Microsoft Word - Walton Dialectical Shiftsv4.doc


© Douglas Walton.  Informal Logic, Vol. 29, No.1, pp. 54-83. 
 

Dialectical Shifts Underlying  
Arguments from Consequences 
 
 
DOUGLAS WALTON 
 
 
Centre for Research in Reasoning, Argumentation & Rhetoric 
Department of Philosophy 
University of Windsor 
Windsor, ON 
Canada  N9B 3P4 
 
dwalton@uwindsor.ca 
 
 
 
Abstract: Eight structural criteria are 
developed as part of a dialogical method by 
testing them against seven examples of 
arguments from negative consequences. The 
aim is to provide a method for evaluating the 
arguments in the examples as fallacious or 
not. It is shown that any method that can be 
satisfactorily used to evaluate such examples 
needs to be based on two techniques. The 
first is careful application of argumentation 
underlying shifts from one type of dialog to 
another schemes. The second is consideration 
of contextual factors concerning.  

Résumé:  On développe huit critères dans 
une approche dialogique pour déterminer si 
des arguments fondés sur des conséquences 
négatives sont fallacieux. On montre que 
n’importe quelle méthode qu’on peut 
employer pour évaluer de façon satisfaisante 
ces types d’arguments doit reposer sur deux 
techniques : une application soignée des 
schèmes argumentatifs ; et l’examen de 
facteurs contextuels concernant des 
changements sous-jacents lorsqu’on passe 
d’un type de dialogue à un autre 

 
Keywords: ad consequentiam fallacy; ad baculum fallacy; types of dialog; practical reasoning; 
value-based reasoning; dialog relevance; persuasion dialog; deliberation. 
 
 
 
The task undertaken by this paper is to extend the dialog methodology of 
(Walton and Krabbe, 1995) to get help with attacking the problem of using the 
notion of a dialectical shift more effectively as a tool to help us to evaluate 
problematic cases of arguments from negative consequences, especially those 
associated with certain informal fallacies. Argument from negative 
consequences has a distinctive argumentation scheme, and so does argument 
from threat, a subspecies of argument from negative consequences famously 
associated with the ad baculum fallacy. But based on seven key examples, it is 
shown how the schemes for these two types of argumentation are insufficient 
as tools to enable us to fairly judge, in any given case, whether such an 
argument is fallacious or not. The schemes required to analyze the examples 
are presented in section 1. The examples are used to show that argument from 



          Dialectical Shifts Underlying Arguments from Consequences 
 
55

negative consequences, including argument from threat, are sometimes quite 
reasonable, but that, in some instances, both can be used as deceptive 
sophistical strategies.1 The problems are (a) to put in place normative structures 
that can be used to analyze and evaluate these arguments, and (b) to use this 
methodology as a basis for determining whether a given instance is fallacious 
or not. It is shown that this procedure needs to be carried out at two levels. At 
an inferential level, it requires employment of argumentation schemes, but also 
at a dialectical level it needs to take dialectical shifts into account (transitions 
as a chain of argumentation moves from one conversational context to 
another).  

It is shown that once these tricky types of argument are clearly defined 
using argumentation schemes that exhibit the forms of reasoning on which 
each is based, each individual instance can be analyzed and evaluated using a 
retrospective model that takes dialectical shifts into account. A retrospective 
analysis is then applied, using formal dialectical models of rational 
argumentation (Wells and Reed, 2006) that have rules that specify when a 
progression over a dialectical shift is legal or not. This task is taken to be 
dialectical, or logical in nature, but it does suggest another problem that is 
rhetorical in nature. This other problem is that of explaining why these 
arguments are often so persuasive even when they are fallacious. Why is it that 
they are so rhetorically powerful in everyday conversational argumentation, 
and commonly effective as widely exploited rhetorical techniques of 
persuasion used in media argumentation? No direct attempt is made to solve 
this rhetorical problem, but at the end of the paper some dialectical resources 
for approaching it are suggested.  
 
 
1. Argumentation schemes 
 
The argumentation scheme for argument from consequences represents a very 
common form of argumentation. In argument from positive consequences, a 
particular policy or course of action is recommended by citing positive 
consequences of carrying it out. Argument from negative consequences cites 
negative consequences of carrying the policy out, and uses that as a reason to 
argue against carrying it out. Such arguments are quite often reasonable. For 
example, your physician might recommend against your nutritional habits by 
arguing, “Eating too much salt has the consequence of raising blood pressure; 
raising blood pressure is a bad consequence for you; therefore you should not 
eat too much salt.” Argument from consequences is also often used in ethical 
and political deliberations on what course of action to take, given a choice. 
Aristotle clearly recognized how this form of argument can be used by both 
sides in a disputation in Rhetorica (1399a14 - 1399a15).2  
                                                 
1 The thesis that fear and threat appeal arguments are often reasonable, but are 
fallacious in some instances of their use, has been supported through the analysis of 
both kinds of argument in (Walton, 2000). 
2 The quotation is from the Loeb Classical Library Edition (311). 



Douglas Walton 56 

 
. . .since in most human affairs the same thing is accompanied by some 
bad or good result, another topic consists in employing the consequences 
to exhort or dissuade, accuse or defend, praise or blame. For instance, 
education is attended by the evil of being envied, and by the good of 
being wise; therefore we should not be educated, for we should avoid 
being envied; nay, rather, we should be educated, for we should be wise.  

 
A topic is a defeasible form of argument of a kind that is nowadays called an 
argumentation scheme. It is not hard to appreciate how common argumentation 
from consequences is, and how it is often used as a reasonable form of 
argumentation. Still, an inherently reasonable form of argument can sometimes 
be used as a sophistical tactic to try to get the best of an opponent unfairly in 
disputation.  

Two argumentation schemes representing the two forms of 
argumentation from consequences have been put forward in (Walton, 1996, p. 
75). The scheme for argument from positive consequences takes the following 
form. 
 

Argumentation Scheme for Argument from Positive Consequences 
 

PREMISE: If A is brought about, good consequences will plausibly 
occur. 

CONCLUSION: A should be brought about. 
 
The scheme for argument from negative consequences takes the following 
form. 
 

Argumentation Scheme for Argument from Negative Consequences 
 

PREMISE: If A is brought about, bad consequences will plausibly 
occur. 

CONCLUSION: A should not be brought about. 
 
According to (Walton, 1996, pp. 76-77), three critical questions match the 
scheme. 
 
CQ1. How strong is the probability or plausibility that these cited 

consequences will (may, might, must) occur? 
CQ2. What evidence, if any, supported the claim that these consequences will 

(may, might, must) occur if A is brought about? 
CQ3. Are there consequences of the opposite value that ought to be taken into 

account? 
 
The argument has a presumptive status, once the positive or negative 
consequences are cited as reasons to support the proposed course of action. But 
the argument is cast into doubt if there is a failure to answer any of these 



          Dialectical Shifts Underlying Arguments from Consequences 
 
57

critical questions adequately, once they have been asked. So conceived, 
argument from consequences can be strong in some cases, weak in others. It 
can be weak if it fails to answer appropriate critical questions that have been or 
might be asked in a dialogue. 

Argument from consequences is closely related to another argumentation 
scheme called practical reasoning. 
 

Scheme for Practical Reasoning 
 
MAJOR 
PREMISE: 

I have a goal G. 

MINOR 
PREMISE:  

Carrying out this action A is a means to realize G. 

CONCLUSION: Therefore, I ought (practically speaking) to carry out this 
action A. 

 
Critical Questions 

 
CQ1: What other goals do I have that should be considered that might conflict 

with G? 
CQ2: What alternative actions to my bringing about A that would also bring 

about G should be considered? 
CQ3: Among bringing about A and these alternative actions, which is 

arguably the most efficient? 
CQ4: What grounds are there for arguing that it is practically possible for me 

to bring about A? 
CQ5: What consequences of my bringing about A should also be taken into 

account? 
 

It can be seen from the last critical question how argument from consequences 
is related to practical reasoning. There is also another argumentation scheme 
that is closely related to argument from consequences. 

Schemes for Arguments from Values 

Variant 1: Argument from Positive Value 
 
PREMISE 1: Value V is positive as judged by agent A (judgment value). 
PREMISE 2: The fact that value V is positive affects the interpretation and 

therefore the evaluation of goal G of agent A (If value V is 
good, it supports commitment to goal G). 

CONCLUSION: V is a reason for retaining commitment to goal G.  
 

Variant 2: Argument from Negative Value. 
 
PREMISE 1: Value V is negative as judged by agent A (judgment value). 



Douglas Walton 58 

PREMISE 2: The fact that value V is negative affects the interpretation and 
therefore the evaluation of goal G of agent A (If value V is 
bad, it goes against commitment to goal G). 

