ISSA Proceedings 2010 – On The Concept “Argumentum Ad Baculum”

1. Introduction
The aim of this paper is to question the value of the concept of the so-called “argumentum ad baculum” (appeal ‘to the stick’). This aim is distinct from the purpose of many earlier works that focused on analyzing whether appeals to threat are or are not fallacious and under which circumstances they might be justified (e.g. Wreen 1989, Levi 1999, Kimball 2006, Walton & Macagno 2007). Instead, this paper investigates whether there is a consistent phenomenon at all that can be called “ad baculum”.[i]

Of course, it must be recognized that any term (such as “ad baculum”) that is established and widely used in argumentation theory and rhetoric has a presumption of usefulness. It is therefore the burden of those who doubt the usefulness of the concept to show that it does significantly more harm than good for the discipline.[ii] Nevertheless, there are circumstances under which this burden of proof can indeed be satisfied. If a term obscures rather than explains the essential qualities of the phenomenon or phenomena it describes, then a discipline may be well advised in changing or abandoning it. One instance in which this might be the case is terms that unite concepts by addressing accidental rather than essential qualities. In the worst case these kinds of terms will unite phenomena under themselves that have very little in common with each other and only share one accidental quality.

To illustrate this point in an extreme case: I might observe that all of my friends by the name of Markus are very thin, nearly anorexic. I might even confirm this observation by looking for more Markuses and finding that most of them are also rather skinny. And I might even be statistically right in my belief that the average Markus is slimmer than the average citizen (due to, for example, the popularity of that name in a certain cohort or social group that is also prone to skinniness or anorexia). Still, I would be ill advised to talk of a “Markus figure” when describing the physique of somebody or analyzing the relationship between “Markusness” and skinniness because the group in question is united only by an accidental quality.

2. Terminology test
In order to analyze whether the term “ad baculum” is of the above kind, one must test it for two qualities:
a) do the phenomena commonly united under the name “ad baculum” share one common essential quality, and
b) could all significant instances of “ad baculum” also be described by other categories that might be more relevant?

The first of these tests can be performed by substituting the proposed essential quality (the occurrence of a warning, threat or other appeals to fear or reference of a potential undesirable outcome – to use the widest possible meaning of “ad baculum”) by another quality. If the phenomenon under scrutiny (i.e. the argument or fallacy) maintains most of its observed relevant aspects, then it is very likely that its ‘stickness’ is not essential and should therefore be avoided as a defining quality of the phenomenon.

The second test can inform us whether any separate term for the phenomena that are commonly referred to as “ad baculum” is needed at all. If “ad baculum” is indeed a term united only by accidental qualities and all phenomena to which it refers can be aptly and better described by other concepts, then one might be well advised to discontinue its use in contemporary argumentation theory.

3. Ad Baculum as Fear Appeals
In order to gain a better understanding of the way the term “ad baculum” is commonly used let us first turn to one of the most famous and perhaps oldest instances of a fear appeal in western rhetoric, the Melian dialogue by Thucydides. This dialogue is a semi-ficticious exchange by two parties (the Melians and the Athenians) during the course of the Peloponnesian war. Thucydides includes it in his history of the Peloponnesian war and gives us the background under which it supposedly occurred (Thucydides 1921, 155-177 / V,84-V,115): The Athenians had just landed with a large military force on the island of Melos and demanded the Melians to accept Athenian rule or else be attacked. The Melians ask the Athenians to discuss the matter with the leaders of Melos. The subsequent discussion contains a number of famous fear appeals that might be labeled instances of “ad baculum”. Three exemplary ones shall be singled out and paraphrased here:

Melian dialogue 1 (Athenians to Melians): If you do not accept our rule, we will forcefully subdue you.

Melian dialogue 2 (Melians to Athenians): If you attack us, Sparta will come to our help and defeat you or revenge us.

Melian dialogue 3 (Athenians to Melians): If we do not subdue you, our current subjects will revolt against Athenian rule.

