105
Fallacies in Ad Hominem Arguments
1
Falacias en argumentos ad Hominem
Christian Dahlman
Faculty of Law, Lund University, Lund, Sweden
[email protected]
David Reidhav
Faculty of Law, Lund University, Lund, Sweden
[email protected]
Lena Wahlberg
Faculty of Law, Lund University, Lund, Sweden
[email protected]
Received: 25-07-2011 Accepted: 14-12-2011
Abstract: Arguments ad hominem are common in political debates, legal argumen-
tation and everyday conversations. In this article, we propose a general denition
of ad hominem arguments. An argument ad hominem is an argument that makes a
claim about the reliability of a person in the performance of a certain function, based
onsomeattributerelatingtothepersoninquestion.Onthebasisofthisdenition,
we examine the different ways that ad hominem arguments can go wrong, and classify
them as seven different ad hominem fallacies: false attribution, irrelevant attribute,
overratedeffect,reliabilityirrelevance,irrelevantperson,insufcientdegreeandir-
relevant function. The various fallacies are illustrated with examples from politics,
law and everyday life.
Keywords: Ad hominem, fallacy, argumentation, personal attack, reliability, testi-
mony.
1
The research presented in this article was funded by Torsten och Ragnar Söderbergs
Stiftelser. We owe thanks to Niklas Arvidsson, Roberta Colonna Dahlman, Eveline Fet-
eris, Åke Frändberg, Tobias Hansson Wahlberg, Patricia Mindus, Antonino Rotolo, Stefan
Schubert, Torben Spaak, Lennart Åqvist and two anonymous referees for helpful com-
ments on earlier drafts.
COGENCY Vol. 3, N0. 2 (105-124), Summer 2011 ISSN 0718-8285
106
COGENCY Vol. 3, N0. 2, Summer 2011
Resumen: Los argumentos ad hominem son comunes en debates políticos, en la
argumentación legal y las conversaciones cotidianas. En este trabajo, proponemos
unadenicióngeneraldeargumentosad hominem. Un argumento ad hominem es
unargumentoquemantieneunapretensiónsobrelaconabilidaddelapersonaenel
ejercicio de cierta función, basada en algunos atributos relacionados con la persona
encuestión.Sobre labasedeestadenición,examinamos lasdiferentesformasde
los argumentos ad hominemquepuedenserdefectuosos,ylosclasicamosensiete
diferentes falacias ad hominem: atribución falsa, atribución irrelevante, efecto de so-
breestimación,irrelevanciadeconabilidad,personairrelevante,gradoirrelevantey
función irrelevante. Estas falacias son ilustradas con ejemplos de la política, la ley y
la vida cotidiana.
Palabras clave:Adhominem,falacia,argumentación,ataquepersonal,conabili-
dad, testimonio.
1. Introduction
One of the most common ways to argue against a claim is to attack it indi-
rectly, by attacking the person who makes the claim. Examples are easy to
ndinpoliticaldebates,courtroomsanddinnerconversations.Apolitician
who claims that the government should cut taxes is countered with the ar-
gumentthatheendorsestaxcutsbecauseitwouldbenethimpersonally.
The testimony of an eye witness in a murder trial is attacked by the defense
attorney, with the argument that the witness is unreliable because he is a
drug addict. A guest at a dinner party who says that people in rich countries
should do more for the people in poor countries is met with the argument
that she does not do much for poor people herself.
Counter arguments of this kind are usually dismissed as bad arguments.
They are rejected as unfair and irrelevant, and it is said that they commit
the ad hominem fallacy. This reaction gives rise to questions. Are these
arguments always fallacious? Is the attack always aimed at the opposing
arguer in the discussion? Are all arguments that commit the ad hominem
fallacy fallacious in the same way? Our answer to each of these three ques-
tions is negative. An ad hominem argument is not always fallacious. Some
ad hominem arguments attack a person who is not a participant in the dis-
cussion. And the ad hominem arguments that are fallacious can commit
a number of different fallacies. This article will explain and develop these
views.
107
That an argument which seeks to undermine a claim by attacking the
person who makes the claim is not always fallacious has been pointed out by
many other scholars. It can, for example, make good sense to challenge the
testimony of an eye-witness with the observation that the witness is almost
blind. If a speaker claims that she should be relied on, it becomes relevant
to challenge that claim with an argument that attacks her reliability. Some
authors, including John Woods (2010, p. xxv), Douglas Walton (1998, p.
125), Alan Brinton (1995, p. 215) and Bruce Waller (2005, pp. 180-192),
say that this shows that an argument ad hominem is not always fallacious.
In their treatment of ad hominem arguments, there are ad hominem argu-
ments that are legitimate and ad hominem arguments that commit the ad
hominem fallacy. Others, notably Irving Copi and Carl Cohen (2002, pp.
143-145), prefer to reserve the term ad hominem for arguments that com-
mit the ad hominem fallacy: If an argument is not fallacious it is not an ad
hominem argument. This difference in terminology can be confusing, and
itisimportanttobearinmindthatitdoesnotreectasubstantialdisagree-
ment over legitimacy. Like Woods, Walton, Brinton and Waller we prefer
to distinguish between ad hominem argument and ad hominem fallacy. We
thereforebeginourinvestigationbydevelopingageneraldenitionofad
hominem argument. We then move on to investigate the conditions under
which these arguments are fallacious.