CONCLUSION: V is a reason for retracting commitment to goal G. 
 
How value-based argumentation works in persuasion dialogue and relates to 
practical reasoning is well explained by (Bench-Capon, 2003) and (Atkinson, 
Bench-Capon and McBurney, 2006). Argument from values is combined with 
practical reasoning in the type of argumentation called value-based practical 
reasoning (Bench-Capon, 2003; Atkinson, Bench-Capon and McBurney, 
2006). A scheme for value-based practical reasoning combining these elements 
is formulated below (Walton, Reed and Macagno, 2008, chapter 9). 

Scheme for Value-based Practical Reasoning 
 
PREMISE 1: I have a goal G. 
PREMISE 2: G is supported by my set of values, V. 
PREMISE 3: Bringing about A is a means for me to bring about G. 
CONCLUSION: Therefore, I should (practically ought to) bring about A. 
 
Thus it can be seen that there is a cluster of argumentation schemes combining 
practical reasoning with argument from consequences and argument from 
values. Precisely how the schemes are related to each other, and how they 
should be classified as forming an organized cluster of schemes, are problems 
for future research  

There is also another group of argumentation schemes that forms part of 
this cluster. These schemes have to do with the making of a threat within an 
argument from consequences, and form part of the structure of the kind of 
argumentation traditionally classified under the heading of argumentum ad 
baculum in logic. The essential condition that differentiates the threat type of 
ad baculum argument from a regular argument from consequences that is not 
an ad baculum argument requires that a special premise must be present. This 
premise is the making of a threat by the proponent. In the speech act of making 
a threat, the speaker declares the intention of carrying out a designated action 
not wanted by the hearer, unless the hearer carries out another designated 
action (Nicoloff, 1989; Guerini and Castelfranchi, 2006). Following this view, 
making a threat can be defined as a speech act (type of move) in a dialog 
exchange. This speech act that takes the following form: I (the proponent) am 
making a commitment to see to it that the negative consequences for you (the 
respondent) come about, unless you bring about the action I request. This 
expression of commitment by the proponent is an essential characteristic of the 
appeal to threat type of ad baculum argument. If the proponent only makes a 
warning to the respondent in a case, but not a threat, the argument should be 
classified under the category of argument from negative consequences, but not 
under the category of the threat appeal subtype. For the speech act to really be 
the making of a threat, the proponent must convey his commitment to seeing to 
it that the negative consequences occur.  



          Dialectical Shifts Underlying Arguments from Consequences 
 
59

Accordingly, argument from threat has the following argumentation 
scheme (Walton, Reed and Macagno, 2008, 333), where A and B represent 
outcomes that can be brought about by an agent. 
 

Scheme for Argument from Threat 
 
PREMISE 1: If you bring about A, some cited bad consequences, B, will 

follow. 
PREMISE 2: I am in position to bring about B. 
PREMISE 3: I hereby assert that in fact I will see to it that B occurs if you 

bring about A. 
CONCLUSION: Therefore you had better not bring about A. 
 
It is clear from the discussion above that argument from threat should be 
classified as a subspecies of argument from negative consequences. It should 
also be clear that although such an argument should be seen as defeasible and 
subject to critical questioning, it is not generally fallacious, but only fallacious 
in certain instances under certain conditions. One such condition is where there 
has been a dialectical shift of a certain sort from one context of dialog to 
another. 
 
 
2. Dialectical shifts 
 
In the dialog typology of Walton and Krabbe (1995), there are six primary 
types of dialog: information-seeking dialog, inquiry dialog, persuasion dialog, 
negotiation dialog, deliberation dialog and eristic dialog. These dialogs are 
technical artifacts called normative models, meaning that they do not 
necessarily correspond exactly to real instances of persuasion or negotiation, 
etc. that may occur in a real conversational exchange. Each model of dialog is 
defined by its initial situation, the participants’ individual goals, and the 
collective goal of the dialog as a whole. 

Dialectical shifts were analyzed in (Walton and Krabbe, 1995, pp. 100-
116). In a common kind of example, there is a shift from deliberation dialogue 
to an information-seeking type of dialogue. For example suppose there is a 
debate in a legislative assembly on whether to pass a bill to install a new dam. 
Before those in the assembly will vote on this question, they will want to find 
out many facts about how the dam will be constructed, what its ecological 
consequences will be, and what it will cost to install it. In order to examine 
these questions they will call in experts, like experts in engineering and 
ecology, and they might also interview accountants who have calculated the 
cost of the dam, and persons whose property might be affected by the dam. 
When this happens there has been a shift from a deliberation dialog on whether 
it would be a good idea to install such a dam to an information-seeking dialog 
about matters like what the consequences of installing the dam would be and 
what its costs would be. In this kind of case the dialectical shift is classified as 



Douglas Walton 60 

an embedding (Walton and Krabbe, 1995, 102), meaning that the goal of the 
first dialogue, the deliberation, is supported by the advent of the second dialog, 
the information-seeking interval.  

However, some dialectical shifts are classified as illicit, meaning that the 
advent of the second dialogue blocks or interferes with the progress of the first 
(Walton and Krabbe, 1995, p. 107). To make this analysis precise, two 
theoretical tasks have to be carried out. First, we need to have a model of 
deliberation as a type of dialog, so that it can be distinguished form other types 
of dialog, like persuasion dialog and negotiation. Second, we need to use these 
models of dialog to get more precise idea of when such shifts occur to help us 
judge whether they are licit or illicit. 
  How a dialectical shift works in the case of a threat appeal argument can 
sometimes be explained in the common kind of case in which there is a shift 
from a critical discussion to a negotiation type of dialog. The making of a 
threat to influence the other side is in many cases a reasonable kind of 
argument strategy in negotiation. Threats are risky in negotiation, and can 
sometimes be inappropriate, but generally they are regarded as a reasonable 
sort of argumentation in negotiation. For example in union management 
negotiations, threats of various kinds are commonly made by both sides, and 
are part of the central fabric of the argumentation used in such cases. However, 
threats are irrelevant and obstructive in a critical discussion. If an argument 
starts out to be a critical discussion, but then shifts to a negotiation dialogue, 
the ad baculum argument used during the sequence of argumentation could be 
fallacious. 

Wells and Reed (2006) constructed two formal dialectical systems to 
model dialectical shifts from persuasion dialog to negotiation dialog based on 
specifying a unified architecture for argumentation. When two participants are 
engaged in a persuasion dialog, and they want to shift to a new sub-dialog, they 
must make a request for the shift to be legal (licit), and several requirements 
need to be met. When these requirements are met by a shift initiated by one 
party, the other party has the option of continuing in the current dialogue or 
agreeing to carry over into the shift. Wells and Reed have designed rules for 
both the persuasion dialog and the negotiation dialog specifying termination 
rules for both. Their rules are designed to allow for a clean progression, as they 
call it, from one type of dialog to another. For example, one party could take 
advantage of the opportunity to shift to a negotiation dialog by making the 
appropriate move, but the respondent could make a counteroffer. Such 
metadialog negotiations can help the two parties reach agreement, because 
even though they have run out of arguments in the persuasion dialogue, still 
they might reach agreement on how to proceed further by making a shift. There 
seems to be no reason why their analysis cannot be extended to other kinds of 
shifts, like the shift from a persuasion dialog to a deliberation dialog. However, 
we need to have a clear account of the characteristics of both types of dialog. 

In a persuasion dialog, one participant puts forward a thesis to be proved, 
and the other puts forward an opposed thesis, or else expresses doubt about the 
first party’s thesis. As in all types of dialog, there are three main stages, the 
opening stage, the argumentation stage and closing stage. The purpose of a 



          Dialectical Shifts Underlying Arguments from Consequences 
 
61

persuasion dialog is to resolve this initial conflict of the opening stage by 
putting forward arguments on both sides at the argumentation stage (Prakken, 
2006). The overarching principle of burden of proof, called the burden of 
persuasion, the requirement that he who asserts must prove, is set at the 
opening stage. Meeting the burden of persuasion is determined by three factors: 
(1) what strength of argument is needed to win the dialog for a participant at 
the closing stage (standard of proof), (2) which side bears the so-called burden 
for producing such an argument, and (3) what kind of argument is required for 
this purpose. ‘Winning’ means producing an argument that is stronger enough 
than the opponent’s argument to lift the burden of persuasion set at the opening 
stage. In contrast to the burden of persuasion that applies over the whole dialog 
from opening to closing, there is also the evidential burden (called the burden 
of production in law) that applies during the argumentation stage when a 
particular claim is made or a particular argument is put forward. Both burden 
of persuasion and evidential burden are burdens of proof.  

