for the side anyway will merely have their enthusiasm boosted. Those who might not have voted may be stimulated to come out and do so. Thus the way this strategy may gain votes is by mobilizing some of the undecided vote. We call this strategy vote-gathering.

2. The other general strategy is to appeal to those voters who lean to the other side but who may be won over. These people are high in involvement, that is, they think the issue is important; but they are low in assurance. Typically, they are deliberating citizens who acknowledge that there are two sides to the issue and that their decision should be based on the weight of the arguments. As we have pointed out, there are often a substantial number of such voters on both sides. We call this strategy vote-shifting.

The distinction between vote-gathering and vote-shifting was one of the perspectives we became aware of in a study of televised public policy debates in Denmark (Jørgensen, Kock, and Rørbech 1994; 1998). In these debates we found voting patterns suggesting that some debaters are particularly good at vote-gathering, others at vote-shifting. For example, in one debate, in front of a hundred representative jurors, one debater gathered no less than 14 votes from the undecided group, but she shifted only one from the opposite side; the opponent gathered just 5, but shifted 9. This is shown in figure 2, where the grey columns show votes gathered and the white ones show votes shifted.

If it is true that some debaters excel at gathering votes, while others are good at shifting votes, then we may ask: What are the essential features of the two types of argumentative strategy that have these distinct effects? Observations from our empirical study have led us to the following hypothesis, which is also consistent with much rhetorical theory. We believe the typical vote-gathering debater will tend to broaden the front between the two opposite sides, while the typical vote-shifter will tend to narrow it.

The typical vote-gatherer will tend to claim fundamental, black-and-white differences and introduce a series of further points of contention that will broaden the front between the two sides. He will claim a fundamental ideological opposition between the two sides; he will impute a series of further claims and positions to the opponent that have not been mentioned by the opponent himself; he will see the opponent’s proposal as “the thin end of the wedge,” as part of a large campaign, or even of a conspiracy; he may attack his opponent’s motives, he may bring in matters that cast doubt on the opponent’s intelligence, ethics, or good will; he will typically attack the weakest arguments made by the opponent, trying to make them out as ridiculous, or as self-contradicting. Front-broadening arguers generally spend much energy on refutations of arguments made by the opponent, and on counter-refutations of refutations, and so on ad infinitum. In all this, the issue at hand will often disappear in a confusing verbal duel. As audience, we may find ourselves turning our heads back from right to left and back again, as if watching a tennis match. Refutation and counterrefutation are what we would call secondary argumentation, as distinct from primary arguments. These are the grounds offered by the debaters in direct support of their standpoints – i.e., the main merits of their own proposal, or the drawbacks of the opponent’s. Throughout, the front-broadening debater introduces topics of disagreement that are not necessary to elucidate the disagreement at hand.

The vote-shifter, on the other hand, will argue so as to narrow the front, concentrating on the specific issue that separates the opponents. He will, for example, concede that the opponent has certain weighty arguments, but he will then try to show that his own arguments are weightier. He will typically narrow or demarcate his claim, stating, for example, that he does not advocate a federal superstate in Europe, but that he does strongly advocate a union of nation states for certain reasons. He will concentrate on his own primary grounds for his claim; for example, he will concentrate on the main reasons why he thinks the Amsterdam treaty is a good idea (or, if he is against it, a bad idea), and he will spend less energy refuting the opponent’s grounds, or counter-refuting the opponent’s refutations. We might add that this emphasis on primary grounds, rather than on refutation, is one point where our normative criteria, based on audience needs, differ from the norms for critical discussion.
Furthermore, the front-narrowing debater will treat his opponent with politeness and respect and avoid face-threatening attacks on his person, ethics, and competence. In all these manouevers, the debater seeks to find and preserve whatever common ground there is between the opposite sides, narrowing the front to what is absolutely necessary.
In terms of the traditional rhetorical appeals, the vote-gatherer will rely heavily on pathos and will, for instance, use Atkinson’s “claptraps” in abundance (Atkinson 1984). As is well known, Atkinson described two principal types of claptrap: the contrast, which is clearly a front-broadening feature, and the list of three, a schematic figure of great dynamism, known from ritual and folk literature. Both are clearly front-broadening devices to enhance the feeling of “us” against “them”. The use of these devices will help the vote-gatherer boost the partisan’s spirit and give the spectators a good show. The vote-shifter, in contrast, relies mainly on logos appeals and avoids devices that may appear cheap or facile. As for ethos, the vote-gatherer will tend to impress by being either sparkling or passionate, while the vote-shifter tends to be a more academic type, perhaps slightly stiff and dry, but serious and knowledgeable.
All in all, it is clear that of these two types of argumentation, the vote-gatering, front-broadening type is by far the more “telegenic”, as media people say. This brings us to the role of TV in public debate.

Now, our point in contrasting the two types is of course not that debaters should become pure vote-shifters and never try to be vote-gatherers. Surely good debaters are those who manage to combine elements from both strategies. Nor do we claim that vote-gathering is bad rhetoric at all times. Many situations call especially for vote-gathering; but issue-oriented debate does not. The problem is that many forces in modern TV-mediated democracy unite in suppressing the kind of political argument that aspires, and inspires, to vote-shifting debate. TV debates, when best, are both entertaining and informative. But at times there is a conflict. What works well as TV is often front-broadening features that leave little opportunity for shifting rhetoric to unfold; what boosts and entertains partisans and spectators often alienates the deliberating citizen looking for guidance. In consequence, the media furthers the transformation of citizens to a body of, in Jamieson’s words, viewers “observing the ‘sport’ of politics” (Jamieson 1992: 191).
Front-broadening, vote-gathering TV debates thus appear to be the modern version of sophistic rhetoric. Sophistic debate is basically a type of combat, with debaters in the role of gladiators, in Gurevitch and Blumler’s term. Such a debate may serve a mobilizing  purpose for us if we are partisans of the gladiators, but that role easily slips into the purely spectatorial role where debaters are as much actors, at whose performance we either applaud or hiss. This audience role echoes Aristotle’s description of the auditor as “spectator” in epideictic speech, vs. the role as “judge” in political and forensic speech. According to Aristotle, the spectator is concerned with the ability of the speaker (Rhetoric III, 1358b). The spectator, as George Kennedy explains, “is not called upon to take a specific action, in the way that an assemblyman or juryman is called upon to vote”; the whole event becomes “an oratorical contest” (p. 48, note 77) – which is also how commentators see it when they discuss which politician “did best” in a TV debate. Thus the deliberative function of debate is suppressed by the simplistic question, so dear to the media, of “who loses and who wins”. While spectators see such debates as a sports event, its effect on partisans may be described in the words of Perelman & Olbrechts-Tyteca on the epideictic genre: “the argumentation in epidictic discourse sets out to increase the intensity of adherence to certain values” (1969: 51).

