ISSA Proceedings 2006 – Questioning The Fallacy Of Many Questions

1. Introduction
In June 2001, BBC Newsnight presenter Jeremy Paxman interviewed the Prime Minister of the UK, Tony Blair, about his past term. During this interview, Paxman posed the following question: “But you said ‘over the five years of a Labour Government we will rebuild the NHS.’ Did you underestimate the task?”. Blair subsequently responded, “I don’t think we underestimated the task”. Paxman, however, continued asking “Why say you could do it in five years?”, to which Blair unceremoniously replied with “We didn’t.” (Newsnight, 2001, June 5).
Interestingly, even though the Prime Minister seems to accept the presupposition that he stated his government would rebuild the National Health Service in five years by means of answering the interviewer’s initial question, he later openly rejects this proposition. The answer to Paxman’s follow-up question reveals an interesting property of the initial question: answering it apparently tricked Blair into committing to a proposition he did not want to be committed to.
Within argumentation studies, these type of questions have been treated in the extant literature under the general title fallacy of many questions. Consequently, Paxman’s initial question could be analysed as an instance of this fallacy. However, given that later on in the interview Blair admits he has said the Labour government would rebuild the NHS in five years on two occasions – Blair: “It [the statement about rebuilding the NHS in five years] is in the manifesto [of my Labour Party].”; and Paxman: “It was a mistake to say it [the statement] then?” Blair: “No. We do have to rebuild the National Health Service. We are doing it.” – Paxman seems to have been wholly justified in presupposing it in his question.
So how do we practically analyse such potential instances of the fallacy of many questions? To answer this question, we will first discuss the few definitions of the fallacy of many questions presented in the extant literature and their accompanying difficulties. Second, we will examine the possible approaches the argumentation analyst could take to evaluating the fallaciousness of these kinds of questions. And third, we will present which approach s/he should take in the practical analysis of potentially fallacious questions and discuss the implications of this approach.

2. Literature Definitions
The fallacy of many questions, also known as the fallacy of the complex question, is traditionally illustrated by the question “Have you stopped beating your spouse?” (see Robinson 1936, p.196; Oesterle, 1963, p.259; Hamblin 1970, p.38; Hintikka 1976, p.28; Walton 1989, pp. 36-75). This example nicely illustrates the difficulties the respondent is confronted with in answering this question: answering it in any fashion, whether affirmatively or negatively, commits the respondent to the presupposition that s/he has beaten his or her spouse. Further, challenging the question might be interpreted as undermining the questioner’s credibility and can be thus impolite or otherwise face threatening. As this leaves the respondent with no reasonable chance to answer (or challenge) the question, this way of asking questions has been treated in the literature as fallacious. However, when it is perfectly clear from the situation that the respondent has been beating his or her spouse in the past, it would be strange to regard the question as a fallacy: the question includes a presupposition the respondent is committed to anyway. Although the fallacy of many questions is, therefore, context dependent, a definition of it would be helpful to analyse question-answer adjacency pairs in argumentative discussions.

Despite the frequent occurrence of the spouse-beating question as the fallacy’s illustration, only a few definitions of the fallacy of many questions can be found in the literature. A rather descriptive definition is provided by Walton who holds that this fallacy occurs when a complex loaded question is asked that, if answered, traps the respondent into “conceding something that would cause him to lose the argument, or otherwise be unfavourable to his side” and is thus interfering with “the respondent’s ability to retract commitments to allegations made by the other party who is asking the questions” (Walton 1999, pp. 379 and 382, resp.). So, “Have you stopped beating your spouse?” is fallacious because, if the respondent answers it directly, s/he becomes committed to the complex presupposition – more specifically, the conjunctive proposition (Walton 1999, p.381) – of having beaten his or her spouse and, presumably, this is a proposition s/he does not want to concede.
Even though this definition emphasises an important characteristic of the fallacy of many questions, namely that answering such a question inherently means the respondent is committed to a proposition that s/he would not like to be committed to, it cannot sufficiently distinguish fallacious questions from non-fallacious questions. In argumentative discourse, one of the most straightforward ways to reasonably convince the opponent of a standpoint is indeed by demonstrating that the opponent’s commitment to a particular proposition is inconsistent with opposing the standpoint at hand – even if this commitment is obtained by the opponent’s answer to a question[i]. Imagine an interlocutor putting forward the standpoint “I have always treated my spouse well”. If the antagonist in a discussion about this standpoint would subsequently ask, “Have you stopped beating your spouse?”, and the protagonist would directly answer it in the affirmative or negative, then this complex loaded question is used to trap the protagonist into committing to the presupposition “I have beaten my spouse in the past”. Since this commitment is disastrous for the defence of the protagonist’s standpoint and the protagonist is unable to retract it, the posed question should be regarded as a fallacy of many questions in terms of Walton’s definition. Yet, taken as the antagonist’s argument – because you have been beating your spouse in the past, you cannot say that you have always treated your spouse well – the antagonist’s question accurately points out the inconsistency that undermines the protagonist’s case: it does not seem to be fallacious or in any way problematic.