In the formal model of deliberation dialog presented by McBurney, 
Hitchcock and Parsons (2007, 100), a deliberation dialog consists of an 
opening stage, a closing stage, and six other stages making up the 
argumentation stage. 

 
Open: In this stage a governing question is raised about what is to be done. A 

governing question, like ‘Where shall we go for dinner this evening?’, is 
a question that expresses a need for action in a given set of 
circumstances. 

Inform: This stage includes discussion of desirable goals, constraints on 
possible actions that may be considered, evaluation of proposals, and 
consideration of relevant facts. 

Propose: Proposals cite possible action-options relevant to the governing 
question 

Consider: this stage concerns commenting on proposals from various 
perspectives. 

Revise: goals, constraints, perspectives, and action-options can be revised in 
light of comments presented and information gathering as well as fact-
checking. 

Recommend: an option for action can be recommended for acceptance or non-
acceptance by each participant. 

Confirm: to participant can confirm acceptance of the recommended option, 
and all participants must do so before the dialog terminates. 

Close: The termination of the dialog. 
 
An important property of deliberation dialog is that an action-option that is 
optimal for the group considered as a whole may not be optimal from the 
perspective of an individual participant (McBurney, Hitchcock and Parsons, 
2007, 98). In a deliberation dialog, a participant must be willing to share both 
his/her preferences and also information with the other participants. The initial 
situation of deliberation is the need for action arising out of a choice between 



Douglas Walton 62 

two or more alternative courses of action that are possible in a given situation. 
The ultimate goal of deliberation dialog is for the participants to collectively 
decide on what is the best available course of action for them to take. 

Both deliberation and persuasion dialogs can be about actions, and 
therefore we have to try to draw a bright line more carefully between these two 
types of dialog. Deliberation is not centrally an attempt by one participant to 
persuade another to become committed to a particular proposal, although it is 
quite common for there to be a shift to persuasion dialog as reasons for or 
against a proposed action are supported were criticized. There appears to be no 
burden of proof in a deliberation dialog, comparable to the central notion of 
burden of proof in persuasion dialog, but this matter has so far not been 
studied. Argumentation in deliberation is primarily a matter of supporting 
one’s own proposal with reasons to accept it, and critiquing the other party’s 
proposal. Also, one’s proposal may need to be abandoned if the reasons given 
against it are strong enough to show that the opposed proposal is better to solve 
the problem posed at the opening stage. Deliberation dialog is different from 
negotiation dialog, which deals with competing interests, because the 
participants evaluate proposed courses of action according to standards that 
may be contrary to their personal interests. 
 
 
3. Problematic examples of arguments from negative consequences 
 

Argument from consequences (argumentum ad consequentiam) has been 
cited as a fallacy in some modern logic textbooks. Rescher (1964, p. 82) 
warned that “logically speaking”, it can be “entirely irrelevant that certain 
undesirable consequences might derive from the rejection of a thesis, or certain 
benefits accrue from its acceptance”. He (1964, p. 82) cited the following 
example, which could be called the classic case, because it illustrates the error 
very well to logic students. 
 

The Mexican War Example 
 

The United States had justice on its side in waging the Mexican 
war of 1848. To question this is unpatriotic, and would give 
comfort to our enemies by promoting the cause of defeatism. 

 
In this instance, argument from consequences is classified as a fallacy on the 
ground that is not relevant to the issue supposedly being discussed. The issue 
was supposedly the historical/ethical conflict of opinions on which side was in 
the right in the Mexican war of 1848. It seems that there has been a shift to a 
different issue when one side argues from negative consequences by saying 
that questioning that the U.S. was in the right would promote defeatism. 

The Mexican-American War, called the Mexican War in the United 
States and the United States Invasion in Mexico, was an armed military 
conflict between the United States and Mexico from 1846 to 1848. The war 
arose because Mexico did not recognize the secession of Texas when the U.S. 



          Dialectical Shifts Underlying Arguments from Consequences 
 
63

annexed Texas in 1845. The Mexican government permitted a few hundred 
American families to settle the area, and English speaking settlers formed a 
majority in Texas. Texans became unhappy with the Mexican government, and 
this situation led to the conflict. 

The Mexican war argument would seem to be an example of a dialectical 
shift that should be classified as illicit, given the judgment that the prudential 
deliberation about whether questioning whether the U.S. had justice on its side 
would give comfort to our enemies is not relevant to resolving the issue of 
which side had justice on its side in the Mexican war. Rescher (1964, 82) 
classified this case as an instance of a fallacy of relevance. On this evaluation 
of the fallacy of relevance, the shift to this prudential issue is merely a 
distraction that interferes with the progress of the original critical discussion by 
distracting the audience away by introducing another issue that may be more 
pressing, but is not useful to resolve the original dispute. It may indeed have 
been true that saying so would have bad consequences for the national interest. 
Still, this assertion is not relevant to the issue of which side was in the right. 
However, to show exactly why not is tricky. We have to have some criterion of 
relevance, and we have to show precisely why the argument in this case 
violated that criterion. 
  Argument from consequences is closely associated with other types of 
arguments traditionally recognized as informal fallacies.  
 

The Riots Example 
 

If the defendant is acquitted, there will be riots. Therefore, he 
should be found guilty.  

 
It may be true that there will be riots if the defendant is acquitted, but if there is 
to be fair trial, it should not follow that the defendant should be found guilty. In 
this case it is easy to recognize the fallacy right away, because of what we 
know about fair trials. The consequence of riots is not evidence of the kind 
required in a fair trial. Spelling out exactly which requirement of a fair trial 
(notions of relevance, admissibility and evidence) is involved, would not be a 
trivial job, however, for an argumentation theorist.  

Another example does not seem fallacious at all when you first look at it.  
 

The Drinking Example 
 

You should stop drinking unless you want to die young like your 
father. 

 
This example seems reasonable enough as a piece of advice, but it depends on 
the circumstances. The parallel between the case of the respondent and his 
father may be open to challenge. As an argument from negative consequences 
it might be reasonable or not. It doesn’t seem to be fallacious, perhaps just 
weak or questionable. 



Douglas Walton 64 

In short, there are problems in evaluating all three examples to judge 
whether they are reasonable or fallacious. A specific problem with the first two 
examples is that there seems to be an underling shift from an issue about 
whether a proposition is true or false, based on the evidence, to a prudential 
issue about actions. In the drinking example, the issue seems to be prudential, 
from all we can judge, and so there seems not to be such a shift. Note that the 
riots example in particular, but perhaps the other two examples well in lesser 
degree, may invoke the traditional fallacy category of argumentum ad 
baculum, since it may suggest something threatening to the respondent. 
Certainly it may act as a fear appeal. 

Other kinds of arguments associated with fallacies, like the slippery slope 
argument, are sometimes based on argumentation from negative consequences. 
Many of the examples of the fallacious argumentum ad baculum found in logic 
textbooks involve the use of a threat of negative consequences. The logic 
textbooks tend to define argumentum ad baculum as the use of force or a threat 
by one party in a dispute, to try to get the other party to accept an argument he 
is advocating (Walton, 2000, chapter 2). Many accounts, however, have a more 
inclusive treatment that includes arguments that appeal to fear, even if no threat 
is made. In this paper, we are not directly concerned with use of force (say, by 
taping your opponent’s mouth shut) or with fear appeals. The type of ad 
baculum argument we are concerned with concerns the use of a threat. The 
scheme for this form of argument is that of argument from negative 
consequences, along with the additional premise that the speaker makes a 
threat to the hearer. When the proponent makes a threat in this type of 
argument, she is not only telling the respondent that the bad state of affairs B 
will or may happen, unless he brings about A. This premise adds a personal 
factor to the effect that the proponent is declaring that she will bring about the 
event B, unless the respondent takes the proposed action with respect to A. The 
fear appeal argument is not based on a threat.  

It is known that whether ad baculum arguments are fallacious depends on 
the context of the dialog in the case at issue (Walton, 2000). In negotiation 
dialog, threats are accepted (within limits) as normal argumentation strategies, 
whereas in a critical discussion, threats are not relevant. In a critical discussion, 
participants are not supposed to prevent the other party from advocating his 
viewpoint or critically questioning the other’s viewpoint (van Eemeren and 
Grootendorst, 2004, 190). The making of a threat as an argument could go 
against this rule of a critical discussion. The making of a threat could also be 
irrelevant by violating the relevance rule of a critical discussion (192). 
 

The Free Will Example 
 

A professor and a student are discussing the issue of free will 
versus determinism in a philosophy seminar, and the professor 
says, “You had better stop advocating that argument against free 
will or I’ll give you a failing grade in this course!” 