What is problematic with the spectator and partisan roles according to the deliberative ideal is that they tend to turn the audience into mere bystanders rather than participants in the political process. Only as deliberating citizens do we become a genuine rhetorical audience in Bitzer’s sense of the word – an audience of decision- makers, “capable of being influenced by discourse and of being mediators of change” (Bitzer 1968, 1992: 7).
We may compare our view here with Walton’s pragmatic approach: Walton is critical of debaters who have fixed positions, so that there is no “genuine chance of either side persuading the other” (1992: 157). However, Walton ignores the trialogic nature of debate, which makes it quite acceptable for debaters to be unwilling to be persuaded by each other. What threatens the legitimacy of debate is when it is conducted in such a way that there is no chance of anyone in the audience shifting to the other side.

To sum up, what we advocate in issue-oriented debate is that vote-shifting argumentation be allowed to unfold – i.e., argumentation strongly characterized by the features we have called front-narrowing. The purpose of course is not the shifting of voters as such. We call for more vote-shifting argumentation for normative reasons.We propose that if debaters argue with the shiftable voters on the opposite side as their primary addressees, this would stimulate them to produce convincing argumentation, i.e., arguments that those on both sides of the boundary who recognise the force of argument would consider weighty – whether they are persuaded by them or not. Thus, the deliberative goal would not be lost, namely that of providing citizens with the best arguments on both sides, to be weighed against each other, in order to reach a decision. The net result at the polling station would perhaps be pretty much the same. But decisions would be made on a firmer basis, and debates would better serve the purpose of informed political argument. They would not degenerate into mere sports events for spectatators or peptalk for partisans, and citizens might remain active participants in the political process.

REFERENCES
Aristotle. On Rhetoric: A Theory of Civic Discourse. Ed. George Kennedy (1991). New York: Oxford University Press.
Atkinson, Max. (1984). Our Masters’ Voices: The Language and Body Language of Politics. London: Methuen.
Bitzer, Lloyd F. (1968). The Rhetorical Situation. Philosophy and Rhetoric 1,1-14 (Reprint Supplementary Issue 1992).
Dieckmann, Walther (1981). Politische Sprache, politische Kommunikation: Vorträge, Aufsätze, Entwürfe. Heidelberg: Carl Winter Universitätsverlag.
Eemeren, Frans H. van & Rob Grootendorst (1992). Argumentation, Communication, and Fallacies. A Pragma-Dialectical Perspective. Hillsdale, New Jersey:: LEA.
Gurevitch, Michael and Jay G. Blumler (1977). Linkages between the Mass Media and Politics: a model for the analysis of political communication systems. James Curran, Michael Gurevitch, and Janet Woollacott (eds.) Mass Communication and Society (pp. 270-290) London: Edward Arnold.
Jamieson, Kathleen Hall (1992). Dirty politics: Deception, Distraction, and Democracy. New York: Oxford University Press.
Jørgensen, Charlotte (in press). Public Debate: An Act of Hostility? Argumentation.
Jørgensen, Charlotte, Christian Kock, & Lone Rørbech (1994). Retorik der flytter stemmer: Hvordan man overbeviser i offentlig debat. Copenhagen: Gyldendal.
Jørgensen, Charlotte, Christian Kock, & Lone Rørbech (1998). Rhetoric That Shifts Votes: An Exploratory Study of Persuasion in Issue-Oriented Public Debates. Political Communication 15, 283-299.
Klein, Josef (1991). Zur Rhetorik politischer Fernsehdiskussionen.
Ueding, G. (ed.). Rhetorik zwischen den Wissenschaften. Geschichte, System, Praxis als Probleme des “Historischen Wörterbuchs der Rhetorik (pp. 353-362). Tübingen: Max Niemeyer.
Perelman, Chaim, & Olbrechts-Tyteca, Lucie. (1969). The New Rhetoric: A Treatise on Argumentation. Notre Dame: Notre Dame University Press.
Tonsgaard, Ole. (1992). A Theoretical Model of Referendum Behaviour.
Gundelach, Peter, & Karen Siune (eds.). From Voters to Participants. Esaays in Honour of Ole Borre. Århus: Politica.
Walton, Douglas N. (1989). Informal Logic: A Handbook for Critical Argumentation. Cambridge: Cambridge University Press.
Walton, Douglas N. (1992). Plausible Argument in Everyday Conversation. Albany: State University of New York Press.
Walton, Douglas N. (1995). A Pragmatic Theory of Fallacy. Tuscaloosa: The University of Alabama Press.

The single most important thing to know about the pragmatics of argumentation is that argumentation is a kind of conversational expansion, a form of repair that kicks in when triggered by a special sort of event. Discourse occurs before a very dense backdrop of assumptions, assertions, and implications, not all of which can be examined for their acceptability or justifiability. Whenever any of us speaks, we evoke for our hearers an indefinitely expandable context of belief and claim, any part of which may be called out and made arguable. Most of what we say, and especially most of what we evoke, passes without close examination.
This willingness to let things pass without examination, though essential to the organization of conversation, is antithetical to what is commonly called “critical thinking.” In educational contexts, at least, we might suppose that what we want is for students to be constantly engaged in reviewing each proposition advanced and considering whether it is to be believed or not. Realizing, however, that a speaker’s “standpoint” is not simply what is asserted but also what must be believed in order to have made that assertion and to have made it in the circumstances in which it was made, we see that it is not in fact possible for students to inspect everything. Like all of us in all contexts, they must pick and choose among propositions to examine. In the classroom as in conversation, most statements pass without inspection.
This paper is about designing discourse for the support of argumentation, both in the sense of stimulating its occurrence and in the sense of regulating its conduct. Argumentation is valuable in educational contexts, and although I do not expect this point to be controversial, I will begin by reviewing in the first section some of what is known about the relationship between argumentation and learning. Unfortunately, however valuable argumentation may be, it is also interpersonally complex, implicating not just our beliefs about impersonal things but also our “standing concerns” for identity, status, and relationship (Jacobs, Jackson, Stearns & Hall 1991). In status-marked settings like the classroom, these interpersonal complexities can create intractable dilemmas for the structuring of argumentation, a point to be elaborated briefly in the second section of the paper. Employing a design methodology described briefly in the third section of the paper, I will describe several explicitly theorized plans for the incorporation of argumentation into teaching and learning. In this respect, the present paper is an instance of the form of practical research my colleagues and I championed in Reconstructing Argumentative Discourse (van Eemeren, Grootendorst, Jackson & Jacobs 1993): research organized by the search for argumentation procedures that take into account the situation of argumentation within real-world constraints and limitations.