Our approach to ad hominem arguments differs from the standard ap-
proach taken in textbooks on fallacies. The main structure in the standard
approach is the division of ad hominem arguments in the subtypes known
as abusive ad hominem and circumstantial ad hominem, along with the
additional subtype known as tu quoque. Our approach is structured around
ageneraldenition:anargumentad hominem is an argument that points
out that a person has a certain attribute and claims that this affects the reli-
ability of the person with regard to the performance of a certain function.
Inthefollowing,wewillrelateourframeworktothestandardclassication
at appropriate points.
In our account, an ad hominem argument does not necessarily attack
the opposing arguer in the discussion. A personal attack on someone who
does not participate in the discussion is also an ad hominem argument.
Such attacks are often used as counter arguments against an authority ar-
gument. Think, for example, of a discussion where one of the participants
Fallacies in Ad Hominem Arguments / C. DaHlman, D. reiDHav and l. WaHlberG
108
COGENCY Vol. 3, N0. 2, Summer 2011
backs up his claim by referring to Professor X as an authority, and the other
participant replies with a personal attack against Professor X, intended to
undermine the Professor’s reliability as an authority. In our understanding
this is clearly an argument ad hominem, but there are scholars who have
a different view. According to Frans van Eemeren, Rob Grootendorst and
other members of the pragma-dialectical school this kind of argument does
not count as an argument ad hominem, as it attacks a person outside the
discussion.Accordingtotheirdenition,anargumentad hominem is an
argumentthatviolatestherstruleofdiscussion,that“partiesmustnot
prevent each other from advancing standpoints or casting doubt on stand-
points” (see, e.g., van Eemeren and Grootendorst, 1992, pp. 153-155). In
our view, this makes the analysis of ad hominem arguments too narrow. As
we understand it, the soundness of a personal attack depends on a number
of factors, but participation in the discussion is not one of them. It makes
no difference for the soundness of the argument whether the person under
attack is a participant in the discussion or not.
In our analysis, an argument that commits the ad hominem fallacy can
be fallacious in seven different ways. We will identify each of these errors
and analyze them as seven separate ad hominem fallacies. This means that
our account of how ad hominem arguments can be fallacious is broader
thantheexpositionyouwouldndinmosttextbooks.Itisoftensaidthat
ad hominem arguments are fallacious because they try to attack a claim by
attacking the reliability of the person who makes the claim, in spite of the
fact that this is irrelevant to the correctness of the claim (see, e.g., Hamblin,
1970, p. 41; Walton, 1987, p. 317; Copi and Cohen, 2002, p. 143). In our
account, this is just one of the seven ways in which an ad hominem argu-
ment can be fallacious. It will be analyzed in section three as “the fallacy of
reliability irrelevance”.
At the outset, our approach is similar in some ways to Brinton’s treat-
ment of ad hominem arguments. Like Brinton we seek to identify different
ways in which ad hominem arguments can go wrong. An important differ-
ence between our approach and Brinton’s is that we analyze all these errors
as false premises. As we shall see, we represent ad hominem arguments as
deductively valid arguments and understand every ad hominem fallacy in
terms of a false premise. This approach is somewhat unorthodox. The ad
hominem fallacy is normally assumed to involve an illegitimate move from
109
premise to conclusion. It is important to understand, however, that this
difference is a matter of presentation. To describe an ad hominem fallacy
as a false premise instead of an illegitimate move from premise to conclu-
sion is just a different way of describing the same thing. To see this, con-
sider the following reasoning: “What the witness says is not true. He is an
alcoholic.” This might be held to commit an ad hominem fallacy because
it makes an illegitimate move from premise to conclusion: the conclusion
(what the witness says is not true) cannot be derived from the premise (the
witness is an alcoholic). A different way to say the same thing is to say that
the argument is fallacious because the premise that would make it legiti-
mate (what an alcoholic says is not true) is false. It is exactly because this
premise is false that some scholars say that we are dealing with an illegiti-
mate move from premise to conclusion.
We believe our approach has several advantages. By conceptualizing
every ad hominem error as the result of a false premise we are able to give
a uniform treatment of the various ways in which ad hominem arguments
can go wrong. We are also in a position to classify these errors not only on
basis of which premise of the argument is false, but also in terms of the
grounds on which the premise is false. This makes it possible to give a com-
prehensive and detailed description of the various forms of error.
2. Denition of Argument ad Hominem
Weproposethefollowinggeneraldenition:anargumentad hominem is
an argument that makes a claim about the reliability of a person, P, based
on an attribute, A, of P. The argument points out that P has A, and it claims
that, because P has A, P’s reliability in the performance of a certain func-
tion, F, amounts to the degree D. Ad hominem argumentation, in its basic
form, can be represented as follows.
Argument ad hominem (basic form):
(i) P has A
(ii) If P has A, the reliability of P in the performance of F is D
(iii) The reliability of P in the performance of F is D
Fallacies in Ad Hominem Arguments / C. DaHlman, D. reiDHav and l. WaHlberG
110
COGENCY Vol. 3, N0. 2, Summer 2011
Wewillspeak oftherst premiseasthe attribution premise and the
second premise as the effect premise.