 



          Dialectical Shifts Underlying Arguments from Consequences 
 
65

It would be evident to the participants in the class that this argument is 
inappropriate. The shift from the critical discussion to some sort of prudential 
act of warning or threat is a clear indicator that the professor has committed an 
ad baculum fallacy. It is an argument from negative consequences, and 
prudentially, it may be wise for the student to act in accord with it, but it is not 
relevant to the discussion supposedly taking place. 

Ad baculum arguments that are based on the use of a threat often take the 
form of an indirect speech act. The following example is a classic case.  
 

The Firebombing Example 
 

The last person who didn’t buy protection from our association was 
the victim of an unfortunate accident. Therefore, it would be 
prudent for you to buy our protection insurance in order to prevent 
such unfortunate consequences of not having it. 

 
This example looks on the surface like a warning, as in the drinking example. 
The indirect threat allows for plausible deniability for the proponent of the 
threat. Later the “insurance salesman” can say it was just a warning, even if, at 
the time, respondent, at the time the threat was made, would recognize it as a 
threat.3 The indirect threat involves the use of a strategy of deception.  

Consider another example made up to represent a common type of case.  
 

The Jury Intimidation Example 
  

In a case of jury intimidation, a jury member realizes quite well 
that a motorcycle gang’s threat to kill him is irrelevant as legal 
evidence that should be considered in the trial. But he asks to be 
taken off the jury because he fears for his life.  

 
In this example, the threat used to intimidate the jury member is direct rather 
than indirect, assuming there was no attempt to disguise it as a warning. Still, it 
is credible and effective because of who has made it, and because of the 
respondent’s normal commitment to his own interests and safety. The logic 
texts would probably classify the argumentation in the example as fitting under 
the heading of the argumentum ad baculum, and as such, it would no doubt be 
taken to commit the ad baculum fallacy. It is similar to the riots example in 
certain respects, but also introduces some new factors that will be considered in 
section 5. 

In his book, The Enemy at Home, right-leaning author Dinesh D’Sousa 
employed a novel and interesting argument against left-leaning politicians, 
celebrities and activists to blame recent terrorist attacks by Islamic 

                                                 
3 Many examples of indirect threat arguments of this sort are studied in (Walton, 
2000). 



Douglas Walton 66 

fundamentalists. The description of the argument presented in the example is a 
very brief summary.4 
 

The Domestic Insurgency Example 
 

D’Sousa denounced America as having sunk into decadent moral 
values, he argued that religious fundamentalists, with some 
justification, judged America as the worst civilization for decadent 
values. Examples of decadent actions he cited include widespread 
use of intoxicants, gambling, and fornication. He argued that the 
attempts to promote gender equality in the developing world can be 
seen as promoting values considered disgusting and deviant by 
traditional cultures. He named more that a hundred left-leaning 
politicians, celebrities and activists whose actions, he argued, were 
responsible for the causing the hatred of the terrorists because of 
their attempts to promote these decadent moral values and impose 
them on the rest of the world. He did not accusing them of being 
terrorists themselves, or of even of actively working to promote the 
interests, of the terrorists. However, he argued that because of the 
consequences of their actions, they were taking part in a “domestic 
insurgency” that was, in effect, “working in tandem” with Osama 
bin Laden to defeat George W. Bush.  

 
This example of the use of argument from negative consequences is highly 
controversial, and no doubt many of those attacked by it are antagonized by it. 
D’Sousa cited many activities to promote their “decadent” views by activists, 
intellectuals and celebrities that, he alleged, had negative consequences for 
America. Can we conclude that the argument in the domestic insurgency 
example is a fallacious instance of argumentation from negative consequences? 
We return to a discussion of this question in section 5. 
 
 
4. Analysis of the Mexican War example 
 
  We now return to a discussion of each of the examples in turn, beginning with 
the Mexican war example. In each case should begin by trying to see which 
argumentation scheme fits the argument in the example. In order to do this, we 
have to analyze the argument in the example and identify its premises and 
conclusion. An analysis of the structure of the Mexican war argument is shown 
in Figure 1. 
 

                                                 
4 A review of the book presents a summary on which some details of the example are 
based: Jerry Adler, ‘America’s Most Wanted’ (review of The Enemy at Home) 
Newsweek, Feb. 5, 2007, p. 46. 



          Dialectical Shifts Underlying Arguments from Consequences 
 
67

       
Figure 1: First Argument Diagram of the Mexican War Example 

 
In the diagram in Figure 1, the missing premise D has been inserted. It is 
shown in a darkened box, indicating that it has been added as an implicit 
premise. The argumentation scheme (argument from consequences) that links 
premises C and D to conclusion A has been represented on the diagram.5  

What is tricky about the argument can be approached by looking at 
Figure 1. Figure 1 represents the routine way that the argument in the Mexican 
war example should be represented by an argument diagram. But when the 
argument is represented in this way, it seems blatantly wrong, because the 
arguments composed of B and C&D do not support the truth of the conclusion 
A. But what has gone wrong? What these two arguments do support is another 
conclusion A+: ‘You (the respondent) should not question (in public) the 
proposition that the United States had justice on its side in waging the Mexican 
war of 1848’. This way of analyzing the argument is represented in the 
argument diagram in Figure 2.  

                                                 
5 In this instance it is argument from negative consequences.  



Douglas Walton 68 

      
Figure 2: Second Argument Diagram of the Mexican War Example 

 
Looking at the argument as shown in Figure 2, it seems reasonable. The two 
arguments can correctly be seen as offering reasons to support the conclusion 
(A+).  

Comparing the two argument diagrams, we begin to get some inkling of 
what has gone wrong. The argument could be classified as an instance of the 
fallacy of arguing to the wrong conclusion. The fallacy can be diagnosed using 
this triad. 

 
Remark 1: B and C&D do not give reasons to support acceptance of 

conclusion A. 
Remark 2: B and C&D do give reasons to support acceptance of 

conclusion A+. 
Remark 3: It is easy to confuse A with A+. 
Remark 3 is a plausible statement because the difference between A and 

A+ is subtle. It is the difference between saying that some 
statement is true (or not), as supported by evidence, and saying that 
saying this statement is true (or not) is commendable (or not). A is 
a statement about something, while A+ is a statement about talking 
about A.  

 
 What could be called the fallacy of arguing to the wrong conclusion is an 
argument fitting the following pattern (Walton, 2004, 35): the arguer is 
supposed to prove conclusion A, but he puts forward an argument for 
conclusion A+, a proposition that looks like (or appears to be the same as) A. 
Aristotle seemed to have something like this fallacy in mind when he made 
the following remark about what he called the fallacy of misconception of 
refutation: “When the argument stated is a demonstration [apodeixis] of 
something, if it’s something other than that leading to the conclusion, it will 



          Dialectical Shifts Underlying Arguments from Consequences 
 
69

not be a syllogism about that thing.”(Topica 162a13-162a16; quoted from the 
translation in Walton, 2004, 35). Seen in this way, the fallacy in the Mexican 
war example can be diagnosed as a failure of relevance of a specific kind, 
namely the fallacy of arguing for the wrong conclusion. But why would 
anyone be fooled by such an error of arguing for the wrong conclusion? More 
must be involved than the simple error revealed by the juxtaposition of the 
inferential structures shown in Figures 1 and 2. 
 What gives the fallacious argument its aura of plausibility is the melding 
together of the reasonable argument from consequences shown in Figure 2 
with the patently unreasonable argument shown in Figure 1. What could lead 
a respondent to blend these arguments together and take them to be the same? 
Why is the argument so highly deceptive because of this concealed blending? 
How could we train students of critical thinking to recognize such a 
deception that conceals the shift from the one argument to the other one? 
How do such arguments have such a powerful persuasive effect as rhetorical 
strategies of strategic maneuvering (van Eemeren and Houtlosser, 2000)? It 
will now be argued, by proceeding to a second level of analysis, that the 
reason for this deceptiveness is a failure of relevance concealed by an 
implicit contextual shift .  

As noted, Rescher (1964, 82) categorized the fallacy as a failure of 
relevance. This diagnosis calls into question why the argument should be 
considered not relevant. This suggests examining the context of dialog in 
which the argument was used. This argument could be relevant if the dialog 
were that of a deliberation on how to act in a given situation. We see from the 
wording of the case that “enemies” are involved, so let us imagine that the 
situation is one in which there is some current conflict or war. In this scenario, 
giving comfort to enemies would be a negative consequence, and might even 
lead to loss of life for the soldiers on our side. This wording suggests a 
deliberation on how to act prudently. But suppose on the other hand the context 
of dialog is that of a persuasion dialog, of the kind one might have in a history 
class or an ethical discussion about which side was in the right in the Mexican 
war. Here the very same argument could be judged to be irrelevant, because the 
original discussion was supposed to be about the ethical/historical issue of 
which side “had justice on its side” in waging the war. Relevant arguments in 
such a persuasion dialog would be historical facts like who started the conflict, 
what were the territorial claims of each side, and so forth. Only these kinds of 
argument can fulfill the original burden of persuasion. 