According to Thucydides the two parties did not find a solution to their difference of opinion, the Melians insisting on their independence and the Athenians on their will to subdue Melos. Soon after, an Athenian military expedition attacked and conquered Melos, killed all Melian men, and sold the women and children into slavery.

Needless to say, there is something evidently revolting about this blatant use of violence. However, this aspect should not obscure the analysis of the dialogue. The three selected fear appeals above illustrate the scope of argumentative moves that can be covered by the term “ad baculum”. An appeal “to the stick” can be a warning (i.e. the potential negative consequence alluded to has not been created by the protagonist: e.g. MD2 & MD3) or a threat (i.e. the potential negative consequence alluded to has been created by the protagonist: e.g. MD1). It can refer to negative consequences independent of whether they will actually happen (e.g. MD1) or not (e.g. MD2). And the potential negative consequence can be a threat to either the protagonist (e.g. MD3) or (probably more commonly) to the antagonist (e.g. MD1 & MD2). These aspects show only part of the scope of what can be referred to as “ad baculum” and are by no means exhaustive. They do however serve as a useful reminder to the variety of different argumentative moves that feature some kind of fear appeal.

4. Extent Treatments of Fear Appeals
Of the three fear appeals above the first one (MD1) is probably the most evident instance of “ad baculum”. It would be treated as fallacious or otherwise problematic by most communication disciplines. The reasons for the negative judgment of this argumentative move are quite different however. This is not the place for an exhaustive comparison of the treatment of “ad baculum” of all disciplines and schools in question. For our purposes, it will be sufficient to illustrate that the very same phenomenon (“ad baculum” in MD1) can be categorized quite differently.

Formal Logic: Copi and Cohen treat the “ad baculum” only briefly. They consider its fallaciousness to be so evident as to make any further discussion of it superfluous: “The appeal to force is the abandonment of reason.” (Copi & Cohen 2002, 148). Their main objection to fear appeals (as well as to related “ad” fallacies such as “ad hominem”, “ad populum” and “ad misericordiam”) seems to be the lack of a relevant argument scheme under which they can be subsumed. That this criticism does not hold true for all kinds of fear appeals will be shown below.

Informal Logic: There are a variety of different approaches to the “ad baculum” in informal logic. These offer different reasons for its fallaciousness and some distinguish between fallacious and non fallacious uses of fear appeals. Perhaps one of the most interesting explanations is offered by Douglas Walton. He treats some instances of “ad baculum” (presumably including MD1) as improper dialectical shifts from persuasive dialogue types into negotiation or bargaining (comp. Walton 2000, 180ff, Walton & Macagno 2007, 72ff.). In this approach the fear appeal itself is constructed as being less problematic than then pretense of engaging in one dialogue type although using the techniques of another type.

Pragma-dialectics: Van Eemeren and Grootendorst treat instances of “ad baculum” as a violation of rule 1 of the set of rules for a critical discussion (Eemeren & Grotendorst 1992, p. 212): “Parties must not prevent each other from advancing standpoints or casting doubt on standpoints” (Van Eemeren & Grotendorst 1992, p. 208). According to this approach instances of “ad baculum” are fallacious because they hinder the solution of a difference of opinion by breaking the rules for a critical discussion. Discussion partners that are interested in an optimal resolution process must therefore avoid resorting to or permitting open threats to the other party.