1. Contributions of Argumentation to Learning
Argumentation here refers not to preparation of an essay or speech that makes a case for a proposition, but to critical engagement in dialogue or dialectic – an interactive, collaborative process. Since the publication of Toulmin’s landmark study The Uses of Argument (1958) theorists have recognized that argumentation unfolds as an answer to questioning, doubt, or contradiction. In contemporary argumentation theory (van Eemeren & Grootendorst 1983; Willard 1989), central importance is assigned to interaction and to the social context in which it occurs. Argumentation’s interactional function – the resolution of disagreement – demands discourse forms in which anything that might be contested can be “externalized” and addressed (van Eemeren, Grootendorst, Jackson & Jacobs 1993). Argumentation expands around disagreement (Jackson 1987; Jackson & Jacobs 1980).
Argumentation is known to contribute to learning across a broad spectrum of educational levels and subjects (Bruffee 1992; Kuhn 1993; Kuhn, Shaw & Felton 1997; Pontecorvo 1993; Meyer & Woodruff 1997; Voss 1991; Zeidler 1997). Argumentation stimulates deeper processing and more critical thinking, and when it is incorporated into instruction it helps students learn. For example, Kuhn, Shaw, and Felton (1997) developed a teaching/learning design in which students met and engaged in discussion on a single topic with peers holding diverse positions over a 5-week test period. As compared with a control group that only had to state an opinion on the topic and write a justification of their opinions at the beginning and at the end of the experimental period, the group engaging in argumentation with others achieved superior topical insight and superior argument quality.

What accounts for the difference in learning? More is involved than the effect of thinking about the topic and writing about it. All students went through these processes. Kuhn et al. did not simply sort students into random pairs but arranged the dyads so as to guarantee encounter with a wide range of discrepant and congruent positions, so that students would be sure of meeting disagreement. Other designs that putatively rely on argumentation, but that fail to ensure controversy, have not had the same effect on learning. For example, Marttunen (1992) found instruction organized around comment on written argumentation to be less effective than “traditional” instruction, but since no mechanism was provided to assure clash of viewpoints, the argumentation design may have omitted its active ingredient.
We know that encountering disagreement stimulates the search for fallacy and other weakness in argumentation, that people are much more competent at evaluating arguments for conclusions they disagree with than at evaluating arguments for conclusions they agree with. Experimental research on “biases” in reasoning (Klaczynski 1996, 1997) has shown that the quality of reasoning and evidence is unlikely to be thoroughly evaluated if the conclusion happens to be congruent with one’s own beliefs. By contrast, disagreement stimulates the search for what is wrong in others’ reasoning and what is needed to bolster one’s own reasoning against challenge (Jackson 1996).
Ideal models of argumentation (e.g., van Eemeren, Grootendorst, Jackson & Jacobs 1993) treat the externalization of contradiction and the expansion of discussion around points of contention as fundamental to rationality. To encourage pervasive occurrence of argumentation, our first requirement is to provide for externalization of disagreement. Contradiction and confrontation should be emphasized and exploration of the grounds for belief and disbelief should be expanded. Externalization is not simply a matter of requiring that students write position statements of their own, but a matter of guaranteeing that each student wrestle with positions discrepant from their own. This will be a key feature of every successful design for argumentation in learning.

2. The Interpersonal Complexity of Argumentation
So why not simply contradict everything students say in the style of Monty Python? Unfortunately, merely confronting speakers with contradictions does not assure critical discussion. (Yes it does. No it doesn’t.) The possibility of critical discussion is also known to rest on various levels of preconditions, including most obviously the abilities and motivations of the arguers and the social and political circumstances surrounding the argument.
Conditions known to threaten critical discussion include artificial limitations on participation, limitations in individual ability, personal identity concerns, and hierarchical social relationships-all of which play prominent roles in classrooom communication. Participation in classroom discussion is generally infrequent and uneven at the postsecondary level (Karp & Yoel 1976; Nunn 1996), with a few individuals accounting for the bulk of student contributions. While the overall level of student participation is linked to instructional design decisions, which individuals in a group participate is linked to gender, self-confidence, and other individual difference variables (Fassinger 1995). Social norms may inhibit expression of controversial opinions or extended argumentation (Fassinger 1995; Lusk 1994), while deference to the authority of the teacher may suppress the occurrence of disagreement or lead to premature closure of debate.
In other words, argumentation is interpersonally complex, having not only intellectual dimensions but also highly-charged relational dimensions. Disagreement is often experienced as threatening, especially under conditions of unequal power or authority; contradiction or challenge by authority figures often simply closes down discussion. Among peers, argumentative exchange has a competititive quality that can make it difficult for arguers to change their minds once committed to a position.
Some of these threats can be handled through sensible design decisions, whether in traditional classrooms or in virtual environments. For example, the dyadic argumentation procedure developed by Kuhn et al. was designed to guarantee controversy by pairing students with others holding discrepant views, and it was further designed to minimize deference by forming peer dyads rather than teacher-student dyads. Knowing that specifiable characteristics of the social situation may suppress argumentation, we can design those characteristics out of the interaction, using whatever resources come to hand. To the extent that interpersonal complexity threatens the occurrence or quality of argumentation, the successful integration of argumentation into teaching and learning will depend on management of its interpersonal complexity.