In order to back up the claim that the argument in the example commits a 
fallacy of irrelevance because the discussion was not originally supposed to be 
a deliberation dialog, it can be hypothesized there has been a shift from one 
type of dialog to another inherent in the argumentation in the example. The 
original dialog was supposedly, from all we can tell from the information given 
in the example, a persuasion dialog on which side had justice on its side in 
waging the Mexican war of 1848. The goal of such a discussion is to resolve 
this conflict of opinions by examining the rational arguments on both sides. 
However, there was a kind of shift made when the premises were brought 



Douglas Walton 70 

forward that to question whether the United States had justice on its side is 
unpatriotic, and that doing so would give comfort to our enemies by promoting 
the cause of defeatism. What needs to be carefully observed is that these two 
statements may very well be true, and the analysis in the argument diagram in 
Figure 2 could be judged to be a reasonable argument. It is based on an 
argumentation scheme that has true premises. In this format, the argument 
would reasonably provide a prudential argument against questioning whether 
the United States had justice on its side by citing negative consequences of 
such questioning, namely that it would give comfort to our enemies by 
promoting the cause of defeatism. This prudential argument only becomes 
irrelevant and fallacious when expressed in its original form in Figure 1. Thus 
there are two levels of analysis that need to be considered together—an 
inferential level and a dialectical one. 

This bi-level analysis depends on how the notion of relevance is defined, 
but it has been argued (Walton, 2004) that an argument should be judged to be 
relevant in a case only if it is part of a connected sequence of argumentation 
used in a dialog leading to an ultimate probandum to be proved in that case. In 
a particular case, whether an argument is judged to be relevant or not depends 
on how far the sequence has gone forward at the stage where it was put 
forward. Such matters need to be judged by examining the text of discourse in 
the given case being analyzed, and asking what the purpose of the discourse is 
supposed to be. Some cases can be difficult to decide, because the argument 
simply is not finished yet, or even if it is finished, because there are a lot of 
gaps and missing premises, and it may not even be clear what type of dialog 
the participants are supposedly engaged in. however in the case is being 
considered here, the key indicator of irrelevance is that of the dialectical shift.  

The issue of the case was whether the United States had justice on its 
side in waging the Mexican war of 1848. Presumably the type of dialog here 
was some sort of persuasion dialog on which country had justice on its side in 
that war. This issue is a historical question, or perhaps an ethical question, of a 
kind commonly disputed by historians, or by anyone interested in this kind of 
issue. The arguer goes on to claim that questioning the initial claim is 
unpatriotic, and would give comfort to our enemies. This argument from 
negative consequences seems like a reasonable one if considered as part of a 
deliberation dialog. Perhaps raising this question might indeed give comfort to 
our enemies, as the claim states. But the question is whether this claim is 
relevant to the original issue being discussed. Using the second (dialectical) 
level of analysis it can be argued that it is not, because there has been a 
dialectical shift to prudential questions concerning the consequences of 
discussing the original issue of whether the United States had justice on its side 
in the Mexican war. This kind of case is quite common one, and all comparable 
examples presumably share this bi-level underlying structure. For example 
comparable arguments commonly occur in cases of freedom of speech.  
 
 
 
 



          Dialectical Shifts Underlying Arguments from Consequences 
 
71

5. Discussion of the other examples 
 
The riots example is interesting because there is clear evidence of a dialectical 
shift. The issue in the case is supposed to be that of whether the defendant 
should be acquitted of the crime he was charged with. This is an issue to be 
decided by a trial in which evidence for and against the charge is considered 
and evaluated by a judge or jury. Acquitting the defendant for the practical 
reason that there will be riots in the streets, while being in principle a 
reasonable argument from negative consequences, is not relevant to the 
conclusion that the defendant should be found guilty as the outcome of the 
trial. This argument is a clear instance of an inappropriate use of public 
pressure, in the form of argument from negative consequences, brought to bear 
on a trial in the judicial system. When the argument is shown in this light, it is 
quite clear that there has been a dialectical shift from a persuasion type of 
dialogue to a prudential argument about matters of public safety, or damage 
caused by riots. This argument is very similar to the Mexican war argument, 
and is fallacious for much the same reasons. 

The drinking example is also interesting because, at least on the surface, 
it seems to be a highly reasonable argument directed to the respondent to 
support the conclusion that he should stop drinking. The reason given is that 
continuing this action may make him die young, like his father. However, we 
can critically question premises and implicit assumptions in the argument. Did 
his father really die young, and if so was it because of his drinking? How 
similar is the case of the father to that of the son? Maybe there was some 
difference, for example, in how each individual reacted to alcohol, or in how 
much alcohol each one was drinking. But failure to substantiate a premise of an 
argument, whether explicit or implicit, should not be sufficient reason to judge 
it fallacious. Here we have a prudential argument that does present a reason for 
the conclusion, based on alleged negative consequences. It is not a fallacious 
argument from consequences. 

In the free will example, even though the dialectical shift is implicit, the 
illegitimate nature of the move in argumentation is obvious to everybody. It is 
seen as shockingly inappropriate. It meets the general requirements for the 
speech act of making a threat as a move in dialog (Walton, 2000). The 
proponent of the threat warns the respondent that something that negative 
consequences may happen to him. The proponent also indicates to the 
respondent that she (the proponent) will see to it that these negative 
consequences come about, unless the respondent complies by carrying out (or 
omitting to carry out) some designated action. In the analysis of (Walton, 2000, 
113-114), there are three essential conditions for the speech act of making a 
threat. 

 
  ►  The preparatory condition states that the respondent believes that the 

proponent can 
   bring about the negative consequences.  



Douglas Walton 72 

  ►  The sincerity condition states that both the proponent and the respondent 
presume that the respondent wants to avoid the negative consequences.  

  ►  The essential condition states that the proponent is making a commitment 
to bring about the negative consequences if the respondent does not 
comply.  

 
In the free will example, the threat is an explicit one, and it is clear to everyone 
that it is inappropriate in relation to the critical discussion that the professor 
and student are supposed to be engaging in. In contrast, in the firebombing 
example, the “insurance salesman” is making an indirect threat. An indirect 
threat is meant to be recognizable to the respondent as a threat, but is also an 
implicit speech act that leaves room open for plausible deniability. 
  An interesting aspect of the free will example is its contextual sensitivity. 
Changes to the circumstances of the example could make what was formerly an 
outrageously irrelevant argument into a perfectly good argument. Consider the 
following two modifications of the original circumstances of this example.6 
First, suppose that the student’s argument was based on racist or other 
pernicious views. In this kind of case the position adopted by the student is 
itself unacceptable, and therefore the professor could be justified in telling the 
student that if he continues to maintain this unacceptable view, he will fail the 
course. Second, suppose the position advocated by the student showed a 
misunderstanding of the material in the course. For example, suppose the 
student had put forward an argument that had already been rejected in the 
course as untenable because it had been shown to be based on a false 
assumption, like 2 + 2 = 5. Similarly, in this kind of case the professor could be 
justified in telling the student that he will fail the course if he continues to 
maintain this unacceptable view. The interesting aspect brought out by 
consideration of these two kinds of examples is that this threat argument that 
was previously fallacious now has become a non-fallacious argument. It shows 
that an argument that is clearly a fallacious instance of argumentation from 
negative consequences can be turned into a reasonable argument provided only 
that the circumstances of the case are slightly different. 

Another interesting aspect of the free will example is that there appears 
to be a shift to a different level of dialog when the professor puts forward his 
counterargument to the student. The student began, in the discussion of the 
issue of free will versus determinism, by putting forward an argument for 
determinism. The professor, at his next move, does not put forward an 
opposing argument for free will, nor does he criticize the student’s argument 
for determinism, at least directly. Instead, he makes a remark about the 
students advocating the argument for determinism, by giving a reason why the 
student should stop advocating that argument. This move by the professor can 
be seen as a dialectical shift to a meta-dialog. If the original persuasion dialog 
about the free will issue can be called a ground level dialogue, then a meta-

                                                 
6 These two modifications of the free will example, and their implications, were 
expressed in comments made by Trevor Bench-Capon on February 13, 2008, during a 
talk at the University of Groningen.  



          Dialectical Shifts Underlying Arguments from Consequences 
 
73

dialog could be defined as a dialog about that original dialog (Krabbe, 2003, p. 
83). In this instance, when the professor puts forward his argument telling the 
student that he had better stop advocating determinism or he will get a failing 
grade, there has been a shift to a meta-dialog. The reason is that the professor is 
now discussing the original dialog, a persuasion dialog, and telling the student 
to stop it, or at least stop his part in it that consists of giving arguments for 
determinism. He is telling the student to either stop it, or he will give him a 
failing grade in the course. This observation would also apply to the Mexican 
war example. 