As is indicated by the term “ad hominem” P is a person. However, tradi-
tionally, arguments in which P is a group of people or an organization have
been categorized as ad hominem as well, and we see no reason to break
with this tradition. By “attribute” we mean a property, in the widest sense
of the term, which can be ascribed to the person P. An attribute can, for
example, consist in having bad eyesight, being a compulsive liar or being in
Lund. Some of the things we here call attributes are normally referred to,
not as attributes or properties, but rather as “circumstances”. For example,
the fact that a witness was in Lund at a certain time would be described
by lawyers as a “circumstance”. However, in order to maintain a coherent
terminology, we will treat any fact that can be ascribed to a person as an
attribute.
An argument ad hominem is an argument about reliability. It is an ar-
gument claiming that a certain attribute has a certain effect on a person’s
reliability. This is not always apparent when one looks at the way ad homi-
nem arguments are described, but a number of other scholars who describe
the effect of ad hominem arguments regard them broadly in these terms.
Thus Brinton (1985, p. 56) says ad hominem arguments typically question
someone’s credentials as a speaker, and Walton (1998, pp. 273-278) ob-
serves that they seek to affect the opponent’s “credibility”.
2
Reliability depends on ability as well as motivation. That a person is
reliable in the performance of a function means that he or she is able to
perform the function and motivated to do so. A witness, for example, is
reliable when he or she has the cognitive ability to observe and remember
events correctly as well as the motivation to tell the truth. As we shall see,
some ad hominem arguments are aimed at ability while others are aimed
at motivation.
Accordingtoourdenitionad hominem arguments target P’s reliability
with respect to a certain function, F. In this formulation F can stand for
2
That the arguments we are interested in are fundamentally about reliability was de-
tected by Bentham. He does not use the term ad hominem at all. Instead he speaks of
arguments that commit this kind of fallacy as modications of the “fallacy of distrust”
(1824/1952, pp. 83-92; pp. 100-102).
111
anyfunctionthatapersoncanbereliedupontofulll.Itcanstandforthe
function of serving as an eyewitness in a court of law, but it can just as well
stand for the function of being Prime Minister. This makes our account of
ad hominem argument broader than Brinton’s. Brinton (1995, pp. 213-215)
describes an argument ad hominem as an argument that attacks someone’s
reliability as an advocate. In our account, F can stand for the function of
advocating a certain proposition, but it can also stand for other functions
that can be related to reliability. When F is the Prime Ministerial function,
itstandsforcondenceinPtoexercisegoodjudgmentonpoliticalissues.
When F is the function of serving as an eye witness, it stands for reliability
to give an accurate account of certain events. In our account “credibility”
(i.e. reliability as a source of information) is thus a sub-class of reliability in
the performance of function.
3
An argument that is not about reliability is not an argument ad homi-
nem. The argument that Pedro is older than me because he was born in
1956 is not an argument ad hominem. Being older than someone is not a
question of reliability. In some cases, it is a matter of interpretation wheth-
er an argument is an argument about reliability. Consider the argument
“Alfred can’t run fast. He has a bad leg”. This is not necessarily an argu-
ment about reliability, but it can be interpreted as such. It can be inter-
preted as an argument to the effect that Alfred cannot be relied on to run
fast. We could, for example, imagine that Alfred is being considered for the
position of courier. Someone is needed to deliver an important message
beforeacertaintime.Asweallknow,interpretationisdifcult.Forpresent
purposes this need not worry us. That an argument ad hominem is an argu-
ment about reliability simply means that an argument that is interpreted
as an argument ad hominem is an argument that is interpreted as an argu-
ment about reliability.
The effect that A has on P’s reliability can be positive or negative: it can
3
According to Brinton arguments directed at functions which are merely accidentally
associated with human beings are not ad hominem. He concludes (1995, pp. 213-214) that
only arguments directed at advocacy qualify as ad hominem arguments. Given the struc-
tural similarities of arguments about reliability in the performance of functions of all kinds,
an account which is applicable to any of these functions is in our view methodologically
preferable. Besides, we are not convinced that advocacy is an essential attribute for being a
person, as Brinton assumes.
Fallacies in Ad Hominem Arguments / C. DaHlman, D. reiDHav and l. WaHlberG
112
COGENCY Vol. 3, N0. 2, Summer 2011
improve P’s reliability, or it can or reduce it. We will talk about arguments
where the effect is positive as positive ad hominem arguments and argu-
ments where the effect is negative as negative ad hominem arguments.
Ad hominem arguments are often used as counter-arguments to claims
aboutP’sreliability.Theyarethereforeoftendenedascounter-arguments.
Inourview,thismakestheanalysistoonarrow.Givenourdenition,anad
hominem argument does not need to be a counter-argument. As we have
seen above, the ad hominem argument in its basic form is an argument in
its own right. The claim that it makes about P’s reliability need not stand in
opposition to some other claim about P’s reliability.
An ad hominem argument that targets a particular claim, and which
draws the conclusion that the effect that A has on P’s reliability makes this
target claim false, can be represented as an extension of the ad hominem
argument in its basic form.
