Using Reasoning to Persuade: Making Sense of “Irrationality” + How Reasoning Helps Groups But Also Drives Bad Behavior


This is long and bit technical excerpt from a great paper — Why do humans reason? Arguments for an argumentative theory — Hugo Merciera and Dan Sperbera — but gives very useful to hints on helping to understand how people think about financial matters and behaviors.  It also offers a theoretical foundation for persuasion and changing behavior.

The Problem:

“The direction reasoning takes is mostly determined by the participants’ initial intuitions. If they have arrived at the conclusion themselves, or if they agree with it, they try to confirm it. If they disagree with it, they try to prove it wrong.  In all cases, what they do is try to confirm their initial intuition.”


“Human reasoning is not a profoundly flawed general mechanism; it is a remarkably efficient specialized device adapted to a certain type of social and cognitive interaction at which it excels.”
“…contrary to common bleak assessments of human reasoning abilities, people are quite capable of reasoning in an unbiased manner, at least when they are evaluating arguments rather than producing them, and when they are after the truth rather than trying to win a debate.”

Key Points:

  • Reasoning can lead to poor outcomes not because humans are bad at it but because they systematically look for arguments to justify their beliefs or their actions.
  • People can be skilled arguers, producing and evaluating arguments felicitously.  This good performance stands in sharp contrast with the abysmal results  found in other, no argumentative, settings
  • We outline an approach to reasoning based on the idea that the primary function for which it evolved is the production and evaluation of arguments in communication. It is to devise and evaluate arguments intended to persuade.
  • Reasoning enables people to exchange arguments that, on the whole, make communication more reliable and hence more advantageous.
  • Reasoning contributes to the effectiveness and reliability of communication by enabling communicators to argue for their claim and by enabling addressees to assess these arguments.  It thus increases both in quantity and in knowledge quality the information humans are able to share.
  • Reasoning is often used to find justifications for performing actions that are otherwise felt to be unfair or immoral. Such uses of reasoning can have dire consequences.

Abstract: Reasoning is generally seen as a means to improve knowledge and make better decisions. However, much evidence shows that reasoning often leads to knowledge distortions and poor decisions. This suggests that the function of reasoning should be rethought. 

Our hypothesis is that the function of reasoning is argumentative.

  • It is to devise and evaluate arguments intended to persuade.  Reasoning so conceived is adaptive given the exceptional dependence of humans on communication and their vulnerability to misinformation.
  • Skilled arguers, however, are not after the truth but after arguments supporting their views.  This explains the notorious confirmation bias. This bias is apparent not only when people are actually arguing but also when they are reasoning proactively from the perspective of having to defend their opinions.  Reasoning so motivated can distort evaluations and attitudes and allow erroneous beliefs to persist.
  • Proactively used reasoning also favors decisions that are easy to justify but not necessarily better. In all these instances traditionally described as failures or flaws, reasoning does exactly what can be expected of an argumentative device: Look for arguments that support a given conclusion, and favor conclusions for which arguments can be found.


  • INFERENCE is the production of new mental representations on the basis of previously held representations.  Examples of inferences are the production of new beliefs on the basis of previous beliefs, the production of expectations on the basis of perception, or the production of plans on the basis of preferences and beliefs.  So understood, inference need not be deliberate or conscious. It is at work not only in conceptual thinking but also in perception and in motor control (It is a basic ingredient of any cognitive system).
  • REASONING, refers  to a very special form of inference at the conceptual level, where not only is a new mental representation (or conclusion) consciously produced, but the previously held representations (or premises) that warrant it are also consciously entertained. The premises are seen as providing reasons to accept the conclusion.  *Such reasoning is typically human. There is no evidence that it occurs in nonhuman animals or in preverbal children. *

Conclusion: Reasoning and Rationality
Reasoning contributes to the effectiveness and reliability of communication by enabling communicators to argue for their claim and by enabling addressees to assess these arguments.  It thus increases both in quantity and in knowledge quality the information humans are able to share.

  • We view the evolution of reasoning as linked to that of human communication.
  • Reasoning, we have argued, enables communicators to produce arguments to convince addressees who would not accept what they say on trust; it enables addressees to evaluate the soundness of these arguments and to accept valuable information that they would be suspicious of otherwise.
  • Thus, thanks to reasoning, human communication is made more reliable and more potent.

From the hypothesis that the main function of reasoning is argumentative, we derived a number of predictions that, we tried to show, are confirmed by existing evidence. True, most of these predictions can be derived from other theories. We would argue, however, that the argumentative hypothesis provides a more principled explanation of the empirical evidence (in the case of the confirmation bias, for instance).

In our discussion of motivated reasoning and of reason-based choice, not only did we converge in our prediction with existing theories, we also extensively borrowed from them. Even in these cases, however, we would argue that our approach has the distinctive advantage of providing clear answers to the why-questions:

  • Why do humans have a confirmation bias?
  • Why do they engage in motivated reasoning?
  • Why do they base their decisions on the availability of justificatory reasons?

Moreover, the argumentative theory of reasoning offers a unique integrative perspective: It explains wide swaths of the psychological literature within a single overarching framework.