As noted in the discussion above of the jury intimidation example, here 
we have a prudential threat appeal argument used to intimidate somebody into 
taking a course of action, but nobody is being deceived by it. Or let’s say, at 
any rate, to make the example interesting, that the jury member is not deceived 
by the threat, but simply acts in accord with it for prudential reasons, because 
he fears for his life. In such a case, although the ad baculum argument was 
clearly irrelevant to the discussion, it did not fool anybody into thinking it was 
a relevant argument. In such cases, the ad baculum could still have quite an 
effect by inhibiting the respondent. In such a case, the ad baculum strategy 
may work as a fallacy, even when there is no deception that it is relevant. Thus 
deception is not the only basis for judging an ad baculum argument to be 
fallacious. In some cases, the respondent is quite aware that the threat appeal is 
irrelevant, but gives in to it anyway on the basis of self-interest.  

But has the motorcycle gang, the jury member, or anyone else, really 
committed a fallacy? You could argue not, on the ground that a fallacy always 
requires deception, and all parties might clearly recognize that the tactic used is 
wrong. Nobody is being deceived. But the gang might use the argument 
anyway, and the jury member might be persuaded to act in accord with it 
anyway. In this kind of case, the fallacy could result from a dialectical shift 
from one setting in which an argument was used to a different setting. It could 
be prudentially justified for the jury member to respond to the threat, but in the 
setting of the trial, the threat is not relevant evidence, or should not be treated 
as such. It should not influence the jury member to decide whether the accused 
party is guilty or not. As in the other examples, the burden of persuasion in the 
trial is not met by the purely prudential argument directed against the juror. 

On the surface, the argumentation in the jury intimidation example looks 
very similar to that in the case of the riots example. There is a trial involved, 
but the argument shifts to prudential concerns by using argument from negative 
consequences. We do want to say in the jury intimidation example that 
something is wrong. It is improper that a jury member should be intimidated, 
and in fact this type of act is illegal, on grounds of obstruction of justice, the 
reason being that it makes a fair trial difficult or impossible. So the tactics used 
by the motorcycle gang are certainly illegal, as well as being unethical. But do 
they commit a fallacy? There could be some grounds for saying so, because 
they are shifting the argument from the kind of critical discussion based on 
evidence used in a trial to purely pragmatic matters of safety because a person 
fears for his life. Although it is possible that nobody is being deceived by this 



Douglas Walton 74 

move, still there could be grounds for classifying the argument as a fallacious 
ad baculum, or use of a threat. These grounds would be that of relevance. As 
stated in the jury intimidation example, the threat to kill the jury member is 
irrelevant as legal evidence that should be considered in the trial. On these 
grounds, a case could be made that a fallacious ad baculum argument is 
committed in this example. 

Still, if this evaluation is correct, it takes us back to the riots argument, 
where there was a similar dialectical shift from a trial to a prudential argument 
based on negative consequences. On these grounds, the riots argument could 
perhaps also be judged to be an instance of a fallacious ad baculum argument. 
The grounds are that the possibility or probability of riots should not be taken 
to be a good or adequate reason to support the conclusion that the defendant 
should be found guilty. It is irrelevant, even though it may be relevant to 
drawing other conclusions as prudential courses of action, like shifting the 
venue of the trial to a different location, or rescheduling the time of the trial. 

What makes the argument in the domestic insurgency example most 
interesting, and different from the other examples, is the added element that a 
kind of practical inconsistency is alleged. It is argued that the liberals that are 
the target of the argument have certain values, like gender equality, that the 
terrorists are opposed to. But then it is argued that their attempts to promote 
these values and impose them on the rest of the world are actually undermining 
the same values by promoting the interests of the terrorists. Thus it is alleged 
that the consequences of the liberals’ attempts to promote their position and 
values are in fact undermining them. This argument against the liberals accuses 
them of being committed to a pragmatic inconsistency of the same kind we are 
already familiar with in circumstantial ad hominem arguments. Thus the 
argument in this example is certainly an interesting and subtle one as an attack, 
but is it fallacious? 

In the domestic insurgency example there has been a dialectical shift 
from a discussion about ethical values to a citing of alleged negative 
consequences of the expression and promotion of these values by certain 
parties. Thus it seems similar to the Mexican war example as an instance of 
argument from negative consequences. However, the argument in the domestic 
insurgency example is more subtle and indirect than that of the Mexican war 
example. D’Sousa did argue at length that the political activism of the persons 
and groups cited had bad negative consequences. He did not argue that that 
these people have no right to express their views or to promote them 
politically, however. But, like the case of the Mexican war example, it was 
argued that the bad consequences involve loss of life. In the Mexican war 
example, loss of morale in war was cited as the bad outcome—giving comfort 
to our enemies. In the domestic insurgency example, the consequences cited 
are even worse. The argument equates the bad consequences with loss of life in 
terrorist attacks, and even appears to partially lay blame for these attacks on the 
parties cited. Thus there is a dialectical shift from a discussion of the political 
views in question to a deliberation dialog about the allegedly bad consequences 
of these views. The secondary dialog even takes the form of laying guilt for 



          Dialectical Shifts Underlying Arguments from Consequences 
 
75

these bad consequences on the parties who are alleged to have contributed to 
them, even if unknowingly. 

The argumentation in this case is much more complex than the one in the 
Mexican war example, because it was put forward in a whole book, and 
because of its politically divisive nature. Those on the right will like the 
argument in the book, while those on the left will be strongly inclined to 
disagree with its argument. But it would be an error to leap too quickly to one 
side or the other. To properly evaluate the argument, one would have to 
examine the specific claims made in the individual cases cited, and the 
arguments offered to back up these claims. There is no space for that here. 
Still, it is interesting to cite the example to show how argumentation from 
negative consequences is used in everyday conversational arguments of the 
most common kind, for example in political rhetoric, in a subtle way. Such 
cases verge on the fallacious, because of the shift concealed within the 
sequence of argumentation, and can certainly be highly deceptive. But it would 
be erroneous to declare them fallacious in a wholesale fashion, and each case 
needs to be judged on its merits or demerits, taking the dialectical shift into 
account.  
 
 
6. Retrospective evaluation 
 
The remaining problem is how one should approach a particular case, like the 
examples presented in section 1, where it appears evident to a reader or viewer 
of the example that argument from consequences may be involved. The initial 
piece of advice would be to look to see if the argument has the wrong 
conclusion. But how could one be expected to know or to prove that it has the 
right conclusion or the wrong conclusion? The next required step is to make 
some determination of what type of dialog the argument is supposed to be 
contributing to. For example, there may be evidence that the argument was 
supposed to originally take place in the context of a persuasion dialog. Yet a 
problem may arise if there is also evidence that the argument has been put 
forward in such a way that it can only be reasonable if it is taken to be part of 
some other type of dialog like a deliberation dialog. What these observations 
suggest is that the ordering of the two dialogs is crucial. We have to approach 
the case by identifying dialog type x that the argument appears to part of, and 
then analyze the whole sequence of argumentation in the case retrospectively, 
tracing back to dialog type y at an earlier point in the sequence.  
  Analyzing the Mexican war example needs to take a retrospective 
approach, an approach based on the assumption that there has been a dialectical 
shift from one type of dialog to another. The alleged shift is from a persuasion 
dialog on an issue to a different type of dialog, a prudential type of deliberation 
dialog. The case is similar to many threat appeal and fear appeal arguments 
where the argument may be reasonable, when considered in itself, but becomes 
fallacious when considered in the context of dialog. Consideration of the 
context may indicate there has been a dialectical shift, and that this shift is 



Douglas Walton 76 

evidence of the committing of a fallacy. However, one problem in this case is 
how we can be sure the original issue was supposed to be part of a critical 
discussion. There is very little evidence in the case. There are no indicator 
words or other textual evidence to show that to participants were originally 
engaged in a critical discussion on the issue of which country had justice on its 
side in the Mexican war of 1848. The only evidence is that the first statement 
in the Mexican war argument is put forward as an assertion. The statement is 
made that the United States had justice on its side in waging the Mexican war 
of 1848. The statement appears to be a claim made by one party in a dispute, 
because the second statement in the Mexican war argument appears to respond 
to it by giving a reason against questioning it. When the first statement is made 
therefore, it would appear that there is a burden of persuasion set on the party 
who made the claim, because presumably it is a claim being made to attempt to 
overcome the doubt of some other party in the dialog. But the requirement of 
lifting this burden was not met, because of the shift to the second type of 
dialog.  
  The structure of this kind of shift is illustrated in Figure 3 below.  
 