Argument ad Hominem (extended form):
(i) P has A
(ii) If P has A, the reliability of P in the performance of F is D
(iii) If the reliability of P in the performance of F is D, the target claim
is false
(iv) The target claim is false
The ad hominem argument in its extended form has three premises: the
attribution premise and the effect premise from the basic form, plus a third
premise that we shall label the rebuttal premise.
The following dialogue serves as an example.
X: “Mrs. Borg is very reliable as an eyewitness.”
Y: “No, she is not. She has poor eyesight.”
The argument advanced by Y is an ad hominem in the extended form. It
is a counter-argument to the claim that Mrs. Borg is very reliable, and it
draws the conclusion that the target claim is false. In full, the argument of
Y in the dialogue is:
113
(i) Mrs. Borg has poor eyesight.
(ii) If Mrs. Borg has poor eyesight, the reliability of Mrs. Borg as an eye
witness is low.
(iii) If the reliability of Mrs. Borg as an eye witness is low, the claim
“Mrs. Borg is very reliable as an eye witness” is false.
(iv) The claim “Mrs. Borg is very reliable as an eye witness” is false.
3. Ad Hominem Fallacies
An argument ad hominem can go wrong in many different ways. We shall
now identify seven kinds of error. We talk about these errors as ad homi-
nem fallacies, and we refer to them by the following names.
1. the fallacy of false attribution
2. the fallacy of irrelevant attribute
3. the fallacy of overrated effect
4. the fallacy of reliability irrelevance
5. the fallacy of irrelevant person
6. thefallacyofinsufcientdegree
7. the fallacy of irrelevant function
Each fallacy results from the presence of a false premise in the ad homi-
nem argument. An argument that commits the fallacy of false attribution
is unsound because the attribution premise is false. In arguments commit-
ting the fallacy of irrelevant attribute or overrated effect the effect premise
is false. Arguments that commit the fallacy of reliability irrelevance, irrel-
evantperson,insufcientdegreeorirrelevantfunctionhaveafalserebuttal
premise. In its basic form ad hominem argumentation can commit only
therstthreefallacies.Sincethelastfourfallaciesconsistinafalserebut-
tal premise, they cannot be committed by arguments in the basic form. Ad
hominem arguments in the extended form, on the other hand, can commit
any of the seven fallacies.
The fallacy of false attribution is committed when the argument claims
that P has an attribute P in fact lacks. Here is an example.
Fallacies in Ad Hominem Arguments / C. DaHlman, D. reiDHav and l. WaHlberG
114
COGENCY Vol. 3, N0. 2, Summer 2011
“Barack Obama shouldn’t be trusted as Commander in Chief. He is a
Muslim, you know.”
The fallacy of irrelevant attribute is committed in cases when A has no
effect on P’s reliability in the performance of F. Here is an obvious example.
“You cannot take him seriously as an expert on foreign policy. He looks
like an aardvark.”
This kind of error in ad hominem argumentsisalsoidentiedbyWoods
(2010, pp. 98, 105). As the example shows, sometimes a negative ad homi-
nem argument that commits the fallacy of irrelevant attribute is not only a
bad argument, but a downright insult. And indeed arguments of this kind
areclassiedas“abusive” ad hominem arguments by Copi and Cohen (2002,
p. 143). However, the fallacy of irrelevant attribute is not always abusive. It
is not abusive in positive ad hominem arguments, of course, and there are
cases where a negative ad hominem argument commits the fallacy of ir-
relevant attribute without being abusive. Consider the following example.
“ItistruethatMr.Berghastestiedthattheplaintiffwasbleeding,but
we should not rely too much on this testimony, since Mr. Berg is not a
medical expert.”
This argument commits the fallacy of irrelevant attribute (you do not need
to be a medical expert to recognize that someone is bleeding), but it is not
abusive. The mere observation that someone is not a doctor is not an insult.
The effect of an attribute depends on the context (see, e.g., Walton,
1987, p. 323). In this article, we treat context as a bundle of attributes. An
ad hominem argument must be understood as being made relative to an as-
sumption about a certain bundle of attributes that includes A. This bundle
determines the effect that the attribute is assumed to have on P’s reliability.
We propose to label this bundle B
AH
. Note that this is the assumed context,
and that it may not correspond with the actual context. For example, the
argument might be made on the mistaken assumption that a person has
poor eyesight. Given this mistaken assumption, the effect of the circum-
stance that the eyewitness was not close to the scene of the crime (A) is
likely to be overrated.
115
We use the label B
A
for the bundle of attributes that actually holds for
P. On our account this actual context determines the real effect of A on P’s
reliability. It is against the background of B
A
that an ad hominem argument
should ideally be assessed. In other words, a fallacy can occur because re-
liability is assessed on the basis of a non-actual bundle of attributes, or
because the relevance of A in the actual bundle is estimated inaccurately
by the arguer.
The following sets of relationships—between A, the degree of reliability
in a particular function F, and a particular person P with his or her actual
bundle of attributes (B
A
) including A — are conceivable.
(a) The attribute A never affects the reliability of a person in the perfor-
mance of F and therefore does not affect the reliability of the particular
person P in the actual context B
A
.