Some of the evidence reviewed here shows not only that reasoning falls short of delivering rational beliefs and rational decisions reliably, but also that, in a variety of cases, it may even be detrimental to rationality.  Reasoning can lead to poor outcomes not because humans are bad at it but because they systematically look for arguments to justify their beliefs or their actions.

The argumentative theory, however, puts such well-known demonstrations of “irrationality” in a novel perspective.  Human reasoning is not a profoundly flawed general mechanism; it is a remarkably efficient specialized device adapted to a certain type of social and cognitive interaction at which it excels.

Even from a strictly knowledge perspective the argumentative theory of reasoning does not paint a wholly disheartening picture. It maintains that there is an asymmetry between the production of arguments, which involves an intrinsic bias in favor of the opinions or decisions of the arguer whether or not they are sound, and the evaluation of arguments, which aims at distinguishing good arguments from bad ones and hence genuine information from misinformation. This asymmetry is often obscured in a debate situation (or in a situation where a debate is anticipated).

  • People who have an opinion to defend don’t really evaluate the arguments of their interlocutors in a search for genuine information but rather consider them from the start as counterarguments to be rebutted.
  • Still, as shown by the evidence reviewed in section, people are good at assessing arguments and are quite able to do so in an unbiased way, provided they have no particular axe to grind.

In group reasoning experiments where participants share an interest in discovering the right answer, it has been shown that truth wins.

  • While participants in collective experimental tasks typically produce arguments in favor of a variety of hypotheses, most or even all of which are false, they concur in recognizing sound arguments.
  • Since these tasks have a demonstrably valid solution, truth does indeed win.
  • If we generalize to problems that do not have a provable solution, we should at least expect good arguments to win, even if this is not always sufficient for truth to win.

This may sound trivial, but it is not. It demonstrates that, contrary to common bleak assessments of human reasoning abilities, people are quite capable of reasoning in an unbiased manner, at least when they are evaluating arguments rather than producing them, and when they are after the truth rather than trying to win a debate.

Couldn’t the same type of situation that favors sound evaluation favor comparable soundness in the production of arguments?

Tips for Group Decision-Making — 

  • Note, first, that situations where a shared interest in truth leads participants in a group task to evaluate arguments correctly are not enough to make them produce correct arguments.
  • In these group tasks, individual participants come up with and propose to the group the same inappropriate answers that they come up with in individual testing.
  • The group success is due to, first and foremost, the filtering of a variety of solutions, achieved through evaluation.
  • When different answers are initially proposed and all of them are incorrect, then all of them are likely to be rejected, and wholly or partly new hypotheses are likely to be proposed and filtered in turn, thus explaining how groups may do better than any of their individual members.

Individuals thinking on their own without benefiting from the input of others can assess only their own hypotheses, but in doing so, they are both judge and party, or rather judge and advocate, and this is not an optimal stance for pursuing the truth.

Wouldn’t it be possible, in principle, for an individual to decide to generate a variety of hypotheses in answer to some question and then evaluate them one by one?  More realistically, individuals may develop some limited ability to distance themselves from their own opinion, to consider alternatives and thereby become more objective. Presumably this is what the 10% or so of people do. But this is an acquired skill and involves exercising some imperfect control over a natural disposition that spontaneously pulls in a different direction.

Here, one might be tempted to point out that; after all, reasoning is responsible for some of the greatest achievements of human thought in the knowledge and moral domains. This is undeniably true, but the achievements involved are all collective and result from interactions over many generations the whole scientific enterprise has always been structured around groups, from the Lincean Academy down to the Large Hadron Collider.

In the moral domain, moral achievements such as the abolition of slavery are the outcome of intense public arguments. We have pointed out that, in group settings, reasoning biases can become a positive force and contribute to a kind of division of cognitive labor.

Still, to excel in such groups,:

  • it may be necessary to anticipate how one’s own arguments might be evaluated by others and to adjust these arguments accordingly.
  • Showing one’s ability to anticipate objections may be a valuable culturally acquired skill, as in medieval disputations.
  • By anticipating objections, one may even be able to recognize flaws in one’s own hypotheses and go on to revise them.
  • We have suggested that this depends on a painstakingly acquired ability to exert some limited control over one’s own biases. Even among scientists, this ability may be uncommon, but those who have it may have a great influence on the development of scientific ideas.
  • It would be a mistake, however, to treat their highly visible, almost freakish, contributions as paradigmatic examples of human reasoning. In most discussions, rather than looking for flaws in our own arguments, it is easier to let the other person find them and only then adjust our arguments, if necessary.

In general, one should be cautious about using the striking accomplishments of reasoning as proof of its overall efficiency, since its failures are often much less visible

  • Knowledge success may depend to a significant extent on what philosophers have dubbed knowledge luck; that is, chance factors that happen to put one on the right track.
  • When one happens to be on the right track and “more right” than one could initially have guessed, some of the distorting effects of motivated reasoning and polarization may turn into blessings.