Rhetoric: Due to the large variety of different approaches referred to as rhetoric it is impossible to chose any one representative rhetorical treatment of “ad baculum”. Rhetoric understood as ars persuadendi or “the art of influencing the Will” (Whately 1963, p. 175) might not be a very good choice for the analysis of fallacies in the first place, because it lacks much of the normative elements of the approaches mentioned above. In its most radical version, rhetorical theory might well endorse any communicative act that leads to persuasive success. But even under these circumstances some argumentative moves might still be considered problematic. If (as is the case in the “ad baculum” in MD1) a potential persuasive effect is very limited in its reach, in other words it is not stable and not replicable, then it might be considered defective. From a rhetoric perspective, MD1 could be considered deficient because its persuasiveness depends on the maintenance of an immediate and credible threat, a quality that makes this form of persuasion very expensive and at the same time less stable than conventional argumentation that does not depend on altering external states.[iii]

5. A Taxonomy of Ad Baculum
What is interesting about the approaches above is that, for most of them, the threat itself is not the main problem but rather a symptom of an underlying issue: (i) a problematic shift in dialogue types; (ii) an obstacle to the free participation of a critical discussant or (iii) an instable form of persuasion, that can be triggered similarly by a number of non-threat related moves. The variety of underlying issues is an indicator for the heterogeneous character of the phenomena referred to as “ad baculum”. Of course the three problems mentioned above are far from constituting a complete list of underlying issues that can be found in instances of “ad baculum”. To start with they do not take into consideration the full breadth of different kinds of “ad baculum”.

In order to understand what kind of issues can be underneath the various instances of “ad baculum”, a brief taxonomy of the most important aspects of fear appeals will be helpful. As indicated above, a fear appeal might take the form of a threat (negative consequence created by persuader) or a warning (negative consequence not created by persuader). Furthermore the threat can be credible (the persuader is planning to bring about the negative consequence in case the addressee responds in the wrong way) or empty.[iv]

The aforementioned division between fear appeals addressed to the protagonist and those addressed to the antagonist is of little significance for the analysis of underlying issues. One further division that should be taken into account, however, is the type of persuasive goal that is being pursued by the protagonist of the fear appeal. It is of central importance insofar as it captures a number of the more absurd examples used in certain (more hostile) treatments of fear appeals in the literature. The persuasive goal is of two basic types: either a change of belief in the antagonist or the performance of an action.

Taken together the three divisions create the basic taxonomy of fear appeals below (Figure 1):

Figure 1

6. Consequences of the Taxonomy of Ad Baculum
This brief taxonomy enables us to give a more complete analysis of the underlying issues in different kinds of appeals ‘to the stick’. If it can be shown that all types of “ad baculum” are either valid arguments following a standard argument scheme or are fallacious due to reasons that are independent of the threat itself, then we can assume that we do not need “ad baculum” as a separate concept to describe any instances of fear appeal. While this would make the concept of “ad baculum” superfluous, showing as much would not yet be sufficient to claim that the use of the concept would actually be harmful. This claim requires additional reasons to be considered in a later step.

Depending on the theoretical starting point and perspective, any one type of fear appeal above might be fallacious or deficient for more than one reason. That reason is independent of the accidental quality of a threat or a fear appeal if it can be easily fulfilled or triggered by non-threat related aspects. Also the fact that a type can generally be subsumed under a valid argument scheme does not of course mean that any instance of that type would be a strong argument. It would rather mean that it can be tested by means of the critical questions associated to that scheme.

The following types of fear appeals are covered by threat independent fallacies or standard argument schemes:
1) Reverse naturalistic fallacy (types 1a / 1b / 1c): Types 1a, 1b and 1c are forms of reasoning from the desirability of a proposition onto the plausibility of a proposition. As such, they are the mirror image to the better-known “is-ought” problems (“It exists in nature therefore it is good”) and are just as fallacious. Any change in the quality of the proposition within an argument leaves that argument worthless. This type of fallacy or argumentative deficit is entirely independent of threats and can be reproduced in any argument scheme with a variety of propositional quality changes. Most of the resulting forms of reasoning are however so blatantly fallacious that they are not at all likely to fool any intelligent addressee. Woods’ example of an “ad baculum” of this type is “If you do not fully and sincerely believe proposition p is true then I will insult your sister” (Woods 1998, 496). It is easy to imagine very similar fallacies without the use of threats such as “If you do fully and sincerely believe proposition p is true then I will buy you a car,” or “Professor Woods says that it would be nice if proposition p were true, therefore you should believe that proposition p is true.” Instances of 1a, 2a and 3a would therefore (i.e. because they are reverse naturalistic fallacies) be fallacious independent of whether they include references ‘to the stick’ or not.