(b) The attribute A sometimes affects the reliability of a person in the per-
formance of F, but does not affect the reliability of the particular person
P in the actual context B
A
.
(c) The attribute A sometimes affects the reliability of a person in the per-
formance of F, and the particular person P in the actual context B
A
rep-
resents a case where A has such an effect.
(d) The attribute A always affects the reliability of a person in the perfor-
mance of F and therefore affects the reliability of the particular person
P in the actual context B
A
.
It can be questioned whether there really are attributes that never af-
fect the reliability of P in performing F. If, however, A never is relevant, it
will be irrelevant irrespective of other attributes that P has. For example,
if the color of the clothes of the person testifying is never relevant to the
assessment of that person’s reliability as a source of information, it does
not matter what other attributes the person has. In any situation of the sort
described in (a) an ad hominem argument to the effect that A is relevant
will commit the fallacy of irrelevant attribute.
The fact that a fallacy is being committed is easy to spot in cases where
it is unthinkable, or at least highly implausible, that the attribute that is
pointed out could ever affect a certain kind of reliability: an example would
be the attribute of an expert on foreign policy looking like an aardvark. It is
Fallacies in Ad Hominem Arguments / C. DaHlman, D. reiDHav and l. WaHlberG
116
COGENCY Vol. 3, N0. 2, Summer 2011
moredifculttospotwhentheargumentpointstoanattributethatisrel-
evant in some contexts but not in the one at hand (situation (b) above). For
example, not wearing one’s glasses normally affects a person’s reliability
as an eyewitness. However, the fact that P was not wearing her glasses is
an irrelevant attribute for her reliability as an eye witness if she has perfect
vision and wears glasses as a fashion accessory.
The fallacy of overrated effect is committed in cases where A has an
effect on P’s reliability in performing F but this effect is less than the effect
premise claims. The argument does not commit the fallacy of irrelevant at-
tribute, as it points to an attribute that is relevant to the reliability issue at
hand. What is wrong in the argument is simply that it exaggerates the effect
oftheattribute.ThiserrorhasalsobeenidentiedbyBrintonasoneofthe
ways that ad hominem arguments can go wrong. Brinton describes it as a
“failure in degree of support” (1995, p. 218).
An example of the fallacy of overrated effect in legal argumentation
occurs in the Swedish case NJA 1986 s. 358. A man, TS, reported to his
insurance company that his car had been stolen and claimed compensa-
tion for his loss in accordance with the insurance policy. The company did
not believe his story and refused to pay. TS sued. The company argued,
before the court, that TS should be considered unreliable because he had
a criminal record: TS had previously been convicted for several crimes, in-
cluding attempted insurance fraud. This argument is clearly ad hominem.
The company claimed that the conviction for insurance fraud had a strong
negative effect on TS’s reliability to give accurate testimony. This negative
ad hominemwasnotacceptedbytheappellatecourt.ConrmingthatTS’s
criminal record did indeed have a negative effect on his reliability, but em-
phasizing that many years had passed since the last conviction, the court
heldthat theeffect waslesssignicantthantheinsurancecompany had
argued. The appellate judges were, in effect, insisting that the ad hominem
argument presented by the insurance company committed the fallacy of
overrated effect.
The fallacy of overrated effect can occur in situations (c) and (d) above.
This fallacy—like the fallacy of irrelevant attribute—can be the result of the
reliability being assessed relative to a non-actual bundle of attributes or an
inaccurate estimate of the effect of the bundle by the arguer. Thus, in the
example from the Swedish Supreme Court the insurance company based
117
theirreliabilityassessmentonanincompletelyspecied(andinthissense
non-actual) bundle in which the time-span between the conviction and the
compensation claim was not included.
Some ad hominem arguments claim that the attribute they invoke cat-
egoricallydisqualiesPfromreliableperformanceofF.Consider,forex-
ample, the following argument.
“There is just one thing you need to know about the next witness. She is
the mother of the accused. As we all know, a mother would say anything
to protect her son. You should therefore disregard everything she has
to say. No other thing that you learn about her could ever change that.”
This kind of argumentation is sometimes referred to as poisoning the well
(see, e.g., Walton, 2006, p. 289). Where it is claimed that A renders P com-
pletely unreliable, no matter what other attributes P has, the situation is of
kind (d). Plainly, any claim imputing complete unreliability will be a very
strong one, and even when the situation indeed is of kind (d) this kind of
argumentation often involves the fallacy of overrated effect.
An attribute’s effect on reliability depends, not just on the bundle, but
also on the function at issue. The fallacies of irrelevant attribute and over-
rated effect can therefore occur if information about an attribute’s effect on
reliability in the performance of one function is unwarrantedly generalized
or extrapolated to another function. This mistake is not uncommon in as-
sessments of the credibility of expert witnesses.
The fallacy of reliability irrelevance is committed when the effect that
A has on P’s reliability does not make the target claim false, because the
target claim is not about reliability. Consider the following example.
X: “The government should cut taxes. It would be good for the economy.”
Y:“You’rejustsayingthatbecauseyouwouldbenetfromtaxcuts.”