To conclude, we note that the argumentative theory of reasoning should be  congenial to those of us who enjoy spending endless hours debating ideas – but this,  of course, is not an argument for (or against) the theory.

Reasoning: Mechanism and Function  – Intuitive Inference and Argument —
Since the 1960s, much work in the psychology of reasoning has suggested that, in fact, humans reason rather poorly,

  • Failing at simple logical tasks
  • Committing egregious mistakes in probabilistic reasoning, and
  • Being subject to sundry irrational biases in decision making

This work has led to a rethinking of the mechanisms for reasoning, but not – or at least not to the same degree – of its assumed function of enhancing human cognition and decision making.

We contend in particular that:

  • The arguments used in reasoning are the output of a mechanism of intuitive inference
  • A process of inference is a process the representational output of which necessarily or probabilistically follows from its representational input.
  • The function of an inferential process is to augment and correct the information available to cognitive system.

An evolutionary approach suggests that:

  • Inferential processes, rather than being based on a single inferential mechanism or constituting a single integrated system
  • Are much more likely to be performed by a variety of domain-specific  mechanisms, each attuned to the specific demands and affordances of its domain
  • The inferential processes carried out by these  mechanisms are unconscious
  • They are not mental acts that individuals decide to  perform, but processes that take place inside their brain, at a “sub-personal” level
  • People may be aware of having reached a certain conclusion – be aware, that is, of the output of an inferential process – but we claim that they are never aware of the process itself.
  • All inferences carried out by inferential mechanisms are in this sense intuitive. They generate intuitive beliefs; that is, beliefs held without awareness of reasons to hold them.

The claim that all inferential processes carried out by specialized inferential  mechanisms are unconscious and result in intuitive inferences may seem to contradict  the common experience of forming a belief because one has reflected on reasons to  accept it, and not, or not only, because of its intuitive force. What characterizes reasoning proper is indeed the awareness not just of a conclusion but of an argument that justifies accepting that conclusion.

We suggest, however, that:

  • Arguments exploited in reasoning are the output of an intuitive inferential mechanism.
  • Like all other inferential mechanisms, its processes are unconscious and its conclusions are intuitive.
  • However, these intuitive conclusions are about arguments; that is, about representations of relationships between premises and conclusions.

Arguments should be sharply distinguished from inferences.

  • An inference is a process the output of which is a representation.
  • An argument is a complex representation.
  • Both an inference and an argument have what can be called a conclusion, but in the case of an inference, the conclusion is the output of the inference; in the case of an argument, the conclusion is a part – typically the last part – of the representation.
  • The output of an inference can be called a “conclusion” because what characterizes an inferential process is that its output is justified by its input; the way however in which the input justifies the output is not represented in the output of an intuitive inference.
  • What makes the conclusion of an argument a “conclusion” (rather than simply a proposition) is that the reasons for drawing this conclusion on the basis of the premises are (at least partially) spelled out.

…it is a common but costly mistake to confuse the causally and temporally related steps of an inference with the logically related steps of an argument. The causal steps of an inference need not recapitulate the logical step of any argument for it to be an inference, and the logical step of an argument need not be followed in any inference for it to be an argument.

All arguments must ultimately be grounded in intuitive judgments that given conclusions follow from given premises.

  • Rather, arguments are the output of one mechanism of intuitive inference among many that delivers intuitions about premise-conclusion relationships.
  • Intuitions about arguments have an evaluative component: Some arguments are seen as strong, others as weak.
  • Moreover there may be competing arguments for opposite conclusions and we may intuitively prefer one to another. These evaluation and preferences are ultimately grounded in intuition.

If we accept a conclusion because of an argument in its favor that is intuitively strong enough, this acceptance is an knowledge decision that we take at a personal level.  If we construct a complex argument by linking argumentative steps, each of which we see as having sufficient intuitive strength, this is a personal-level mental action.  If we verbally produce the argument so that others will see its intuitive force and will accept its conclusion, it is a public action that we consciously undertake.

  • The mental action of working out a convincing argument,
  • The public action of verbally producing this argument so that others will be convinced by it,
  • The mental action of evaluating and accepting the conclusion of an argument produced by others correspond to what is commonly and traditionally meant by reasoning (a term that can refer to either a mental or a verbal activity).

We see three complementary explanations for the saliency of reasoning.

  • First, when we reason, we know that we are reasoning, whereas the very existence of intuitive inference was seen as controversial in philosophy before its discovery in cognitive science.
  • Second, while an inferential mechanism that delivers intuitions about arguments is, strictly speaking, highly domain specific, the arguments that it delivers intuitions about can be representations of anything at all. Thus, when we reason on the basis of these intuitions, we may come to conclusions in all theoretical and practical domains. In other words, even though inferences about arguments are domain specific (as evolutionary psychologists would expect), they have domain general consequences and provide a kind of virtual domain generality (without which traditional and dual-process approaches to reasoning would make little sense).
  • Third, as we will now argue, the very function of reasoning puts it on display in human communication.

The Function of Reasoning —
We use function here in its biological sense (Put simply, a function of a trait is an effect of that trait that causally explains its having evolved and persisted in a population: Thanks to this effect, the trait has been contributing to the fitness of organisms endowed with it.