The pragma-dialectical definition of the fallacy of many questions, by van Eemeren and Grootendorst (1992), may be more acceptable then as it enables the distinction between fallacious and non-fallacious questions. The pragma-dialectical theory proposes that this fallacy occurs when a questioner falsely presupposes a proposition as a common starting point by wrapping up this starting point in the question’s presupposition (see van Eemeren and Grootendorst 1992, p. 152). Common starting points are defined as the ground shared by the questioner and respondent that determines whether their agreement is sufficiently broad to conduct a successful discussion (see van Eemeren and Grootendorst 2004, p.60). Therefore, in the spouse-beating question, the presupposition that the respondent has beaten his or her spouse is included as though it is a common starting point that is accepted by the respondent if s/he answers the question.
Hence, the pragma-dialectical definition of the fallacy of many questions enables the distinction between the non-fallacious version of “Have you stopped beating your spouse?” and its fallacious counterpart: the questioner does not falsely presuppose that the respondent has beaten his or her spouse if it is perfectly clear that this proposition belongs to the starting points of the respondent, while it is fallacious if it does not belong to the respondent’s starting points. However, can it always be determined if a proposition is correctly presupposed by the questioner? In other words, is it possible to establish every single starting point of the respondent in order to determine when the fallacy of many questions has occurred?

3. Possible Approaches
Since the fallaciousness of the fallacy of many questions comes down to falsely presupposing a starting point as a shared point of departure in a question’s presupposition according to the pragma-dialectical theory, it needs to be examined when a presupposition is falsely assumed. In order to do so, we will outline the three distinct ways in which the argumentation analyst could account for this fallacy when analysing question and answer argumentation and show that only one of them, if refined, is appropriate.

First, it could be argued that a presupposition always falsely assumes a shared starting point because, in everyday communication, the questioner is never entirely sure of the respondent’s starting points as they can never be made fully explicit: it is simply impossible to proceed in an argumentative discussion if every commitment were to be expressed before engaging in the discussion. Thus, it is possible to view all presuppositions contained in questions as fallacious.

As should already be clear from the previously noted context dependency of the fallacy of many questions, this view is too restrictive. The pragma-dialectical theory regards fallacies as communicative moves that frustrate the reasonable resolution of a difference of opinion (see van Eemeren and Grootendorst 1992, pp. 104-105), but not all questions that contain presuppositions hinder such a resolution. To take it to the extreme, is it even possible to have questions without presuppositions? Even the simplest yes-no question[ii] presupposes that the question can either be affirmed or negated; meaning that, under this possible approach, all questions could be analysed as fallacious, which would render it rather meaningless.
Moreover, in this view, the question “But you said ‘over the five years of a Labour Government we will rebuild the NHS.’ Did you underestimate the task?”, in the Newsnight interview with Tony Blair mentioned earlier, should be treated as fallacious because it contains a presuppositions that was not put forward at the beginning of the interview. This seems rather odd when taking into account that, in the remainder of the interview, the Prime Minister admits he is committed to saying that his government would rebuilt the NHS over five years. This demonstrates, in fact, that Paxman, as the interviewer, was right in presupposing the NHS proposition in the first place. Therefore, the initial question about rebuilding the NHS in five years could not have hindered the reasonable resolution of the difference of opinion between Paxman and Blair and could consequently not have been fallacious.
These issues might be avoided by taking the second possible approach in the analysis of question and answer argumentation. Accordingly, the argumentation analyst could assume that a question’s presupposition can never falsely presuppose a respondent’s starting point, since the respondent always has the opportunity to challenge the question if s/he does not want to be committed to the presupposed proposition. For instance, if ordinary non-spouse-beaters are asked “Have you stopped beating your spouse?”, they could reply by asserting that they have never beaten their spouse and therefore that the question is not applicable to them.