The argument advanced by X has an unstated major premise to the ef-
fect that the government should cut taxes if tax cuts would be good for the
economy,aminorpremiseafrming thattaxcutswouldbegoodforthe
economy, and the conclusion that the government should cut taxes. None
of these three claims are about X’s reliability. Y may well be right when he
saysthatXismakingtheargumentbecausetaxcutswouldbenethimper-
Fallacies in Ad Hominem Arguments / C. DaHlman, D. reiDHav and l. WaHlberG
118
COGENCY Vol. 3, N0. 2, Summer 2011
sonally, but that does not falsify any claim in X’s argument, and therefore
does not affect its soundness.
The fallacy of reliability irrelevance is a special case of the fallacy known
as ignoratio elenchi (cf. Walton, 1998, p. 65). It presents an argument on
an issue as a counter-argument to a claim on a different issue. This fal-
lacy is sometimes used deliberately, with rhetorical intent, to divert atten-
tion from the main argument and shift discussion to a different issue (van
Eemeren and Grootendorst, 1992, pp. 152-153). The latter is sometimes
referred to as a red herring.
It should be noted that an argument that is not about reliability can be
backed up by an argument about reliability. This means that an ad homi-
nem argument that commits the fallacy of reliability irrelevance in relation
to a certain argument could be a legitimate counter-argument against a
backing argument. Consider the following dialogue.
X: “The government should cut taxes. It would be good for the econo-
my.”
Y: “I’m not sure. I don’t understand that much about economics.”
X: “You can trust me. I have a PhD in economics. That should make me
reliable as an expert on the economic impact of tax cuts.”
Y:“Sorry.Youwouldalsobenetfromtaxcuts,andthereforeyou’renot
reliable as an expert on the economic impact of tax cuts.”
X is now backing up the main argument with an authority argument.
He says he is a reliable expert on the economic impact of tax cuts because
he has a PhD in economics. This is a positive ad hominem argument. The
argument advanced by Y is also an ad hominem argument—a negative one
that stands as a counter-argument to the positive argument advanced by X.
This second argument targets the conclusion in X’s ad hominem argument
and says that this conclusion makes a false claim. In this way, the counter-
argument adds another attribute, X’s personal interest in tax cuts, to the
bundleandclaimsthatX’scredibility,givenhisdoctoralqualicationand
this additional attribute, differs considerably from his credibility given the
doctoralqualicationalone.Sincethenegativead hominem argument in-
voked by Y draws the conclusion that X’s claim about his own reliability is
false, Y’s argument is an ad hominem argument in the extended form. This
argument does not commit the fallacy of reliability irrelevance.
119
Here lies the crucial difference between Y’s response that X would bene-
tpersonallyfromtaxcutswhenweregarditasacounter-argumenttoX’s
backing argument and the same response when we regard it as a counter-
argument to X’s main argument. As we saw above, the response commits
the fallacy of reliability irrelevance against the main argument. This shows
the importance of picking the right target claim. An ad hominem argument
committing the fallacy of reliability irrelevance vis-à-vis one claim can be
perfectly sound as a counter-argument to another claim.
Noticethatthismakesitdifculttoassessad hominem arguments in
many of their applications to actual events and issues. In real life it is often
unclear what the target claim is. If the rebuttal premise is such that it would
come out false if the argument is taken to target one claim and true if it
is taken to target another, the soundness of the argument may be equally
unclear. This only goes to show that if we are to determine whether an ar-
gument is successful, it must be clear what the argument is meant to target.
Some scholars have suggested that ad hominem arguments do not com-
mit a fallacy of irrelevance when there is doubt about the truth of a premise
invoked by the opposing arguer. Lawrence Hinman (1982, p. 339) present-
ed this idea in the early 1980s, and a similar argument has recently been
made by Christopher Johnson (2009, pp. 256-259). When there is doubt
about the truth of a premise invoked by the opposing arguer it is not irrele-
vant, their argument goes, to question whether the opposing arguer should
be relied on in his claim that the premise is true. There is some truth in this
observation, but it overlooks the fact that a counter-argument will always
haveaspecictarget,anditfailstoseparatecounter-argumentsthattarget
a claim in the main argument from counter-arguments to a claim in the
backing argument. That an ad hominemargumentisrelevantasafalsier
against a claim in a backing argument does not mean that it is relevant as a
falsieragainstaclaiminthemainargument.Thead hominem argument
will still commit the fallacy of reliability irrelevance against the main argu-
ment if the main argument is not about reliability.
One type of argument that commits the fallacy of reliability irrelevance is the
argument known as tu quoque. The standard treatment of ad hominem argu-
mentsidentiestu quoque as a special kind of ad hominem fallacy. Tu quoque is
a counter-argument to the argument that a certain way of acting is wrong which
attacks the opposing arguer by pointing out that he has acted in that very way.