In principle, several effects of a trait may contribute to fitness, and hence a trait may have more than a single function.  Even then, it may be possible to rank the importance of different functions, and in particular to identify a function for which the trait is best adapted as its main function.

…we are not arguing against the view that our reasoning ability may have various advantageous effects, each of which may have contributed to its selection as an important capacity of the human mind. We do argue, however, that reasoning is best adapted for its role in argumentation, which should therefore be seen as its main function.

There have been a few tentative attempts in dual-process approaches to explain the function and evolution of reasoning.

  • The majority view seems to be that the main function of reasoning is to enhance individual cognition.  However, reasoning itself is a potential source of new mistakes.  Moreover, there is considerable evidence that, when reasoning is applied to the conclusions of intuitive inference, it tends to rationalize them rather than to correct them
  • According to another hypothesis, conscious reasoning “gives us the possibility to deal with novelty and to anticipate the future.” But giving an organism the possibility to deal with novelty and to anticipate the future is less a characterization of reasoning than it is of learning (or even, it could be argued,  of cognition in general).  After all, learning can be defined as “the process by which we become able to use past and current events to predict what the future holds.”  The issue is not whether, on occasion, reasoning can help correct intuitive mistakes or better adapt us to novel circumstances. No doubt, it can. The issue is how far these occasional benefits explain the costs incurred, and hence the very existence of reasoning among humans, and also explain its characteristic features.

In any case, evolutionary hypotheses are of little help unless precise enough to yield testable predictions and explanations. To establish that reasoning has a given function, we should be able at least to identify signature effects of that function in the very way reasoning works. Here we want to explore the idea that the emergence of reasoning is best understood within the framework of the evolution of human communication.

  • Reasoning enables people to exchange arguments that, on the whole, make communication more reliable and hence more advantageous.
  • The main function of reasoning, we claim, is argumentative (or communication to be stable, it has to benefit both senders and receivers otherwise they would stop sending or stop receiving, putting an end to communication itself.

What are the options of a communicator wanting to communicate a piece of information that the addressee is unlikely to accept on trust?

  • One option may be for the communicator to provide evidence of her reliability in the matter at hand (for instance, if the information is about health issues, she might inform the addressee that she is a doctor). But what if the communicator is not in a position to boost her own authority?
  • Another option is to try to convince her addressee by offering premises the addressee already believes or is willing to accept on trust, and showing that, once these premises are accepted, it would be less coherent to reject the conclusion than to accept it. This option consists in producing arguments for one’s claims and in encouraging the addressee to examine, evaluate, and accept these arguments.  Producing and evaluating arguments is, of course, a use of reasoning.

Reasoning contributes to the effectiveness and reliability of communication by allowing communicators to argue for their claim and by allowing addressees to assess these arguments. It thus increases both in quantity and in knowledge quality the information humans are able to share.

Communication plays an obvious role in human cooperation both in the setting of common goals and in the allocation of duties and rights.  Argumentation is uniquely effective in overcoming disagreements that are likely to occur, in particular in relatively equalitarian groups.

The main function of reasoning is argumentative: reasoning has evolved and persisted mainly because it makes human communication more effective and advantageous.

Argumentative Skills: Understanding and Evaluating Arguments —
The demonstration that people are skilled at assessing arguments seems to stand in sharp contrast with findings from the psychology of reasoning.

Various experiments have shown that participants are generally able to spot other argumentative fallacies.  Not only do they spot them, but they tend to react appropriately: rejecting them when they are indeed fallacious, or being convinced to the degree that they are well grounded.  When researchers have studied other skills specific to argumentation, performance has proved to be satisfactory.

Thus, participants are able to recognize the macrostructure of arguments (to follow the commitments of different speakers, and to attribute the burden of proof appropriately). On the whole, the results reviewed in this section demonstrate that people are good at evaluating arguments both at the level of individual inferences and at the level of whole discussions.

Producing Arguments —
The conclusions of these studies were quite bleak and highlighted three main flaws.

  • The first is that people resort to mere explanations (“make sense” causal theories) instead of relying on genuine evidence (data) to support their views.  (However, later research has shown that this is mostly an artifact of the lack of evidence available to the participants: When evidence is made available, participants will favor it (in both production and evaluation). )
  • A second flaw is the relative superficiality of the arguments used by participants.

–  In a normal argumentative setting, a good argument is an argument that is not refuted. As long as they are not challenged, it makes sense to be satisfied with seemingly superficial arguments.

–  On the other hand, people should be able to generate better arguments when engaged in a real debate.

–  This is exactly what observed: Participants who had to debate on a given topic showed a significant  improvement in the quality of the arguments they used afterwards

  • The third flaw is the most relevant one here.  Participants had generally failed to anticipate counterarguments and generate rebuttals.  Seen from an argumentative perspective, however, this may not be a simple flaw but rather a feature of argumentation that contributes to its effectiveness in fulfilling its function.