3. Design Methodology within Normative Pragmatics
We might or might not be able to make students indifferent to authority, identity, and peer pressure. Normative pragmatics accepts the circumstances of ordinary discourse and searches for ways to regulate their impact on argumentation, employing a design methodology adapted to its general theoretical program (van Eemeren, Grootendorst, Jackson & Jacobs 1993). Normative pragmatics approaches the study of argumentation empirically, but with questions motivated by normative considerations and with analytic tools tailored to criticism and intervention. Argumentative practices are examined with an eye to their improvement. The blending of empirical and normative considerations is made explicit in our design methodology.
This design methodology has four components: an empirical examination of discourse practices, a critical analysis based on comparison of practices with an ideal model, a specification of designable features, and a proposed redesign.
Empirical analysis of discourse practices is aimed at developing conjectures about participant goals and about the obstacles participants face in accomplishing these goals. Often this analysis involves direct inspection of records of interaction, but empirical analysis may also extend to experimental investigation of communication behavior and outcomes. In the present case, our focus is on the occurrence of argumentation and on the impact of its occurrence on learning. This being a topic of very active concern, there is a rich literature that documents such facts as the uneven application of critical standards to congruent and discrepant positions, the general social inhibitions against disagreeing, especially with authority, and the unevenness of participation from student to student. In other contexts our central concern might be for management of relevance or for regulation of the impact of authority; in the discourse of teaching and learning, our first concern is for conditions that limit the very occurrence of argumentation.
Neither our participants (teachers and students) nor the conditions under which they interact are ideal. In ideal critical discussion (van Eemeren, Grootendorst, Jackson & Jacobs 1993), arguers engage in full, free, and impersonal exploration of potential disagreement without limitations on either total talk time or rights to speak. In ideal critical discussion, the contestability of every proposition is fundamental and participants are expected to shoulder a “burden of rebuttal” rather than to let potentially controversial points pass. Not all classroom discussion falls far short of this ideal, but much does.
A specification of potentially designable features will normally be grounded in comparison of actual empirical circumstances with conditions defined by ideal models. Against an ideal standard of full, free, impersonal explorations of ideas, certain features of the classroom situation present themselves as possible “culprits”: finite talk time, unequally distributed speaking rights, unequally distributed authority, identity-relevance of speech, and so on. From these noticed features we begin the process of designing discourse to encourage rather than discourage argumentation. To the extent that they are malleable, we can alter them through design and document the result.
Gaps between ideal models and actual practices present opportunities for engineering of argument. We search for ways to eliminate, compensate, or work around design features that promote bad practices and to inject or emulate design features that promote good practices. In the discourse of teaching and learning, with a first objective of simply increasing the occurrence of argumentation, we must find ways to minimize the impact of authority and identity, and also, of course, scarcity. One of many ways to do this is through invention of what we are calling ‘discussion protocols.’

4. Argumentation Protocols for Teaching and Learning
The trick in designing plans for argumentation in instruction is to preserve argumentation’s cognitive advantages while managing its interpersonal complexities. Let’s begin by trying to devise an all-purpose argumentation protocol to use in teaching physics. The role of argumentation will not be to arrive at resolution of disagreement, but to exploit disagreement to induce deeper thinking about problems whose answers are known. So presumably what is wanted is a method for moving a student from a wrong answer to a right answer through exposure of incorrect assumptions or faulty reasoning.
A useful device that meets this challenge is the ‘confrontation sequence’ in which less sophisticated ways of thinking are brought into confrontation with predicaments that call for more sophisticated reasoning. In a confrontation sequence (Bleiberg & Churchill 1975; Jacobs 1986), one speaker (the confronter) helps another (the confronted) to recognize weaknesses or self-contradictions by calling out commitments one at a time and juxtaposing those that are in contradiction – a straightforward dialectical structure. The confrontation sequence has three ‘stages’: an opening in which some statement triggers a decision to confront; an exploration in which question/answer pairs or challenge/response pairs establish commitments; and a punchline or predicament in which the confronter draws out the contradiction or inconsistency in the confronted’s various commitments.
1. Statement
2. Exploration (Challenge/Response, Refutation/Concession, Question/Answer)
3. Predicament

Confrontation might prove very useful in teaching if deployed in such a way as to bring less sophisticated ways of thinking into dilemmas that motivate progression to more sophisticated reasoning. However, by its very design the confrontation sequence exacerbates the conditions that seem to suppress the occurrence of argumentation in the classroom. Its oppositional structure is corrective rather than collaborative, and the final predicament, the punch line, puts the confronted ‘on the spot,’ compelled to respond and unable to do so without repudiating something previously asserted. The classic confrontation subjects a student’s reasoning to public critique and potential loss of face.
The feature we want is opposition. The features we don’t want are the face implications associated with being in the public role of the confronted – what an interaction analyst might call a ‘one-down’ position. A skillful teacher can find ad hoc strategic solutions to how to confront without face threat, but it is also possible to design structures of this kind that are independent of the skill of the confronter.
My own design work has depended heavily on computer mediation of dialogue. Computer mediation allows for asynchrony in interaction (meaning that people can engage in conversational exchanges without being in the same place at the same time) and for a high degree of individualization (meaning that what a teacher says to students can be tailored differently to each one). However, for purposes of managing the interpersonal complexity of argumentation, the most important attribute of computer mediated communication is that it allows for anonymity. Students can be engaged, through interactive computer technology, in argumentation with anonymous others whose characteristics are known only through what they write or through what is written about them.

One of my tasks at the University of Arizona over the past several years has been to design tools to support instruction on the worldwide web, and in particular to design tools that allow for incorporation of argumentation into web-based instruction.  I’ve createdand implemented a web course authoring system known as POLIS, most of whose capabilities are not relevant to the present discussion. What is relevant within POLIS is the repertoire of argumentation protocols offered to instructors to assist them in using argumentation effectively. Instructors in any subject use POLIS to create online argumentative dialogues for students to use as “lessons.” Shortly I’ll have to produce evidence that the POLIS repertoire has measurable impact on learning; POLIS is collecting data on itself every time an instructor creates an online lesson or a student submits a response to it. What I can give so far is a progress report on the creation of the learning protocols themselves.
Unlike otherwise comparable systems of web authoring tools, POLIS is highly theorized. Its protocols can be described structurally in terms of speech act sequences, and the structures it generates are heavily influenced not only by speech acts theory but also by those strands of discourse analysis that have been concerned with conversational sequencing and conversational expansion. I want to describe and contrast three POLIS protocols (Recitation, Adversary, and Virtual Peer) to illustrate the way in which features known to affect argumentation can be managed at a structural level. (The entire web kit is open to public examination at http://emma.comm.arizona.edu.)
Standard classroom recitations have three moves: question, candidate answer, and assessment. The teacher poses a question, a student answers, and the teacher either affirms the answer or, if it is incorrect, offers a correction. The most interesting answers are the wrong ones; those are the opportunities a teacher could use to initiate confrontations or other more obviously argumentative processes. POLIS makes a very slight improvement over the standard form of recitation, presenting not an authoritative assessment but a “model answer” which the student uses to make a self-assessment. So the POLIS Recitation have four moves: question, candidate answer, model answer, and self-assessment. Notice how this minor variation affects the overall quality of the exchange: the standard recitation closes the sequence with assertion of an authoritative answer, while the POLIS Recitation invites expansion around any difference between the submitted answer and the model answer. Though not designed specifically for argumentation, the POLIS Recitation illustrates an important point about protocol design, that the interactional sequence and the framing of contributions might matter.