2) Truth claim negotiations (types 1b / 1c): In addition to being reverse naturalistic fallacies, types 1b and 1c have another significant argumentative deficit. In introducing potential negative consequences into the discourse that are created purely for the purpose of persuasion, the protagonist leaves the discourse type of pure argumentation and enters the type of negotiation or bargaining. Negotiation or bargaining are, however, inadequate discourse types when it comes to truth claims. Once again this deficit or fallacy is quite independent of the involvement of threats. There is no relevant structural difference between “If you do not believe proposition x then I will hit you” (“ad baculum”) and other forms of negotiation, such as “If you believe in proposition x then I will believe in proposition y.”[v]

3) Empty threats (types 1c / 2c): Types 1c and 2c are appeals to threats that are unlikely to materialize even if the addressee of the threat does not act in accordance with the persuader’s interests. Put another way, these empty threats are blatant lies that try to create a wrong appearance for the purpose of persuasion. As such they are once again essentially independent of the threat itself. Most conversational standards or normative systems include a rule or regulation that bans putting forwards standpoints or arguments that the protagonist believes to be false or for which he lacks sufficient evidence (e.g. Grice’s Quality Maxim, Grice 1975, 46). Any blatant lie, whether it refers to an empty threat, an empty promise, or any other faulty statement would be a breach of those rules.

4) Freedom of speech violations (types 2b / 2c): Some instances of types 2b and 2c can be attempts to stop an antagonist in a discussion from advancing standpoints or casting doubt on standpoints of the opponent. In that case, they violate rule 1 of the critical discussion, and according to the pragma-dialectical theory would hinder the effective solution of a difference of opinion. While threats can certainly lead to a violation of the pragma-dialectical freedom rule, they are by no means the only (and probably not even the most important) form of a rule 1 violation. Many forms of diminishing the freedom of speech of the opponent are easy to imagine that do not involve any form of fear appeal. A hearty laughter at any word of the opponent would be just one example of this kind of fallacy that is independent of any appeal ‘to the stick’.

5) Instable persuasion / dialectical shifts (types 2b / 2c): Seen from a rhetorical perspective, many instances of types 2b and 2c will also be deficient forms of persuasion. As noticed above, argumentative moves which force the protagonist to alter external states for the purpose of persuasion are generally less stable and considerable more expensive than pure argumentation. While not necessarily constituting a fallacy, this fact makes any avoidable shift from pure argumentation to negotiation and bargaining undesirable. A similar concern can be expressed from an informal logical perspective about inappropriate dialectical shifts away from a critical discussion to a negotiation (Walton 1992, 141ff.). In both cases the underlying problem (the inappropriate shift) is independent of the presence of a threat or warning and can be caused by a variety of other factors as well.

6) Causal argumentation (effect to cause, type 2a): The only remaining type of fear appeal that is not covered by one or more threat independent kinds of fallacy[vi] is type 2a. The obvious reason for this is the fact that although this kind of reasoning does indeed include an appeal “to the stick,” it follows a perfectly valid argument scheme. Depending on the taxonomy of argument schemes one wants to employ, type 2a might be called an argument from consequences (Walton 1995, 218ff., Walton 2000, 132ff.), argument from prudence (Woods 1998, 496) or simply a type of weak causal argument (in this case an argument from effect to cause, Herrmann et al. 2010, 58ff.).[vii]  The questions which of those (very similar) argument schemes best represents fear appeals does not need to be settled for the purpose of this paper because the testing procedure would be similar for all of them. In order to test whether any particular fear appeal of type 2a is a strong or weak argument one only needs to employ the set of critical questions for that scheme (as well as potentially additional critical questions that are scheme independent). Those critical questions (e.g. Kienpointner 1996, 156f., Walton 1996b, 75ff., Walton 2000, 137ff., Herrmann et al. 2010, 58ff.) are a sufficient testing tool for any given argument scheme and do not need any “ad baculum” specific supplement.