If one’s goal is to convince others, one should be looking first and foremost for supportive arguments.  Looking for counterarguments against one’s own claims may be part of a more sophisticated and effortful argumentative strategy geared to anticipating the interlocutor’s response.  If this is a correct explanation of what need not be a flaw after all, then the difficulty that people seem to have in coming up with counterarguments should be easily overcome by having them challenge someone else’s claims rather than defending their own.

Indeed, when mock jurors were asked to reach a verdict and were then presented with an alternative verdict, nearly all of them were able to find counterarguments against it .  In another experiment, all participants were able to find counterarguments against a claim (which was not theirs) and to do so very quickly.

When people have looked at reasoning performance in felicitous argumentative settings, they have observed good results.

To sum up, people can be skilled arguers, producing and evaluating arguments felicitously.  This good performance stands in sharp contrast with the abysmal results  found in other, no argumentative, settings, a contrast made particularly clear by the  comparison between individual and group performance.

Group Reasoning —
Many types of tasks have been studied in group settings, with very mixed results.  The most relevant findings here are those pertaining to logical or, more generally, intellective tasks “for which there exists a demonstrably correct answer within a verbal or mathematical conceptual system”.

  • Intellective tasks allow for a direct comparison with results from the individual reasoning literature, and the results are unambiguous.
  • The dominant scheme is truth wins, meaning that, as soon as one participant has understood the problem, she will be able to convince the whole group that her solution is correct.
  • This can lead to big improvements in performance.

Several challenges can be leveled against this interpretation of the data. It could be suggested that the person who has the correct solution simply points it out to the others, who immediately accept it without argument, perhaps because they have recognized this person as the “smartest”. The transcripts of the experiments show that this is not the case:

  • Most participants are willing to change their mind only once they have been thoroughly convinced that their initial answer was wrong.
  • More generally, many experiments have shown that debates are essential to any improvement of performance in group settings.
  • Moreover, in these contexts, participants decide that someone is smart based on the strength and relevance of her arguments and not the other way around.  Indeed, it would be very hard to tell who is  “smart” in such groups – even if general intelligence were easily perceptible, it only  correlates .33 with

How groups – even when no member had the correct answer initially – can facilitate learning and comprehension of a wide variety of problems:

  • People are skilled arguers:  They can use reasoning both to evaluate and to produce arguments.
  • This good performance offers a striking contrast with the poor results obtained in abstract reasoning tasks.
  • Finally, the improvement in performance observed in argumentative settings confirms that reasoning is at its best in these contexts. We will now explore in more depth a phenomenon already mentioned in this section: the confirmation bias.

The Confirmation Bias: A Flaw of Reasoning or a Feature of Argument Production? —
The confirmation bias consists in the “seeking or interpreting of evidence in ways that are partial to existing beliefs, expectations, or a hypothesis in hand”.  It is one of the most studied biases in psychology.  While there is some individual variation, it seems that everybody is affected to some degree, irrespective of factors like general intelligence or open mindedness.

For standard theories of reasoning, the confirmation bias is no more than a flaw of reasoning.  For the argumentative theory, however, it is a consequence of the function of reasoning and hence a feature of reasoning when used for the production of arguments.

  • In cases that deserve the label of confirmation bias, people are trying to convince others. They are typically looking for arguments and evidence to confirm their own claim, and ignoring negative arguments and evidence unless they anticipate having to rebut them. While this may be seen as a bias from a normative epistemological perspective it clearly serves the goal of convincing others.
  • In another type of case, we are dealing not with biased reasoning but with an absence of reasoning proper. Such an absence of reasoning is to be  expected when people already hold some belief on the basis of perception, memory,  or intuitive inference, and do not have to argue for it.

The theory we are proposing makes three broad predictions.

  • The first is that the genuine confirmation bias (as opposed to straightforward trust in one’s intuitive beliefs and their positive consequences) should occur only in argumentative situations.
  • The second is that it should occur only in the production of arguments. The rationale for a confirmation bias in the production of arguments to support a given claim does not extend to the evaluation of arguments by an audience that is just aiming to be well informed.
  • The third prediction is that the confirmation bias in the production of arguments is not a bias in favor of confirmation in general and against disconfirmation in general: It is a bias in favor of confirming one’s own claims, which should be naturally complemented by a bias in favor of disconfirming opposing claims and counterarguments.

… The direction reasoning takes is mostly determined by the participants’ initial intuitions. If they have arrived at the conclusion themselves, or if they agree with it, they try to confirm it. If they disagree with it, they try to prove it wrong.  In all cases, what they do is try to confirm their initial intuition.

Rehabilitating the Confirmation Bias —

  • Participants have intuitions that lead them towards certain answers.  If reasoning is used at all, it is mostly used to confirm these initial intuitions. This is exactly what one should expect of an argumentative skill, and so these results bolster our claim that the main function of reasoning is argumentative.
  • By contrast, if people were easily able to abstract from this bias, or if they were subject to it only in argumentative settings, then this would constitute evidence against the present theory.

According to a more standard explanation of the confirmation bias, it is an effect of limitations in cognitive resources and in particular in working memory but it is hard to reconcile this explanation with the fact that people are very good at falsifying propositions when they are inclined to disagree with them.  In those cases, people are not held back by limited resources even though the tasks are not cognitively easier.