POLIS offers a much more explicitly argumentative protocol, known simply as Adversary. Adversary builds and conducts online debates with students. It has a minimum of six moves:
1. Statement of controversy (by teacher, via POLIS)
2. Statement and defense of [initial] standpoint (by student)
3. Statement and defense of opposing standpoint (by POLIS)
4. Rebuttal of opposing standpoint (by student)
5. Invitation to reconsider (by POLIS)
6. Statement and defense of [terminal] standpoint (by student)

The two middle turns, a counterargument/rebuttal pair, can be repeated for additional counterarguments. POLIS selects what to present at that step using the student’s initial position as data. Adversary is an automated system and (because it is built to deal with any subject, not with some fixed body of content) it has no knowledge base to use in planning its contributions. Its opposing arguments are chosen from a store supplied by the teacher or by previous students. However, it allows for an online simulation of the sort of experience students might have had in the Kuhn et al. experiment reviewed earlier. Students are presented with one or more arguments against their own initial positions and must answer these before making a final decision on the controversy. Important features to notice are the open-endedness of the sequence (no suggestion that the controversy is in fact settled) and the use of disagreement per se to motivate deeper reflection on the controversy.
In use, Adversary appears to function also as a kind of modelling exercise for students; their defenses of their initial positions frequently give elaborations of their personal beliefs rather than justfications for those beliefs, but when presented with models of argumentation in the counterargument passages they quickly accommodate to the normative requirements of the exchange.
The last of the three protocols considered here is modelled after a very sophisticated design used in physics instruction (Mazur 1997). In its classroom version, argumentation takes place synchronously between peer dyads within a large group. The teacher presents a problem, each student develops an individual answer and then tries to persuade a neighbor that their answer is correct, and then the correct answer is shown and explained.

The online version within POLIS, known as Virtual Peer, differs from both Recitation and Adversary in terminating with a correct answer to a question. It has a minimum of seven moves:
1. Statement of problem (by teacher, via POLIS)
2. Candidate answer and explanation (by student)
3. Proffering of alternative answer/explanation (by POLIS, presented as peer reasoning)
4. Response to peer reasoning (by student)
5. Invitation to reconsider (by POLIS)
6. Final answer and explanation (by student)
7. Presentation of correct answer and explanation (by POLIS)

Again, the middle subsequence is selected for discrepancy with student’s own position, and it can be repeated as many times as necessary to work through all of the alternative positions presented to students at the first step. Virtual Peer is explicitly argumentative, despite the existence of a correct answer known in advance. This protocol more than any other draws attention to the role argumentation can play in teaching and learning, forcing deeper examination of the reasoning behind even correct answers. Students who get the problem right on the first try have the same sequence of argumentative tasks as students who get the problem wrong on the first try. And importantly, this is framed in such a way as to carry no implication that the counterconsiderations are reasonable: Students get discrepant positions represented as what another classmate argued. (Compare this with another common strategy for probing the reasoning behind a correct response: Devil’s advocacy by the teacher.)
Even in online protocols, it should be noticed that interpersonal considerations must be managed. Recitation and Virtual Peer differ most significantly in the framing of counterconsiderations presented to the student. Recitation presents a model answer to be used by the student as a standard for his or her own writing. Virtual Peer presents alternative answers treated as equal competitors to the student’s own answer, enjoying no presumption grounded in the teacher’s authority. Empirically, students write more in response to the counterconsiderations of Virtual Peer than they do in response to the model answer of Recitation. The pragmatics of Recitation favor narrow self-assessment (“My answer did not mention conditional probability”) while the pragmatics of Virtual Peer favor argument criticism (“This answer looks reasonable at first, but …”).
Argumentation protocols of these kinds appear to be effective in both promoting more argumentation and in leading students to think more critically about their own reasoning. Since POLIS captures student responses pervasively, it is possible to review the arguments students make at the beginning and end of an argumentation sequence and to note the quality of argumentation offered. Although in any given online debate, relatively few students change their positions, many show progression toward more critical examination of evidence. For example, in one application of the Adversary protocol, students were asked to use statistical summaries of their classmates’ codings of a presidential address to decide whether the speech was or was not ‘liberal.’ Initial responses tended to treat the statistical material uncritically: some students argued that the speech was liberal because over half of its paragraphs contained liberal themes, while others argued that the speech was not liberal because the split between liberal paragraphs and neutral/conservative paragraphs was too even. However, after being presented with arguments that challenged the validity and interpretability of the coding, many students wrote position statements that dealt explicitly with the quality of evidence and offered independent grounds for an overall judgment of the speech. At the low end of sophistication, these responses simply exhibited awareness that seemingly scientific evidence might or might not be trustworthy, as in these unedited examples:
1. the speech is liberal. however, if there is confussion in the coding of the document then the results are not reliable. if there is no confussion then the results are correct and the majority of the speech is liberal.
2. I changed my mined because of the last argument concerning the point that there are no reliable grounds because of the statistics about the coding of everyone’s opinions. It is too hard to determine what type of speech it reflected because the results were all so different. At the high end, students were able to transcend the original terms of the problem and challenge the relevance of the evidence given, as in the following excerpt from a student answer:
3. This speech cannot be deemed liberal, as it has the presence of strong conservative assertions as well as weak, or rather, mild liberal statements. While one many deem this liberal using only the micro and statistical view, I believe strongly that one must take the text as a whole into account. The overall essence of teh text is …

Only with accumulation of more data for other uses of these advanced protocols will we be able to thoroughly analyze their impacts on learner outcomes, but the promise in both protocols is clear. Our limited experience to date shows that it is possible to create challenging online dialogues with the capacity to engage students in higher-order reasoning, especially self-criticism and critical evaluation of evidence and reasoning for a position.

5. Conclusion
Individuals vary greatly in their tendency to examine what is said and in their willingness to call out potential arguable threads. This tendency is variously described in terms of “critical thinking ability,” “need for cognition,” or plain “argumentativeness.” At least the first of these is often considered an important intellectual skill, something to be cultivated through education. Important in and of itself, critical thinking is also the means by which students come to deep understandings of any subject.
However, critical thinking needs cultivation in argumentative practice. It might be better to say that critical thinking is itself a form of argumentative practice. Encountering disagreement and interacting with an informed antagonist is the surest way to trigger “central processing.” For this reason, it is worthwhile to build designs that inject disconfirmation, contradiction, and confrontation into teaching and learning dialogues and that do this in a fashion that limits the interpersonal consequences of disagreeing.
In experience to date with online argumentation protocols, we have found it useful to differentiate designs in terms of their capacity to expand around disagreement and in terms of the distribution of authority they presume. While computer technology is in no way essential to the incorporation of argumentation into teaching and learning, it does provide very convenient means for managing these important design features. In particular, it solves in a very generic way many of the dilemmas associated with the interpersonal complexity of argumentation.