 
 
 

Figure 3: Retrospective Evaluation of Argumentation from Consequences 
 
Figure 3 portrays the progress of a sequence of argumentation from a given 
argument in an example that has been identified as fitting the scheme for 
argument from consequences. But when we evaluate how this argument has 
been used in the example retrospectively, we need to trace it back to the 
original persuasion dialog it was supposedly part of. To critically analyze an ad 
baculum argument used in such a case, or other type of argument coming under 



          Dialectical Shifts Underlying Arguments from Consequences 
 
77

the heading of argument from consequences, we have to adopt a backwards or 
retrospective viewpoint, starting with the bottom box in Figure 3. From the 
bottom box, where the argument was put forward in the example, we need to 
look backwards to the original rules of the persuasion dialog. From this 
perspective, the argument can be judged to be fallacious if there was an illicit 
shift from the original persuasion dialog represented in the top box to another 
type of dialog represented beneath it. Thus looking backwards we can 
reconsider the argument in its original context of dialog, and judge it to be 
irrelevant if the text and context of the argument supports that interpretation of 
the example. This methodology is called a retrospective analysis of the 
argumentation, as shown in Figure 3. 
  The point was made in (Walton and Krabbe, 1995, p. 104) that shifts 
from persuasion dialog to another type of dialog are not always fallacious, but 
that the shift is fallacious, or can be associated with a fallacy, when it is a shift 
of an illicit or inappropriate type. In a case cited there (p. 104), a minister of 
finance argued for a certain kind of tax exemption, and he went on to propose 
that if his critics would abstain from moving for penalties for these exemptions, 
he will refrain from opposing a bill that these critics would profit from. In this 
case, the minister evaded the issue by illicitly abandoning the original 
persuasion dialog and moving to a negotiation dialogue. The shift is illicit 
because the negotiation dialog does not support the original persuasion dialog, 
but moves away from it leaving it unfinished, and perhaps even preventing it 
from continuing in a more productive manner towards its goal. It is this kind of 
shift, as shown by the examples above, that is characteristic of many instances 
of fallacious argumentation from consequences, including the ad baculum 
fallacy. 

In the system of formal persuasion dialog set out by Wells and Reed 
(2006, 7) a shift to a different type of dialogue, like a negotiation dialogue, can 
take place only if the party making the shift has requested this sort of move to 
the other party. They also make it a rule of their formal persuasion dialog that a 
progression to another type of dialog is only allowed if the party requesting the 
shift first provides the defense of his thesis. In their system, a shift is only licit 
after the party requesting the shift has first discharged the burden of proof to 
support his previous argument in the persuasion dialog. These requirements are 
put in place because a shift is held by Wells and Reed to be useful in a 
persuasion dialog only if one party is getting nowhere with trying to persuade 
the other party, with appropriate arguments that supports his thesis. If the 
persuasion dialog is failing because of a deadlock, where one party simply 
cannot go further to persuade the other, that is the appropriate kind of situation 
where a shift to a negotiation might prove to be useful. They see the rules of 
their system of persuasion dialog as blocking the inappropriate type of 
dialectical shift, where the party making the shift is using the argument in a 
secondary dialog to fallaciously fulfill a burden of proof required for a prior 
argument put forward in the primary dialog.  
  How can we support the hypothesis that we should interpret the text of 
discourse in cases like of the Mexican war argument in such a way that it 



Douglas Walton 78 

shows evidence of an illicit dialectical shift? The invocation of the normative 
model of the critical discussion, by itself, is insufficient as a basis for such an 
interpretation. In addition, what is required is an analysis of the text to back 
up the hypotheses that this normative model fits the argumentation in this 
case and that there has been a shift from this prior dialog to a deliberation 
dialog. On this view of the matter, when we examine the text of discourse, 
we can see that the argumentation in the case fits a certain type of dialog, in 
this case a kind of historical discussion about values of the familiar kind. 
Since all of us as participants in everyday conversational argumentation can 
recognize this type of dialog, we can then fit it into a normative model like 
that of the critical discussion. As shown above, just before Figure 3, the 
evidence is that the first statement in the Mexican war argument (that the 
United States had justice on its side in waging the Mexican war of 1848) is 
put forward as a claim made by one party in the dialog. When this statement 
is made, a burden of proof is set on the party who made it. When this claim is 
in question by the other party, the burden needs to be lifted by the claimant, 
and that is how the persuasion dialog needs to properly proceed. However, 
the shift to the second type of dialog, put in place by the reason the claimant 
gave to support his claim, interfered with the proper continuation of the 
dialog by taking it in a different direction. The requirement of lifting this 
burden was therefore avoided, or at least that was the sophistical strategy 
used to try to avoid meeting the requirement at the next step in the persuasion 
dialog.  
  The dialectical structure of this kind of fallacious argumentation is 
displayed in Figure 4. 
 

 
Figure 4: Dialectical Structure Underlying Shifting from Burden of Proof 

 
The fallacy is a failure to lift a burden of persuasion that was set in place 
against the proponent when he made a claim in the original persuasion 

P makes claim. P incurs burden of 
proof to support claim. 

P puts forward 
prudential 
argument to 
support claim. 

Persuasion 
Dialog 

Argument gives a 
prudential reason 
for not questioning 
claim. 

Deliberation 
Dialog 

Dialectical Shift 



          Dialectical Shifts Underlying Arguments from Consequences 
 
79

dialog. Such a failure is not in itself fallacious, but the problem was that, at 
his next move, the proponent offered a prudential argument for the 
respondent to accept the claim he had just made. The offering of such a 
prudential argument did give the respondent a practical reason not to dispute 
or question the claim, but unfortunately it only functions in this way if there 
is a shift to a deliberation dialog. The problem is that proponent is shifting 
away from fulfilling the requirement of burden of proof in the persuasion 
dialog by presenting a kind of argument that moves to a deliberation dialog. 
This move interferes with the progress of the persuasion dialog, because it 
cannot move forward towards realizing its goal until the proponent either 
backs up his claim by giving evidence that it true or withdraws it.  
 
 
7. Conclusions 
 
The examples studied suggest that evaluation of cases of supposed informal 
fallacies needs to be carried out both by identifying the type of argument that is 
involved, and by examining how that argument is used in the context of dialog. 
This way of proceeding is supported by the theory of fallacy in (Walton 1995, 
255). People generally recognize kinds of arguments that are typically 
reasonable (in principle) but defeasible, and best seen as heuristics to guide 
action under conditions of uncertainty even though they are subject to failure in 
some instances. People also recognize that these kinds of arguments can be 
used very effectively as deceptive tactics, and thus they are on their guard 
against them if they see any evidence at all of exaggeration or sophistical 
maneuvering. These arguments have two sides. They are necessary and useful, 
but at the same time slippery and dangerous. When it comes to evaluating such 
arguments, instead of routinely dismissing them as fallacious, we need to 
realize that they can fall into one of three standard categories: (a) reasonable, 
when considered in the context of dialog; (b) weak but not fallacious; (c) 
fallacious. To take a given argument in a real case, like the examples 
considered above, and to marshal the textual evidence to justify placing them 
in one of these categories, attention should be paid both to the form of 
reasoning and to the appropriateness of using this reasoning in the context of 
dialog for the given case. 

On this approach, a fallacy is an argumentative move that goes counter to 
the direction of a dialog and poses an obstacle to the realization of the goal of 
the dialog it is supposed to contribute to. An instance is the group of fallacies 
associated with dialectical shifts. 

The evaluation of arguments from consequences depends not only on the 
logical form of the argument, abstracted from its context, but also on the 
dialectical context of how the argument was used for some purpose in a dialog 
setting. The very same argument that was appropriate and reasonable when 
used as part of a deliberation dialog can be irrelevant, and for this reason 
fallacious, when used in a persuasion dialog. Such arguments need to be 
evaluated using a retrospective model in which one looks backwards to identify 



Douglas Walton 80 

the original type of dialog that the argumentation was supposed to be a 
contribution to. It is from the viewpoint of this original type of dialog that the 
argument should be evaluated, according to the methodology presented above. 
Using the formal model of Wells and Reed (2006), we have to ask whether the 
rules specifying requirements for starting a new sub-dialog have been met. 
Only if these rules are followed can the shift properly be made from the initial 
dialog that was already under way to a new sub-dialog. Only then can there be 
a clean and visible progression from the one type of dialog to the other. The 
problem, however, in the cases we have examined, is that in an ordinary 
conversational argument, the underlying shift may be invisible. It lies under the 
surface of the text, and the participants are quite likely not even to be aware of 
its occurrence.  