Yet, the kind of difficulties that arise with this perspective resembles those of the previously discussed approach: it fails to make a distinction between those questions that are intuitively felt to be fallacious and those that are not. Some questions that are considered non-fallacious under this approach might frustrate the reasonable resolution of the argumentative dispute. If it is, for example, known that a respondent has never beaten his or her spouse, but this respondent does not reject the spouse-beating question – because s/he does not know how to reject it or does not want to make the impression of being uncooperative – the reasonable resolution of an argumentative dispute is hindered as the respondent was, originally, not committed to the presupposition that s/he now accepts by answering the question. The presupposition is hence smuggled in.
As a last option then, the argumentation analyst could regard questions as instances of the fallacy of many questions only under certain conditions. This view allows for several specifications like: a question is fallacious if the questioner knows that the respondent does not want to commit him- or herself to a proposition but intentionally presupposes it in a question anyway; a question is fallacious if the questioner does not ask whether the respondent agrees with the question’s presupposition before posing the question; or a question is fallacious if the presupposition is not part of the starting points explicitly agreed upon before starting question and answer argumentation.

Even though explaining the fallaciousness by these specifications is less restrictive than the previously discussed two (all-or-nothing) perspectives, difficulties with them might arise as well. For instance, the first specification we used to exemplify possible fallaciousness conditions – in which the fallacy of many questions occurs only if the questioner intentionally uses presupposition to which s/he knows the respondent does not want to be committed to – is infeasible: the argumentation analyst cannot possibly know the intentions of a questioner. If it is perfectly clear that a person has never beaten his or her spouse but a questioner still asks “Have you stopped beating your spouse?”, it cannot be established whether this was asked with intent or just accidentally. Even explicitly establishing it afterwards provides the questioner with an opportunity to blatantly, but undetectably, lie about his or her intent. Moreover, it is questionable whether an intended falsely assumed presupposition would affect the resolution of a difference of opinion in a different manner than an unintended falsely assumed presupposition. What the respondent sees is what s/he gets: either a presupposition that s/he is committed to or a presupposition that s/he is not committed to.
The second example used to illustrate possible condition specifications – in which the fallacy of many questions is committed if the questioner does not ask for the respondent’s agreement with the question’s presupposition before posing the question – is both in agreement with Hamblin’s idea that the spouse-beating question is “perfectly proper” if it has been preceded by the affirmatively answered question “Did you used to beat your spouse?” (Hamblin 1967, p.52) and with Krabbe’s suggestion that, in a dialogue, insertion of the question “Do you have a habit of beating your spouse?” before the question “Have you stopped beating your spouse?” would render the question non-fallacious (Walton 1989, p.68).
Nevertheless, such a confirmation requirement is problematic. As stated before, each question has a presupposition and thus, when asking if the respondent agrees with the question’s presupposition, another question is raised about which the questioner has to ask for the respondent’s agreement with its presupposition accordingly. In other words, the questioner is confronted with the undesirable choice between continuing this confirmatory process infinitely or cutting it off at some arbitrary point.