REFERENCES
Bleiberg, S. & L. Churchill (1975). Notes on confrontation in conversation. Journal of Psycholinguistic Research 4, 273-278.
Bruffee, K. (1992). Science in a postmodern world. Change, 24(5),18-26.
Eemeren, F. H. van & R. Grootendorst (1984). Speech Acts in Argumentative Discussion. Dordrechts-Holland: Foris.
Eemeren, F. H. van, R. Grootendorst, S. Jackson & S. Jacobs (1993). Reconstructing Argumentative Discourse. Tuscaloosa, AL: U. of Alabama Press.
Fassinger, P. A. (1995). Understanding classroom interaction: Students’ and professors’ contributions to students’ silence. Journal of Higher Education 66, 82-96.
Jackson, S. (1987). Rational and pragmatic aspects of argumentation. In: F. H. van Eemeren, R. Grootendorst, J. A. Blair & C. A. Willard (Eds.), Argumentation Illuminated (pp. 260-269), Amsterdam: SICSAT.
Jackson, S. (1996). Fallacies and heuristics. In: J. van Benthem, F. H. van Eemeren, R. Grootendorst & F. Veltman (Eds.), Logic and Argumentation (pp. 101-114), North-Holland: Royal Netherlands Academy of Arts and Sciences.
Jackson, S. & S. Jacobs (1980). Structure of conversational argument: Pragmatic bases for the enthymeme. Quarterly Journal of Speech 66, 251-265.
Jacobs, S. (1986). How to make an argument from example in discourse analysis. In: D. G. Ellis & W. A. Donohue (Eds.), Contemporary Issues in Language and Discourse Analysis (pp. 149-168), Hillsdale, NJ: Erlbaum.
Jacobs, S., S. Jackson, S. Stearns & B. Hall (1991). Digressions in argumentative discourse: Multiple goals, standing concerns, and implicatures. In: K. Tracy (Ed.), Understanding Face-to-Face Interaction: Issues Linking Goals and Discourse (pp. 43-62, Ch. 3), Hillsdale, NJ: Erlbaum.
Karp, D. & W. Yoel (1976). The college classroom: Some observations on the meanings of student participation. Sociology and Social Research 60, 421-439.
Klaczynski, P. A. (1996). Self-serving influences on adolescents’ evaluations of belief-relevant evidence. Journal of Experimental Child Psychology 63, 317-339.
Klaczynski, P. A. (1997). Bias in adolescents’ everyday reasoning and its relationship with intellectual ability, personal theories, and self-serving motivation. Developmental Psychology 33, 273-283.
Kuhn, D. (1993). Science as argument: Implications for teaching and learning scientific thinking. Science Education 77, 319-337.
Kuhn, D., V. Shaw & M. Felton (1997). Effects of dyadic interaction on argumentative reasoning. Cognition and Instruction 15, 287-315.
Lusk, A. B. (1994). Discussing controversial topcis in the classroom: Creating a context for learning. Teaching Sociology 22, 301-308.
Marttunen, M. (1992). Commenting on written arguments as a part of argumentation skills – comparison between students engaged in traditional vs on-line study. Scandinavian Journal of Educational Research 36, 289-302.
Mazur, E. (1997). Peer Instruction: A User’s Manual. Upper Saddle River, NJ: Prentice Hall.Meyer, K. & E. Woodruff (1997).
Consensually driven explanation in science teaching. Science Education 81, 173-192.
Nunn, C. E. (1996). Discussion in the college classroom: Triangulating observational and survey results. Journal of Higher Education 67, 243-266.
Pontecorvo, C. (1993). Social interaction in the acquisition of knowledge. Educational Psychology Review 5, 293-310.
Toulmin, S. E. (1958). The Uses of Argument. Cambridge: Cambridge University Press.
Voss, J. F. (1991). Learning to reason via instruction in argumentation. Learning & Instruction 1, 337-350.
Willard, C. A. (1989). A Theory of Argumentation. Tuscaloosa, AL: University of Alabama Press.
Zeidler, D. L. (1997). The central role of fallacious thinking in science education. Science Education 81, 483-496.

1. The Problem
My primary goal today is to introduce you to a pioneering project undertaken to see how extensive mapping of arguments can be accomplished and whether such a mapping would be useful to students, teachers, and scholars. Why would you want to map an extensive argument? Let me start with a hypothetical story. In the 1930s Alan Turing, the great British mathematician, invented the ideas on which the modern computer is based. In 1950, he wrote, “I believe that at the end of the century the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted.” He certainly thought the computers would be able to think. However …. there are less than two years left before the end of the century. Unfortunately, Turing died in 1954 at the age of 42. Suppose he came back from the dead after 44 years to find out whether his prediction had come true. Suppose he asked you, “What’s happened since I died? Was I right? Does everybody agree that computers can think?”

In the first place we could tell him that he certainly should have expected to be contradicted. That almost 400 scholars have engaged in a 48-year argument that he started. That the argument was worldwide. That it has taken place in almost 300 journals and books and consists of more than 800 major “moves” – claims and rebuttals and counterrebuttals. He would find out that some of the greatest physicists, philosophers, computer scientists, and psychologists in the contemporary world have taken part in it.
Suppose Turing said: “Right now I don’t have time to read 300 journals and books. What is the status of the argument? Where does it stand now?” Stop for a moment. How would you give him a serious answer? Suppose you managed to answer his question. Then suppose he asked another: “Where can I get an overview of the history of the arguments so I can decide which I want to read.” Where would you direct him?

2. The Problem As We Saw It
For great debates like this one about machine intelligence, there is:
– no comprehensive map of this major debate
– no way to get an up-to-date briefing on its current status
– no way to link positions to rebuttals (so that proposed refutations of data and positions can be easily compared)
– no efficient way to navigate through the argument
– no way to visually inspect its structure and direction

These are also the problems of every beginning student in any major subject-matter debate. And these problems are not only true of the artificial intelligence debate but also of most of the great discussions in which humanity is involved. While the argumentation maps I will talk about today show the substance of this decades-long, worldwide debate, I will not so much focus on the substance of that particular argument. Rather, I want to discuss with you the argumentation analysis format we developed, the implications that our maps have for the study of argumentation analysis, the problems we encountered, and the kinds of solutions we came up with.
I should add an historical note here. Credit must go to Stephen Toulmin who, as far as I know, developed the modern ideas of argumentation analysis in 1957. I worked in the mid-80s on a variety of graphic approaches to mapping extensive argumentation. A chapter of my 1989 book, Mapping Hypertext, is devoted to the progress I made. But in the end I felt I hadn’t quite got a useful enough approach. Four-and-a- half years ago I took up the problem again when I went to Stanford.We wanted to map a major philosophical argument. The debate Turing started qualifies, as it involves one of the major questions about which human beings puzzle, worry, and debate – our identity. Who are we? While the history of this debate goes back at least to Hobbes, Leibniz, and Descartes, as I said, the modern debate starts with Turing’s 1950 article in the journal Mind.
Our goal was to map a whole argument and a big one, not some little piece of a broader debate. The Turing argument was an ideal choice. It turned out to be an even bigger challenge than I thought. The debate was far more extensive than I knew or than what usual book-length summaries indicated. But it was an ideal testbed for the visual methodology we were developing. And we faced the challenge of designing a useful tool–simple enough to be educationally sound yet detailed enough to help scholars.