Fallacies in Ad Hominem Arguments / C. DaHlman, D. reiDHav and l. WaHlberG
120
COGENCY Vol. 3, N0. 2, Summer 2011
An interesting example can be taken from the Nuremberg Trials (see,
Yee, 2004, pp. 103-116; Heise, 2009, pp. 10-14). The German Admiral Karl
Dönitz was accused of unrestricted submarine warfare in violation of the
London Naval Treaty of 1936, to which Germany was a party. Dönitz had
ordered his submarines to attack British merchant vessels without warn-
ing. Before the court, Dönitz made the counter-argument that allied sub-
marines had done the same thing. For example, the American Admiral
Chester Nimitz had torpedoed Japanese merchant vessels without warn-
ing. This is an argument ad hominem, although in this case the opposing
arguer being responded to is not a person but rather a group of people that
he represents. The argument challenges the reliability of the allies to pass
judgment on the unrestricted submarine warfare ordered by Dönitz.
The tribunal found Dönitz guilty on the charge of illegal submarine war-
fare. This can be taken to show that the tribunal recognized the tu quoque
argument as fallacious. It should be mentioned, however, that no sentence
was imposed on Dönitz for this crime, in addition to his sentence for other
crimes. This strange decision suggests that the tribunal was, to some de-
gree, taken in by the argument.
The fallacy of irrelevant person is committed when the target argu-
ment is about someone’s reliability, but that person is not P. An argument
that commits the fallacy of irrelevant person presents an argument against
P’s reliability as a counter-argument to a claim about P*s reliability. Here
is an example.
X: “Miss Limpany is highly reliable as an expert witness. She has a PhD
inreengineering.”
Y: “With all due respect, let me just remind the court that every argu-
ment that X makes is biased. As we all know, he is paid by the plain-
tiff to argue this case.”
Since Y’s argument is about the reliability of the opposing advocate (X),
it does not falsify X’s claim about Miss Limpany’s reliability. Waller (2005,
p. 188) has called attention to this kind of fallacy in legal argumentation.
As is probably already apparent, the fallacy of irrelevant person is similar
to the fallacy of reliability irrelevance. It is also a special case of ignoratio
elenchi.
121
The fallacy of insufcient degree is committed when the target argu-
ment is about P’s reliability, but the effect that A has on P’s degree of reli-
abilityisinsufcienttofalsifythetargetclaim.Thefallacycanbeillustrated
by the following counter-argument.
X:“Idon’tthinkthatJaneJetsonislessqualiedthantheaveragepoli-
tician to be Prime Minister.”
Y: “Well, she never took a PhD...”
The fact that a politician does not have a PhD may have a small nega-
tive effect on the degree to which she can be relied upon to serve as Prime
Minister,buttheeffectofnothavingaPhDdoesnotmakeherlessqualied
than the average politician. After all, the average politician does not have
aPhD.Thefallacyofinsufcientdegreeisalsoaspecialcaseofignoratio
elenchi. It involves presenting an argument that is unable to falsify the tar-
get claim.
Lastly, an error occurs if the effect premise in the extended ad hominem
argument concerns a function other than the one in the target claim. The
following exchange illustrates this fallacy of irrelevant function:
X: “I would rather trust Bill Clinton than Barack Obama to handle the
economic crisis.”
Y: “No way! I wouldn’t trust Bill Clinton to date my daughter.”
4. Talking about Errors as Fallacies
As we have seen, ad hominem arguments can go wrong in seven different
ways. We have referred to these errors as seven different ad hominem falla-
cies. Some scholars object to our analysis and claim that some of the errors
thatwehaveidentiedshouldnotbecalled“fallacies”.Theyacceptthatwe
haveidentiedsevendifferentwaysinwhichad hominem arguments can
go wrong, but deny that all of them qualify as “fallacies”.
There are different opinions among scholars on how the term fallacy
should be used. Some scholars argue, for instance, that an error must be
deceptive to qualify as a fallacy (see, e.g., Hitchcock, 2006; Woods, 2007,
Fallacies in Ad Hominem Arguments / C. DaHlman, D. reiDHav and l. WaHlberG
122
COGENCY Vol. 3, N0. 2, Summer 2011
p. 110). We prefer not to employ this psychological criterion. Perhaps some
oftheerrorsthatwehaveidentiedarenotverydeceptive,butthismakes
no difference in our analysis. When we talk about an argument as fallacious
we make no claim about its psychological effect. It has also been insisted
that a fallacious error must occur with some frequency in real arguments
(see, e.g., Hitchcock 2006; Woods 2007, p. 110). Some of the errors that we
haveidentiedareperhapsnotcommonenoughtoqualifyasfallacieson
this view. However, we do not subscribe to this kind of empirical criterion
either.
Brintonhasidentiedsomewaysthatad hominem arguments can go
wrong, but he is only prepared to talk about one of them as a fallacy. For
Brinton(1995,pp.216-217),onlythekindoferrorthatwehaveclassied
as the fallacy of reliability irrelevance comes close to qualifying as a fallacy.
Whenheidentiesthekindoferrorthatwehavecalledthefallacyoffalse
attribution he says that this error should not be called a fallacy, as it is not a
“logical failure”, but merely consists in a false premise (1995, pp. 217-218).
ThisreectsafundamentaldifferencebetweenBrintonandus.Inourap-
proach, where the ad hominem argument is given a deductively valid form,
all errors in this argument will manifest themselves as false premises.
5. Conclusions
In this paper we have elaborated a framework which can be summarized
as follows. Arguments ad hominem note that a person P has a certain at-
tribute and claim that this attribute affects P’s reliability in performing a
certain function. P can be another participant in the discussion, but this is
not necessarily the case.