Even if we would assume that we could theoretically warrant the idea that the respondent’s confirmation only needs to be obtained about the presuppositions made in the questioner’s primary question, this might not always be practically possible and, worse case scenario, elongate the discussion to such an extent that its reasonable resolution is frustrated. First, a question like “Have you stopped beating your spouse?” not only presupposes that the respondent has been beating his or her spouse in the past, but also that s/he has a spouse, is capable of beating this spouse, understands what is meant by the question, understands how to answer the question, etc. As several such presuppositions can be distinguished, the respondent should be asked to confirm each of them. Yet, this is not always practically possible due to time constraints, social conventions, or politeness considerations. Additionally, in the worst-case situation, the questioner asking each of these questions might obscure the point of his or her discussion with the respondent to the extent that its resolution is hindered, rather than promoted.
In reaction, the argumentation analyst could regard the fallacy of many questions to be committed under the conditions specified in the thirdly mentioned example – the fallacy of many questions occurs if the question’s presupposition is not part of the common starting points explicitly agreed upon before starting question and answer argumentation – amounting to a compromise between the first two possible all-or-nothing approaches. Although this definition will keep argumentation analysts from an infinite workload, it touches upon the core problem regarding the fallaciousness of the fallacy of many questions: not all the propositions the respondent wants, or does not want, to be committed to can be made explicit before starting an argumentative discussion.

4. Practical Approach
So, we seem to have arrived at an impasse: our discussion of the possible approaches the argumentation analyst could take in his or her analysis of potential instances of the fallacy of many questions shows that it cannot always be determined whether a proposition is correctly presupposed by the questioner as it is not always possible to establish every single starting point of the respondent; yet, knowing the propositions that constitute the common starting points of the discussants is essential to determining whether the fallacy of many questions has been committed. However, although the analyst cannot know all the propositions that do and do not belong to the commitment set of the discussants, s/he can distinguish two cases in which the parties’ starting points are perfectly clear.
The first case occurs when one party explicitly states, before starting a discussion, to which starting points s/he wants or does not want to be committed to. For example, if a respondent stated that s/he has never beaten his or her spouse, it is fallacious to presume the opposite in a question’s presupposition. Yet, since starting points may remain implicit in everyday discussions – leaving the questioner to assume the respondent’s commitment to the presupposition of his or her question – this situation occurs only limitedly.
Yet there is another situation where it is clear what the starting points are, or better said, what starting points the parties are not committed to. The parties involved in a critical discussion know that their standpoint is never part of the opposite party’s starting points and thus presupposing these standpoints in questions directed to the other party would be fallacious. Imagine having a discussion about the standpoint “He has beaten his spouse”. Here, the protagonist would act fallaciously when asking “Has he stopped beating his spouse?”, because the antagonist is not committed to this presupposition – if s/he were, there would not have been a difference of opinion in the first place.
This second situation is described substantially in the extant literature as the begging the question fallacy – also known as circular reasoning or petitio principii. According to the pragma-dialectical theory, the discussion party committing the begging the question fallacy hinders the reasonable resolution of a difference of opinion by presenting a proposition that amounts to his or her standpoint as a common starting point in the argumentation advanced in support of this standpoint (see van Eemeren and Grootendorst 1992, p.153). Consequently, in question and answer argumentation, this takes place when a discussant presents his or her standpoint as an agreed upon presupposition in a question posed to the respondent.
Because the fallacy of many questions only occurs in argumentative interactions between two or more discussants and can always be distinguished if the presupposition comes down to the same thing as the questioner’s standpoint – rather than on the limited occasion that a question’s presupposition constitutes the opposite of the respondent’s starting points that were made explicit beforehand – the fallacy of many questions is, in the practical analysis of argumentative discourse, essentially the dialectical version of the begging the question fallacy.

5. Implications
With respect to the Newsnight interview question: “But you said ‘over the five years of a Labour Government we will rebuild the NHS.’ Did you underestimate the task?”; and Blair’s subsequent challenge of the question’s presupposition, reconstructing the argumentation shows that Blair defended a standpoint like “Labour’s policies did not fail”. Consequently, the presupposition “Blair said ‘over the five years of a Labour Government we will rebuild the NHS’” is not the direct opposite of Blair’s standpoint – it can only be used to argue against this standpoint by demonstrating, for a start, that rebuilding the NHS has failed in the five year time period. Therefore, Paxman did not beg the question in posing his question.
However, the question could still be an instance of the fallacy of many questions if Blair made explicit in the present discussion that he is not committed to the starting point that he pledged to rebuild the NHS in five years before the question was asked. Yet, he did not do so, which means that Paxman did not commit this fallacy either.
Interestingly, although Paxman’s question cannot be analysed as fallacious, the question and answer argumentation concerning rebuilding the NHS in five years seems to obstruct the resolution of the argumentative dispute. As elucidated by the interviewer’s questions, the Prime Minister’s contradictory commitments – Blair first accepts the proposition about pledging to rebuild the NHS over five years of his Labour government by saying that he had not underestimated this task, then explicitly denies this proposition, and subsequently accepts it again by both admitting that it is in his party’s manifesto and asserting that stating it was not a mistake – hinder the reasonable resolution of the dispute between Paxman and Blair. Consequently, a fallacy is committed, but the present analysis shows that it is Blair who acts fallaciously by misusing unclearness about his commitment – or absence of commitment – to the proposition about rebuilding the NHS, rather than Paxman in asking the question that clarified this.