In this paper eight structural characteristics help to explain how the 
argument in the Mexican war example, and other cases studied, can be 
identified as a fallacious use of argument from consequences. First, there was a 
dialectical shift from a persuasion dialog to a deliberation dialog. Second, 
when we evaluate the argument in the deliberation dialog retrospectively, by 
looking back to the rules and requirements of the original persuasion dialog, 
we can see that the deliberation dialog is not helping the original persuasion 
dialog move forward towards realizing its goal of resolving the original 
conflict of opinions. Third, there was no agreement made between the parties 
that the shift to the second type of dialog was acceptable to both. Fourth, no 
rule that allows such a shift from the first dialog the argument was cited or met. 
Fifth, the argument put forward in the deliberation dialog has a different 
conclusion from the one that the arguer was originally supposed to prove in the 
persuasion dialog. It is the wrong conclusion. Sixth, there was a shift to a meta-
dialog, visible in the contrast between conclusions A and A+ in the two 
argument diagrams for the Mexican war example (Figures 1 and 2). 
Proposition A+ is about questioning proposition A in public. Seventh, the 
argument put forward in the deliberation dialog does not fulfill the 
requirements for burden of persuasion. Eighth, a structural characteristic that 
holds for the examples studied in this paper is that the shift is specifically from 
a persuasion dialog to a deliberation dialog. Whether this characteristic holds 
for all examples of fallacious use of argument from negative consequences is 
merely a hypothesis, however, and remains to be studied in future work. 

The analysis of the structural characteristics of the argumentation in these 
relatively simple examples puts us in a position to analyze much more complex 
and subtle cases. These cases are common, controversial and interesting. They 
typically involve arguments citing the negative consequences of artistic or 
political communication. For example, the film Clockwork Orange, and many 
other films and literary works as well, have been criticized on the grounds that 
they tend to promote violence. Indeed, murderers have sometimes admitted that 
their violent acts were inspired by books or movies. It may be true in some 
instances that these works of art have had unfortunate consequences. But is that 
the basis of a good argument for condemning them? It might be, in some 
respects, but one has to be careful not to commit the fallacy of improper 
argument from consequences. For such a work of art may have literary or 



          Dialectical Shifts Underlying Arguments from Consequences 
 
81

artistic merit, even though it did have the bad consequences cited. Still, from a 
prudential or deliberative point of view, the argument may have some worth, 
provided it is not directed to condemning the work as bad art, but rather to 
arguing for restricting its communication to a public audience on grounds of 
danger to life. Classic cases of freedom of speech typically tread on the 
borderline of this kind of fallacy.  

The examples studied in this paper reveal that arguments from 
consequences are often reasonable in shifting the burden of proof under 
conditions of uncertainty and lack of knowledge. But these examples also show 
that such arguments are often dangerous in that they have a powerful 
persuasive effect that goes far beyond the force of rational persuasion that they 
should be properly taken to carry. The question is how they work as rhetorical 
strategies of strategic maneuvering (van Eemeren and Houtlosser, 2000). It 
may be suggested as a hypothesis, based on what has been shown, that the 
reason for this deceptiveness is that failures of relevance associated with these 
kinds of fallacies are concealed by dialectical shifts. According to this 
hypothesis, rhetorical persuasiveness of these arguments depends on concealed 
dialectical shifts that need to be brought out by analysis of cases, which show 
how criteria that need to be used to judge such arguments fallacious or not in 
specific cases must take such dialectical shifts into account.  

One problem with all the examples studied in this paper is that they are 
short, and appear to be quite simple on the surface, but once a critic attempts to 
analyze the example, it is found that there is a lot of implicit communication 
under the surface. All of the examples can be analyzed at two levels. At the 
inferential level, the fallacy of argumentum ad consequentiam can 
straightforwardly be identified as one of wrong conclusion. But to pinpoint in 
depth what is wrong, to teach students to properly identify and analyze the 
fallacy in such cases, and to grasp rhetorically how such arguments can be 
powerfully deceptive, one needs to proceed to the dialectical level. One needs 
to ask what conclusion the given argument actually does support. A problem 
exposed here is that the given argument may fit the argumentation scheme for 
value-based practical reasoning, and thereby give the respondent a reason for 
taking action or not. Looked at in this way, the argument seems reasonable, and 
indeed it may be a reasonable, provided the dialectical shift is not taken into 
account, and the prudential conclusion is taken as the real one to be proved. 
Once the shift is taken into account, the diagnosis of the failure is that meeting 
the requirement of fulfilling the burden of persuasion has not been met by the 
prudential argument that has been put forward. 
 
 
Acknowledgements 
 
The author would like to thank the Social Sciences and Humanities Research 
Council of Canada for a research grant that helped to support the work in this 
paper. An earlier version of the paper was read on February 14, 2007 as a 
keynote address at the International Conference on Strategies of 



Douglas Walton 82 

Argumentation held at the University of Groningen, held in conjunction Erik 
Krabbe’s Valedictory Lecture. I would like to thank Trevor Bench-Capon, Erik 
Krabbe, Henry Prakken and Bart Verheij for especially useful questions and 
comments on this earlier version. 
 
 
References 
 
Aristotle (1937). The Art of Rhetoric. Trans. John Henry Freese. Loeb Classical 

Library. Cambridge, MA: Harvard University Press. \ 
Atkinson, Katie, Trevor J. M. Bench-Capon and Peter McBurney (2006). 

Computational Representation of Practical Argument. Synthese 152, 157-
206.  

Bench-Capon, Trevor J. M. (2003). Persuasion in Practical Argument Using 
Value-based Argumentation Frameworks. Journal of Logic and 
Computation 13, 429-448. 

Eemeren Frans H. van and Rob Grootendorst (2004). A Systematic Theory of 
Argumentation. Cambridge: Cambridge University Press. 

Eemeren, Frans H. van and Peter Houtlosser (2000) Strategic Maneuvering in 
Argumentative Discourse: a Delicate Balance. In Frans H. van Eemeren 
and Peter Houtlosser (eds), Dialectic and Rhetoric: the Warp and Woof of 
Argumentation Analysis. Dordrecht: Kluwer, 131-159.  

Guerini, Marco and Cristiano Castelfranchi (2006). Promises and Threats in 
Persuasion. 6th Workshop on Computational Models of Natural Argument 
(CMNA 6). European Conference on Artificial Intelligence (ECAI 06). 
Riva del Garda, Italy: August 28. 

Jacobs Scott and Sally Jackson (2006). Derailments of Argumentation. In Peter 
Houtlosser and Agnes van Rees (eds), Considering Pragma-Dialectics. 
Mahwah, NJ: Erlbaum, 121-133. 

Krabbe, Erik C. W. (2003). Metadialogues. In Frans H. van Eemeren, 
J.Anthony Blair, Charles A. Willard and A.Francisca Snoek Henkemans 
(eds), Anyone Who Has a View: Theoretical Contributions to the Study of 
Argumentation. Dordrecht: Kluwer, 83-90. 

McBurney, Peter, David Hitchcock and Simon Parsons (2007). The Eightfold 
Way of Deliberation Dialogue. International Journal of Intelligent Systems 
22, 95-132. 

Nicoloff, Franck (1989). Threats and Illocutions’. Journal of Pragmatics, 13, 
501-522.  

Prakken, Henry (2006). Formal Systems for Persuasion Dialogue. The 
Knowledge Engineering Review, 21, 163-188.  

Rescher, Nicholas (1964). Introduction to Logic. New York: St. Martin's Press. 
Walton, Douglas (1995). A Pragmatic Theory of Fallacy. Tuscaloosa, AB: 

University of Alabama Press.  
Walton, Douglas (1996). Argumentation Schemes for Presumptive Reasoning. 

Mahwah, NJ: Erlbaum.  
Walton, Douglas (2004). Relevance in Argumentation. Mahwah, NJ: Erlbaum. 



          Dialectical Shifts Underlying Arguments from Consequences 
 
83

Walton, Douglas Scare Tactics: Arguments that Appeal to Fear and Threats. 
Dordrecht: Kluwer. 

Walton, Douglas (2006). How to Make and Defend a Proposal in Deliberation 
Dialogue. Artificial Intelligence and Law, 14, 177-239.  

Walton, Douglas and Erik C. W. Krabbe (1995). Commitment in Dialogue. 
Albany: State University of New York Press. 

Walton, Douglas, Chris Reed and Fabrizio Macagno (2008). Argumentation 
Schemes. Cambridge: Cambridge University Press.  

Wells, Simon and Chris Reed (2006). Knowing when to Bargain: the Roles of 
Negotiation and Persuasion in Dialogue. 6th Workshop on Computational 
Models of Natural Argument (CMNA 6). European Conference on 
Artificial Intelligence (ECAI 06), Riva del Garda, Italy, August 28.