What do the maps look like? Figure 1 shows a complete map and figure 2 shows detail.

3. Basic Structure of Argumentation Maps
The basic framework of our mapping generally follows Toulmin.

3a. Major topics of the debate
One of the consequences of our taking on such a large, sprawling argument was that we needed to subdivide it into different issue areas. Debates frequently divide into topic areas which can be shown as regions in the mapping of the debate by putting them all together in one area and giving them a title. The example here shows the initial claims boxes of three regions, identified with the questions in bold face. Table 1 lists the issue areas. It shows the breadth of the more than 50 philosophical issues that have become involved in the debate over Turing’s question. It provides a kind of table of contents or subject index of the issues. Within each issue area, the arguments are presented chronologically.

3b. Focus box
The focus box introduces and summarizes the core dispute of each issue area, sometimes as an assumption and sometimes as a general claim with no particular author. The lowest-numbered box in each issue area is an introductory focus box.

3c. Claims
Debates start with claims, which have been defined by Toulmin as “assertions put forward publicly for general acceptance with the implication that there are underlying ‘reasons’ that could show them to be ‘well founded’ and therefore entitled to be generally accepted.” (Toulmin et. al., l979) Claims as we have written them are brief summaries, often accompanied by explanatory illustrations.
Some readers have been thrown off in expecting the claim boxes to be abstracts of published works. But claims summarize individual arguments. As such, a given published article may be broken down into numerous claims on the maps; alternately, a given claim may draw on information in several published chapters and articles. Each claim is connected to the next by one of three links: supported by, disputed by, or interpreted as.

3d. Supported by
We defined the “supported by” relationship slightly differently than Toulmin. These are arguments that uphold or defend another claim. Examples include: supporting evidence, further argumentation, thought experiments, extensions or qualifications, and implemented models.

3e. Disputed by
These are charges made against another claim. Examples include: logical negations, counterexamples, attacks on an argument’s emphasis, potential dangers an argument might raise, thought experiments, and implemented models.

3f. Support and dispute carry a range of meanings
Support and dispute are used in an argumentative sense rather than in a strict logical or epistemic sense. They structure the map into chains of agreement and disagreement where claimants respond to one another in a variety of affirmative and negative way. As such, the relations of support and dispute cover a wide range of cases, which fall into “fuzzy categories” or “families” of supportive and disputative responses.

3g. Interpreted as
Sometimes an argument is reframed by one of the disputants. If there was a distinctive reconfiguration of an earlier claim, we used this icon.

3h. Anticipated by
Where this phrase appears in a box, it identifies a potential attack on a previous argument that is raised by the author so that it can be disputed.

3i. Links as arrows direct the eye
After experimenting with a number of formats, we decided to use arrows to show the paths of arguments, with icons showing whether the relationship was one of support, dispute, or interpretation. The directionality of a link, represented by an arrow, represents the direction in which the reader should read the claims for maximal effectiveness. The arrows direct the eye. Thus, links do not necessarily correspond to direct evidential support, logical negation, or any more crispy defined logical relation (though in particular cases a link may be any one of these). I should point out that we sometimes include what Toulmin would call grounds, warrants and backing in our claim boxes. We did this primarily to avoid more of a tangle of boxes and arrows than we already had.

3j. Rebuttals and counterrebuttals
The rebuttal presents the possible exceptions or objections to the claim. There is no such thing as a debate without at least one rebuttal. And we followed that guideline as a criteria for choosing arguments to map.
Debates then continue through a series of contributions that dispute previous claims and other rebuttals. The counterrebuttals may or may not be made by the original claimant.

4. What’s the Answer? Can Computers Think?
The argumentation maps do not attempt to evaluate the arguments summarized. They map the debate without taking a stand. They are, as much as possible, neutral. It is left to readers to be the jury, to evaluate the “weight” of the arguments and evidence and draw their own conclusions. Many students have been frustrated by this. Indeed many scholars who have seen the maps say, “So, what’s the answer?” The maps do not provide the answer. Hopefully they do not even reveal the mapmakers’ views.
Of course, the maps are to some extent interpretive. In writing and linking arguments, we had to condense incredible amounts of information, often on the basis of highly obscure or technical literature. We also had to make decisions about placement and emphasis. The way these maps organize the debate is not necessarily the only possible organization, but it was carefully considered and weighed against alternatives. The argument summaries themselves, which is where the real dialogue takes place, stick closely to the words of the authors, the better to avoid interpretation.

5. Criteria for Inclusion of Arguments
Over the course of the project, we have developed 11 criteria for deciding whether to include a particular argument.

5a. Use published arguments
Only those arguments were included that have been published in an established print or electronic medium: journals (including reputable electronic journals and white papers), magazines, and books. Arguments made in Usenet newsgroups, electronic forums, e-mail exchanges, or in interpersonal debate were excluded as too ephemeral and as representing positions still in development. Such arguments will be excluded until they appear in a more established medium.

5b. Use arguments that lie within the scope of the map
The major claim – that machines can or will be able to think – determines the scope of these maps. Many threads of argument drift away from the central issue into such related territories as the mind-body problem, functionalism, and the philosophy of science. Such claims were set aside until a chance arises to map neighboring territories with maps of their own.

5c. Seek out the historically earliest or best-known version of an argument
When different authors make similar arguments, we chose the version which was either historically earliest, or the best-known version of the argument. When the best-known version is used, the historically earliest version is usually mentioned in a note. In the few cases in which differing versions of an argument are sufficiently unique or separately disputed, each is summarized separately.

5d. Avoid loosely drawn arguments
Sometimes an author makes an argument loosely, at the end of a paragraph, as an aside, or in a footnote. In general, such arguments are not included unless they are developed further in follow-up articles or are the focus of further debate.

5e. Avoid repetitive, nitpicking, or duplicative arguments
One goal of the maps is facilitation of productive debate. Ad hominem arguments, redundant rounds of back-and-forth, and tediously nitpicky arguments were left out.

5f. Avoid forbiddingly technical discussion
Highly technical arguments, which are based on extensive symbolic notation and formalisms, could not be represented with the cartographic conventions we developed, or at the scale we chose to work at. However, summaries of many technical and symbolic discussions were included. Only the most forbidding had to be excluded.