However, the idea that the confirmation bias is a normal feature of reasoning that plays a role in the production of arguments may seem surprising in light of the poor outcomes it has been claimed to cause.  In such cases, reasoning tends not to be used in its normal context: that is, the resolution of a disagreement through discussion. When one is alone or with people who hold similar views, one’s arguments will not be critically evaluated.

This is when the confirmation bias is most likely to lead to poor outcomes. However, when reasoning is used in a more felicitous context – that is, in arguments among people who disagree but have a common interest in the truth – the confirmation bias contributes to an efficient form of division of cognitive labor.

When a group has to solve a problem:

  • It is much more efficient if each individual looks mostly for arguments supporting a given solution.
  • They can then present these arguments to the group, to be tested by the other members.
  • This method will work as long as people can be swayed by good arguments, and the results reviewed in section show that this is generally the case.
  • This joint dialogic approach is much more efficient than one where each individual on his or her own has to examine all possible solutions carefully.
  • The advantages of the confirmation bias are even more obvious given that each participant in a discussion is often in a better position to look for arguments in favor of his or her favored solution (situations of asymmetrical information).

So group discussions provide a much more efficient way of holding the confirmation bias in check.  By contrast, the teaching of critical thinking skills, which is supposed to help us overcome the bias on a purely individual basis, does not seem to yield very good results.

For the confirmation bias to play an optimal role in discussions and group performance, it should be active only in the production of arguments and not in their evaluation.  Of course, in the back-and-forth of a discussion, the production of one’s own arguments and the evaluation of those of the interlocutor may interfere with each other, making it hard to properly assess the two processes independently.

People tend to be more objective in evaluation than in production.

Proactive Reasoning in Belief Formation —
According to the argumentative theory, reasoning is most naturally used in the context of an exchange of arguments during a discussion. But people can also be proactive and anticipate situations in which they might have to argue to convince others that their claims are true or that their actions are justified. We would say that much reasoning anticipates the need to argue.

Many of our beliefs are likely to remain unchallenged because

  • They are relevant only to ourselves and
  • We don’t share them
  • Or because they are uncontroversial among the people we interact with
  • Or because we have sufficient authority to be trusted when we assert them.

While we think of most of our beliefs – to the extent that we think about them at all – not as beliefs but just as pieces of knowledge, we are also aware that some of them are unlikely to be universally shared, or to be accepted on trust just because we express them.

  • When we pay attention to the contentious nature of these beliefs, we typically think of them as opinions.
  • Opinions are likely to be challenged and may have to be defended. It makes sense to look for arguments for our opinions before we find ourselves called upon to state them.
  • If the search for arguments is successful, we will be ready. If not, then perhaps it might be better to adopt a weaker position, one that is easier to defend. Such uses of reasoning  have been intensively studied under the name of motivated reasoning

Motivated Reasoning —
Other studies have demonstrated the use of motivated reasoning to support various beliefs that others might challenge.

  • Participants dig in and occasionally alter their memories to preserve a positive view of themselves.
  • They modify their causal theories to defend some  favored belief

All these experiments demonstrate that people sometimes look for reasons to justify an opinion they are eager to uphold.

  • From an argumentative perspective, they do this not to convince themselves of the truth of their opinion but to be ready to meet the challenges of others.
  • If they find themselves unprepared to meet such challenges, they may become reluctant to express an opinion they are unable to defend and less favorable to the opinion itself, but this is an indirect individual effect of an effort that is aimed at others.

Consequences of Motivated Reasoning —
…participants used reasoning not so much to assess the studies objectively as to confirm their initial views by finding either flaws or strengths in similar studies, depending on their conclusion.   This phenomenon is known as biased assimilation or biased evaluation.

Several other experiments have studied the way people evaluate arguments depending on whether they agree or disagree with the conclusions.

When people disagree with the conclusion of an argument, they often spend more time evaluating it.

  • This asymmetry arises from the trivial fact that rejecting what we are told generally requires some justification, whereas accepting it does not.
  • Moreover, the time spent on these arguments is mostly devoted to finding  counterarguments
  • Participants tend to comb through arguments for flaws and end up finding some, whether they are problems with the design of a scientific study, issues with a piece of statistical reasoning, or argumentative fallacies.
  • In all these cases, motivated reasoning leads to a biased assessment: Arguments with unflavored conclusions are rated as less sound and less persuasive than arguments with favored conclusions.

Sometimes the evaluation of an argument is biased to the point where it has an opposite effect to the one intended by the arguer:

  • On reading an argument with a counter-attitudinal conclusion (one that goes against their own beliefs or preferences), interlocutors may find so many flaws and counterarguments that their initial unfavorable attitude is in fact strengthened.
  • This is the phenomenon of attitude  polarization, which has been studied extensively since it was first demonstrated have demonstrated that, in the domain of politics, attitude polarization  is most easily observed in participants who are most knowledgeable. Their knowledge makes it possible for these participants to find more counterarguments, leading to more biased evaluations.