7. Conclusions
The analysis of the taxonomy of fear appeals above indicates a few important conclusions: First, all instances of appeals ‘to the stick,’ be they fallacious or non-fallacious ones, can be covered and analyzed by categories that are entirely threat- or warning independent. Second, not only is the category of “ad baculum” superfluous, but it might be positively obscuring the analysis of a given fallacious move because it offers too simple an answer to questions about the underlying reasons for the fallaciousness of the move at hand. Third, refraining from labeling a certain argumentative move “ad baculum” facilitates the distinction between fallacious and non-fallacious fear appeals because the latter are not already stigmatized by a negatively laden term.

It is this last point that also answers the remaining question: Even if it might be the case that all phenomena that are commonly united under the label “ad baculum” can be sufficiently (and perhaps even more precisely) covered by other categories, does that mean that the use of the term “argumentum ad baculum” in contemporary argumentation theory is positively harmful? Yes. Inasmuch as the term unites phenomena by only accidental qualities, it obscures the analysis of potential underlying problems in different types of fear appeals and most importantly produces a ‘guilt by association’ type prejudice against proper uses of fear appeals one might be well advised to avoid using this term for the purposes of contemporary argumentation analysis or at least supplement any use of the term with a more detailed description of the specific type referred to and the theoretical perspective used.[viii]

This conclusion, which suggests the abandonment of the term “argumentum ad baculum” as an umbrella term for very different kinds of fear appeal only extends to this particular fallacy. The method of dividing a particular fallacy (in this case the “ad baculum”) into its underlying types and analyzing each type independently, might well be useful for criticizing other “ad” fallacies as well.[ix] The result of these analyses would probably be different for different fallacies.

[i] This paper assumes a basic familiarity with the idea of a fear appeal or “argumentum ad baculum”, literally translated as appeal “to the stick”. For a historical introduction to the concept see, among others, Hamblin (2004) p. 135ff., Woods (1998) p. 494ff., Walton (2000) p. 31ff., van Eemeren (2001) p. 135ff. and van Eemeren et al. (2009) p. 2ff.
[ii] It must, of course, also be acknowledged that terminology in argumentation theory and rhetoric does not always have to follow the same standards of rigidity as do similar concepts in some hard natural or mathematical sciences.
[iii] In other words it constitutes a shift away from pure argumentation into negotiation. For the purposes of this paper „argumentation“ is used in the sense of „mean of enforcing the will against resistance by changing the state of information in a reasonable way“; „negotiation“ is used as „mean of enforcing the will against resistance by exchanging costs and benefits“.
[iv] Technically speaking the same division holds true for warnings, but for reasons given below, this division is practically irrelevant because the plausibility of the manifestation of negative effect is one of the components of the critical testing of the argument scheme that is used for a warning.
[v] This is not to be confused with a negotiation of the type of “If you refrain from challenging proposition x in this discussion then I will refrain from challenging proposition y.” which, despite sounding rather similar and producing comparable practical results, aims at a particular action rather than a belief of the discussants.
[vi] The list above should be more than sufficient for the purpose of the main claim, namely that all types are covered by at least one fear appeal independent underlying deficient. This should be no means suggest the completeness of the list of reasons however. A number of further reasons for the fallaciousness of certain types of fear appeals have been suggested of which Woods concept of a “veiled intimidation ad baculum” is probably one of the most prominent (Woods 1998, 497).
[vii] Strictly speaking all of the types above can be reconstructed as moves resembling a causal argument from effect to cause. The general structure of all types would be: “Action / belief C will lead to consequence E. E is undesirable. Therefore C is undesirable.” The first part “Action / belief C” would in that case form the y-axis and the second part “will lead to” would form the x-axis of our taxonomy. Since the critical testing of all types other than 2a would however very quickly reveal grave deficits it seems more useful for the purpose of this paper to treat as separate types straight away.
[viii] This conclusion does not mean to suggest that certain types of threat appeals cannot be fruitfully analyzed or researched. It merely suggests that in order to be consistent one must limit oneself the one of the types (e.g. a freedom rule violation, compare van Eemeren et al. 2009, 85ff.) rather than consider the heterogeneous field of so-called “ad baculum” moves.
ix A similar approach is being employed in the distinction between the three kinds of “ad hominem” (abusive, circumstantial and tu quoque). Without this distinction the term “ad hominem” would also be too heterogeneous to be useful for contemporary argumentation theory.