6. Conclusion
So, in order to establish how the argumentation theorist should analyse potential instances of the fallacy of many questions, we established what this fallacy practically comes down to in argumentative discourse.
In accordance with the pragma-dialectical theory, it was argued that questioners commit the fallacy of many questions if they falsely presume the respondent’s commitment to a proposition in the questions’ presupposition: questions that presuppose propositions the respondent is not committed to hinder the resolution of a difference of opinion between the questioner and respondent.
Practically seen, a clear-cut case in which the argumentation analyst can establish that this fallacy is committed is when a question’s presupposition is identical to the standpoint of the questioner in the discussion at hand. Because, from an analytical perspective, this is the only straightforward case in which fallacy of many questions can always occur in argumentative exchanges and because there needs to be a communicative interaction between a questioner and respondent for this fallacy to be committed, the fallacy of many questions is essentially the dialectical version of the begging the question fallacy in argumentation analysis. So, if a question’s presupposition is essentially the same as the questioner’s standpoint, then this presupposition is falsely assumed and the fallacy of many questions is committed.
However, to take all the situations into account in which questions can be analysed as instances of this fallacy, the exceptional case in which the respondent made clear s/he does not want to be committed to the question’s presupposed starting point before the question is asked in the discussion also needs to be regarded as an instance of the fallacy of many questions. This is provided for by, next to recognising that this fallacy occurs as the dialectical variant of begging the question, recognising that the fallacy of many questions is committed if the respondent put forward the starting points to which s/he does and does not want to be committed to before a question is posed, while the questioner presupposes the opposite of these commitments in his or her question anyway.

NOTES
[i] Elucidating commitments by means of questions is empirically shown to be part of the standard pattern of confrontation in argumentative discourses (van Eemeren et al. 1993, p.43).
[ii] A yes-no question in which the pragmatic presupposition is the same as its semantic presupposition.

REFERENCES
Eemeren, F.H. van & Grootendorst, R. (1992). Argumentation, Communication, and Fallacies. Hillsdale, New Jersey: Lawrence Erlbaum Associates, Inc.
Eemeren, F.H. van, Grootendorst, R., Jackson, S., & Jacobs, S. (1993). Reconstructing Argumentative Discourse. Tucsalosa / London: The University of Alabama Press.
Eemeren, F.H. van & Grootendorst, R. (2004). A Systematic Theory of Argumentation. Cambridge: Cambridge University Press.
Hamblin, C.L. (1967). Questions. In: P. Edwards (Ed.) Encyclopedia of Philosophy (vol. 7, pp. 49-53). New York: Macmillan.
Hamblin, C.L. (1970). Fallacies. London: Methuen.
Hintikka, J. (1976). The Semantics of Questions and the Questions of Semantics. Amsterdam: North-Holland.
Newsnight [BBC Interview Transcript]. (2001, June 5). Retrieved June 19, 2006, from http://news.bbc.co.uk/1/hi/events/newsnight/1372220.stm
Oesterle, J.A. (1963). Logic: The Art of Defining and Reasoning. Englewood Cliffs, NJ: Prentice-Hall.
Robinson, D.S. (1936). The Principles of Reasoning. New York: Appleton-Century-Crofts.
Walton, D.N. (1989). Question-Reply Argumentation. New York: Greenwood Press.
Walton, D.N. (1999). The Fallacy of Many Questions: On the Notions of Complexity, Loadedness and Unfair Entrapment in Interrogative Theory. Argumentation 13, 379-383.