Polarization, Bolstering, And Overconfidence —
Attitude polarization can also occur in simpler circumstances.

  • Merely thinking about an object may be enough to strengthen attitudes towards it (polarization).
  • Polarization increases with the time spent thinking about an item, and showed that it increases with the motivation to think.
  • As in the case of polarization following biased evaluation, such polarization occurs only when participants are knowledgeable.  And the effect can be mitigated by providing a reality check: The simple presence of the target object will dramatically decrease polarization.

These results demonstrate that reasoning contributes to attitude polarization and strongly suggest that it may be its main factor.

  • When people are asked to think about a given item towards which they intuitively  have a positive or negative attitude, what happens, we suggest, is that they reflect less  on the item itself than on how to defend their initial attitude.
  • Many other experiments have shown that, once people have formed an attitude to a target, they will look for information that supports this attitude (a phenomenon known as selective exposure and try to put any information they are given to the same use, which leads them to choose inferior alternatives

According to the argumentative theory, reasoning should be even more biased once the reasoned has already stated her opinion, thereby increasing the pressure on her to justify it rather than moving away from it. This phenomenon is called bolstering. Thus, when participants are committed to an opinion, thinking about it will lead to a much stronger polarization.  Accountability (the need to justify one’s decisions) will also increase bolstering.

Finally, motivated reasoning should also affect confidence.

  • When participants think of an answer to a given question, they will be spontaneously tempted to generate reasons supporting that answer.
  • This may then cause them to be overconfident in the answer.

Having to think of arguments against their answer enabled them to see its limitations – something they would not do on their own. It is then easy to see that overconfidence would also be reduced by having participants discuss their answers with people who favor different conclusions.

Belief Perseverance —
Motivated reasoning can also be used to hang on to beliefs even when they have been proved to be ill-founded. This phenomenon, known as belief perseverance, is “one of social psychology’s most reliable phenomena.

The involvement of motivated reasoning in this effect can be demonstrated by providing participants with evidence both for and against a favored belief.  If belief perseverance were a simple result of some degree of psychological inertia, then the first evidence presented should be the most influential, whether it supports or disconfirms the favored belief.  On the other hand, if evidence can be used selectively, then only evidence supporting the favored belief should be retained, regardless of the order of presentation.

Violation Of Moral Norms —
The results reviewed so far have shown that motivated reasoning can lead to poor knowledge outcomes. We will now see that our ability to “find or make a reason for everything one has a mind to do” can also allow us to violate our moral intuitions and behave unfairly.

Reasoning is often used to find justifications for performing actions that are otherwise felt to be unfair or immoral. Such uses of reasoning can have dire consequences.

Perpetrators of crimes will be tempted to blame the victim or find other excuses to  mitigate the effects of violating their moral intuitions which can in turn make it easier to commit new crimes.  This view of reasoning dovetails with recent theories of moral reasoning that see it mostly as a tool for communication and persuasion.

These results raise a problem for the classical view of reasoning.

  • In all these cases, reasoning does not lead to more accurate beliefs about an object, to better estimates of the correctness of one’s answer, or to superior moral judgments.
  • Instead, by looking only for supporting arguments, reasoning strengthens people’s opinions, distorts their estimates, and allows them to get away with violations of their own moral intuitions. I
  • N these cases, knowledge or moral goals are not well served by reasoning.
  • By contrast, argumentative goals are: People are better able to support their positions or to justify their moral judgments.

Proactive Reasoning in Decision Making —
Much reasoning is done in anticipation of situations where an opinion might have to be defended, and we have suggested that work on motivated reasoning can be fruitfully reinterpreted in this light. It is not just opinions that may have to be defended: People may also have to put forward arguments to defend their decisions and actions, and they may reason proactively to that end.  We want to argue that this is the main role of reasoning in decision making.

This claim stands in sharp contrast to the classical view that reasoning about possible  options and weighing up their pros and cons is the most reliable way – if not the only  reliable way – to arrive at sound decisions. This classical view has in any case been vigorously challenged in much recent research.

To What Extent Does Reasoning Help In Deciding? —
It has been consistently observed that attitudes based on reasons were much less predictive of future behaviors (and often not predictive at all) than were attitudes stated without recourse to reasons.

  • This lack of correlation between attitude and behavior resulting  from too much reasoning can even lead participants to form intransitive preferences
  • it  was found that providing reasons led participants to choose items that they were later  less satisfied with or that were less in line with the ratings of  experts
  • People  who think too much are also less likely to understand other people’s behavior
  • In only a few experiments was unconscious thought significantly superior to conscious thought, amounting to a null result when all the experiments were taken into account.  Even so, there was no significant advantage of conscious thought over immediate choice. This is typically  the kind of situation where, according to classical theories, reasoning should help:
    –  A new choice has to be made, with the options well delimited and the pros and cons exposed. It is therefore quite striking that reasoning (at least for a few minutes) does not bring any advantage and is sometimes inferior to intuitive, unconscious processes.
    –  Finally, studies of decision making in natural environments converge on similar conclusions: Not only are most decisions made intuitively, but when conscious decision-making strategies are used, they often result in poor outcomes.