5g. Summarize the author’s published claim
Many authors hold views today that are different from those they expressed at the time they entered into the debate. We include authors’ claims as published. If an author later changed his or her position, and published the change, the new claim was included and the change of position was noted. But if no new contribution has been made, then the original published view stands.

5h. Avoid tentative arguments
It became clear as we wrote the summaries of the arguments that one current, tentative style of academic writing made it extremely difficult to understand exactly what was being argued. In some way, authors had to be definitive in their arguments to qualify for a spot on the map. To use a geographical analogy, a road or a lake or a mountain that “may exist” is rarely mapped.

5i. Include some historical arguments
In order to properly situate the debate in its historical context, we included a sampling of notable historical supports of contemporary arguments.

5j. Include some experimental results
To situate the debate in a context of concrete experimental and computational results, we included some implemented systems and empirical results. Again, we only included a small sample of such results, sticking to famous and notable computer models and experiments.

5k. Include a small sample of outrageous and humorous arguments
Some of the stronger and stranger claims were worth including just to liven things up and have some fun. Such claims also provide “targets” for what we anticipate will be lively threads of response.

6. Why are argumentation maps important to teaching?
The biologist Lewis Thomas has written, “College students, and for that matter high school students, should be exposed very early, perhaps at the outset, to the big arguments currently going on among scientists. Big arguments stimulate their interest, and with luck engage their absorbed attention… But the young students are told very little about the major disagreements of the day; they may be taught something about the arguments between Darwinians and their opponents a century ago, but they do not realize that similar disputes about other matters, many of them touching profound issues for our understanding of nature, are still going on, and, indeed are an essential feature of the scientific process.”

This is the overarching reason that we created these maps – to illustrate for students the dynamic nature of a debate that is active today.

6a. Watching Contemporary, Interdisciplinary, Global Debates Unfold
The intelligent machines debate is a prime example of the type of argument that benefits particularly well from argumentation mapping. From its beginning, the debate has been a truly interdisciplinary and global discussion, with philosophers, cognitive scientists, artificial intelligence researchers, and others joining in, from around the world. Nevertheless, great parts of the debate have taken place in journals that are isolated by the boundaries of particular academic disciplines. As a result, it has been difficult until now to see the structure of the debate as it unfolds.
Argumentation maps provide a picture, more detailed than previously available, of how such a vast debate can take place across disciplinary and geographic distances. By creating an accessible map of the conceptual territory our hope is to facilitate more global interdisciplinary debate, to bring the various sources to light, and to illuminate how the pieces of the puzzle fit together. Perhaps the very existence of the maps will provide incentive and opportunity for more interdisciplinary and international discussion.
In a world of global interdisciplinary discussion, effective communication and productive dialectical exchange are key. We need to elevate the coffee-house discussions and the Usenet dialogues into cooperative and productive exchanges that push our understanding forward. It is all too easy to repeat an argument that has already been made in a distant or obscure location, to talk past one another in the heat of conflict, or to ignore important context. Moving a serious debate forward requires a disciplined interdisciplinary and international dialectic.

6b. Work with Great Minds
It is of benefit to students to observe and engage with great minds at work. The Can Computers Think? arguments have attracted some of the greatest and most subtle minds of the 20th century. I could mention Herbert Simon, Nobel Prize–winning economist; Kurt Gödel, who with Turing was among the greatest mathematical minds of all time; Roger Penrose the great physicist; John Searle, former president of the American Philosophical Association; Herbert Dreyfus, one of the world’s leading Heidegger scholars; John McCarthy, who named the field artificial intelligence; and many more whom I don’t have time to list here.

6c. Learning philosophy dialectically
Argumentation maps illustrate the value of learning philosophy dialectically. Individual arguments are presented in clear summary form and are followed by chains, or threads, of dispute and support. By watching philosophers lock horns and wrestle in an interdisciplinary arena of open debate, readers can better appreciate the subtlety and complexity of the issues they themselves are struggling with.
The dialectical method has ancient roots and remains valuable today. Thousands of years ago Socrates grappled with the best minds of Athens in public debate, and Plato recorded those dialogues as a means of teaching philosophical concepts. Today, contemporary issues are battled out in televised forums and in Internet newsgroups, where everyone from big-name pundits to coffee-shop philosophers chew through issues in round after round of back-and-forth. Argumentation maps harness the full communicative and instructional power of dialectical exchange.

7. More specific educational possibilities
How can these maps aid education and, in particular, education in argumentation analysis? I am sure that many of you will come up with creative uses that we on the project have never thought of. But here are a few possibilities, using the Can Computers Think? series as an example, that we would offer for your consideration.

7a. Excellent hook for student interest
It is easier to get into a subject that has some connection to currently hot topics in the culture. The maps can be used to introduce questions of philosophy in a way that is attractive and compelling. Many students will have heard of the IBM computer Deep Blue that recently beat the human grandmaster champion at chess. The chess-playing arguments are represented appear on Map 3, Can Physical Symbol Systems Think?

7b. Touches many subjects
One of the important things about the Can Computers Think? debate is that it touches on so many of the ongoing topics in philosophy : the mind-body problem, consciousness, free will, etc. This permits the instructor to show how one set of arguments relate to other sets of arguments in related areas.

7c. Provide project opportunities in creative argumentation
Since the maps provide the thread of existing arguments and also show where they have ended (as of now), they provide the opportunity for assigning students to select one thread or topic of an argument and try to add to it with an original argument, or write a critical essay about it, or read the original sources of one or more issue areas and critique them. Since the maps clearly mark the frontiers of arguments, students have a chance to engage in real debates and contribute their critical assessments as well as new arguments.

7d. Save time and provide context
One graduate student in the philosophy of mind said to us: “These maps would have saved me 500 hours of time my first year in graduate school. For almost two semesters, I had to keep reading article after article without enough context to see how they fit in to the bigger picture. The maps would have made my whole experience a much more rewarding one.”

8. Other Topics
We are proceeding on maps of several other major debates and have proposals out for still others. We believe that this mapping approach will serve education by providing a general methodological tool and by providing authoritative maps in substantive areas.

REFERENCES
Horn, R.E. (1989). Mapping Hypertext: Analysis, Linkage, and Display of Knowledge for the Next Generation of On-line Text and Graphics. Lexington, Mass.: Lexington Institute.
Thomas, L. (1981). Debating the unknowable. Atlantic Monthly, July, 49-50.
Toulmin, S. (1958). The Uses of Argument, Cambridge: Cambridge University Press.
Toulmin, S., R. Rieke, & A. Janik. (1979). An Introduction to Reasoning. New York: Macmillan.
Turing, A. (1950). Computing machinery and intelligence. Mind 59, 434-460.