Arguments ad hominem are often used as counter-arguments, but they
need not be counter-arguments. In its basic form, an ad hominem argu-
ment merely makes a claim about someone’s reliability. It need not stand
in opposition to some other claim. When an ad hominem argument says
that some other claim about the person’s reliability is false it becomes an
ad hominem argument in the extended form.
By including all kinds of function that a person can be relied upon to
fulll,thisframeworkcanbeusedtoanalyzeandexplainawiderangeof
123
arguments about reliability. It is not limited to the subcategory of argu-
ments about credibility.
As we have seen, our understanding of ad hominem arguments cov-
ers positive as well as negative variants. An argument where the attribute
pointed out is said to increase reliability is a positive ad hominem, and an
argument where the attribute is said to decrease reliability is a negative ad
hominem.
Once we give the ad hominem argument a deductively valid form, all
errors manifest themselves as false premises. Every possible error either
involves a false attribution premise, or a false effect premise, or (in the
extended form of the argument) a false rebuttal premise. As we have seen,
these premises can be false for various, rather different reasons. We have
identiedsevengroundsoffalsityhere,andwehavesuggestedthatthese
grounds create, in effect, seven distinct ad hominem fallacies: false attribu-
tion, irrelevant attribute, overrated effect, reliability irrelevance, irrelevant
person,insufcientdegreeandirrelevantfunction.
The presented framework allows questions about when, and to what ex-
tent, different kinds of error occur to be addressed rigorously. It also helps
to explain why such errors occur. Thus we have seen that the falsity of the
effect premise, whether it consists in the fallacy of irrelevant attribute or
the fallacy of overrated effect, can be due to a misjudgment of some kind
about the function. And when we discussed the relevance of the assumed
bundle of attributes we saw that this error can result either from a mistaken
assumption about the actual bundle or from a faulty estimate of the effect
of an attribute in that bundle.
The framework that has been presented in this article provides a widely
applicable tool for approaching and analyzing ad hominem arguments. We
recommend that judges use it in their assessment of legal argumentation.
As we have seen, ad hominem arguments are often used to attack witness
testimony, and we believe that our framework would help judges to identify
the various fallacies that are committed in such arguments.
We also hope, of course, that our framework will be used for assessing
political argumentation and everyday conversation. We believe, in general,
that our framework makes it easier to spot ad hominem fallacies, and eas-
ier to avoid them.
Fallacies in Ad Hominem Arguments / C. DaHlman, D. reiDHav and l. WaHlberG
124
COGENCY Vol. 3, N0. 2, Summer 2011
Works Cited
Bentham, J. The Handbook of Political Fallacies. New York: Harper & Broth-
ers, 1824/1952.
Brinton, A. “A Rhetorical View of the Ad Hominem.” Australasian Journal of
Philosophy 63 (1) (1985): 50-63.
Brinton, A. “The Ad Hominem.” In Hansen, H. and Pinto, R. (eds.), Fallacies.
Classical and Contemporary Readings (pp. 213-222). University Park:
Pennsylvania State University Press, 1995.
Copi, I. and Cohen, C. Introduction to Logic. 11
th
ed. Upper Saddle River: Pren-
tice Hall, 2002.
van Eemeren, F. and Grootendorst, R. “Relevance Reviewed: The Case of Argu-
mentum ad Hominem.” Argumentation 6 (2) (1992): 141-159.
Hamblin, Ch. Fallacies. London: Methuen, 1970.
Heise, N. “Deciding not to Decide: Nuremberg and the Ambiguous History of
the Tu Quoque Defense.” Social Science Research Network. Social Science
Electronic Publishing, Inc., 2009. Web. 5 July 2011. <http://ssrn.com/ab-
stract=1354048>
Hinman, L. “The Case for Ad Hominem Arguments.” Australasian Journal of
Philosophy 60 (4) (1982): 338-345.
Hitchcock, D. “Why there is no Argumentum ad Hominem Fallacy.” McMaster
University, 2006. Web. 1 April 2011. <http://www.humanities.mcmaster.
ca/~hitchckd/ adhominemissa.htm>
Johnson, Ch. “Reconsidering the Ad Hominem.” Philosophy 84 (2) (2009):
251-266.
Waller, B. Critical Thinking. Consider the Verdict. 5
th
ed. Upper Saddle River:
Prentice Hall, 2005.
Walton, D. “The Ad Hominem Argument as an Informal Fallacy.” Argumenta-
tion 1 (3) (1987): 317-331.
Walton, D. Ad Hominem Arguments. Tuscaloosa: University of Alabama Press,
1998.
Walton, D. “Poisoning the Well.” Argumentation 20 (3) (2006): 273-307.
Woods, J. “Lightening up on the Ad Hominem.” Informal Logic 27 (1) (2007):
109-134.
Woods, J. The Death of Argument. Dordrecht: Kluwer Academic Publishers,
2010.
Yee, S. “The Tu Quoque Argument as a Defence to International Crimes, Prose-
cution or Punishment.” Chinese Journal of International Law 3 (1) (2004):
87-134.