Copi, I. M & Cohen, C. (2002). Introduction to Logic. Upper Saddle River: Pearson.
Eemeren, F. H. van (2001). Fallacies. In: Ibid. (ed.), Crucial Concepts in Argumentation Theory (pp. 135-164), Amsterdam: Amsterdam University Press.
Eemeren, F. H. van & Grootendorst, R. (1992). Argumentation, Communication and Fallacies. Hillsdale: Lawrence Erlbaum Associates.
Eemeren, F. H. Van, Garssen, B. & Meuffels, B. (2009). Fallacies and Judgments of Reasonableness. Dordrecht et al.: Springer.
Grice, H. P. (1975). Logic and conversation. In: P. Cole & J.L Morgan (Eds.), Syntax and Semantics, III: Speech Acts (pp. 41-58), New York: Academic Press.
Hamblin, C. L. (2004). Fallacies. Newport News: Vale Press.
Herrmann, M., Hoppmann, M, Taraman, J. & Stölzgen, K. (2010). Schlüsselkompetenz Argumentation. Parderborn: UTB Schöningh.
Kienpointner, M. (1996). Vernünftig argumentieren. Hamburg: Rowohlt.
Kimball, R. H. (2006). What’s Wrong with Argumentum ad Baculum? Reasons, Threats, and Logical Norms. Argumentation 20, 89-100.
Levi, D. S. (1999). The Fallacy of Treating the Ad Baculum as a Fallacy. Informal Logic 19, 145-159.
Thucydides (1921). History of the Peloponnesian War. Vol. III. Transl. by C. F. Smith, Cambridge: Harvard University Press.
Van de Vate, D., Jr. (1975). The Appeal to Force. Philosophy & Rhetoric 8, 43-60.
Walton, D. (1992). Types of Dialogue, Dialectical Shifts and Fallacies. In: F. H. van Eemeren, R. Grootendorst, J. A. Blair & C. A. Willard (Eds.), Argumentation Illuminated (pp. 133-147), Amsterdam: SICSAT.
Walton, D. (1995). A Pragmatic Theory of Fallacy. Tuscaloosa & London: University of Alabama Press.
Walton, D. (1996a). Practical Reasoning and the Structure of Fear Appeal Arguments. Philosophy and Rhetoric 29, 301-313.
Walton, D. (1996b). Argument Schemes for Presumptive Reasoning. Mahwah: Lawrence Erlbaum.
Walton, D. (2000). Scare Tactics: Arguments that Appeal to Fear and Threats. Dordrecht: Kluwer.
Walton, D. & Macagno, F. (2007). The Fallaciousness of Threats: Character and ad Baculum. Argumentation 21, 63-81.
Whately, R. (1963). Elements of Rhetoric. D. Ehninger (Ed.), Carbondale: Southern Illinois University Press.
Woods, J. (1998). Argumentum ad baculum. Argumentation 12, 493-504.
Wreen, M. J. (1989). A Bolt of Fear. Philosophy & Rhetoric 22, 131-140.