Reason-Based Choice —

  • A choice  based on reasons should be reinforced when participants have to justify themselves
  • A choice based on reasons will be perceived as easier to justify and less likely to  be criticized
  • A choice based on reasons should give rise to more elaborate explanations
  • Participants who made choices based on reasons tended to make choices that fitted less well with their own preferences as stated before the  choice was made
  • When  participants were able to use their intuitions more, because they were familiar with the alternatives or because the descriptions of these alternatives were more detailed, they were less prone to the attraction effect

What Reason-Based Choice Can Explain —
Many other inappropriate uses of reasons have been empirically demonstrated.

  • Investors’ decisions are guided by  reasons that seem good but are unrelated to real performance
  • People will use a rule such as “more variety is better” or “don’t pick the same things  as others” to guide their decisions, even when less variety or more conformity would  actually be more in line with their preferences
  • When forecasting their affective states, people rely on explicit lay theories, which will often lead them astray because “it’s better to keep options open,” people will be reluctant to make an unalterable decision even when they would be better off making it.
  • When indulging in a hedonic act, people feel they need a reason for such indulgence, even though this does not actually change the quality of the experience.
  • Reason-based choice has  also been used to explain effects related to loss aversion the effect of attribute balance, the tendency to be overwhelmed by too much choice the feature creep effect, the endowment effect, aspects of time  discounting, and several other departures from the norms of  rationality
  • Another sign that reason-based choice can lead to non-normative outcomes is that sometimes reasons that are not relevant to the decision will nonetheless play a role. For instance, the same irrelevant attribute will sometimes be used as a reason for choosing an item and sometimes as a reason for rejecting it depending on what decision it makes easier to justify.
  • People will also be influenced by irrelevant pieces of  information because they find it hard to justify ignoring them

All of these experiments demonstrate cognitively unsound uses of reasoning.  There are two ways to explain these findings.

  • One could argue that these are instances of a mechanism designed for individual cognition, and in particular for decision making, that sometimes gets misused.
  • According to the argumentative theory, however, the function of reasoning is primarily social: In particular, it allows people to anticipate the need to justify their decisions to others.
  • This predicts that the use of reasoning in decision making should increase the more likely one is to have to justify oneself. This prediction has been borne out by experiments showing that people will  rely more on reasons when they know that their decisions will later be made public or when they are giving advice (in which case one has to be able to justify oneself
  • By contrast, when they are choosing for others rather than for themselves, they are less prone to these effects because there is then less need for a utilitarian, justifiable decision.  Finally, it should be stressed that the picture of reasoning painted in these studies may be overly bleak: Demonstrations that reasoning leads to errors are much more publishable than reports of its successes.

Indeed, in most cases, reasoning is likely to drive us towards good decisions.  This, we would suggest, is mostly because better decisions tend to be easier to justify.  The reasons we use to justify our decisions have often been transmitted culturally and are likely to point in the right direction – as when people justify their avoidance of sunk-cost mistakes by using the rule they have learned in class.

In such cases, the predictions of the argumentative theory coincide with those of more classical theories. However, what the results just reviewed show is that, when a more easily justifiable decision is not a good one, reasoning still drives us in the direction of ease of justification.  Even if they are rare, such cases are crucial to comparing the present theory (reasoning drives us to justifiable decisions) with more classical ones (reasoning drives us to good decisions).

Dishonest Signals –
But stability is often threatened by dishonest senders who may gain by manipulating receivers and inflicting too high of a cost on them.  Is there a way to ensure that communication is honest?

Some signals are reliable indicators of their own honesty. Costly signals such as a deer antlers or a peacock tail both signal and show evidence that the individual is strong enough to pay that cost saying “I am not mute” is proof that the speaker is indeed not mute.

However, for most of the rich and varied informational contents that humans communicate among themselves, there are no available signals that would be proof of their own honesty. To avoid being victims of misinformation, receivers must therefore exercise some degree of what may be called knowledge vigilance. The task of knowledge vigilance is to evaluate communicator and the content of their messages in order to filter communicated information.

Several psychological mechanisms may contribute to knowledge vigilance. The two most important of these mechanisms are trust calibration and coherence checking.

People routinely calibrate the trust they grant different speakers on the basis of their competence and benevolence.  Rudiments of trust calibration based on competence have been demonstrated in 3-year-old.  The ability to distrust malevolent informants has been shown to develop in stages between the ages of 3 and 6

  • The interpretation of communicated information involves activating a context of previously held beliefs and trying to integrate the new with old information.
  • This process may bring to the fore incoherencies between old and newly communicated information.
  • Some initial coherence checking thus occurs in the process of comprehension. When it uncovers some incoherence, an knowledge ally vigilant addressee must choose between two alternatives.
  • The simplest is to reject communicated information, thus avoiding any risk of being misled. This may, however, deprive the addressee of valuable information and of the opportunity to correct or update earlier beliefs.
  • The second, more elaborate, alternative consists in associating coherence checking and trust calibration and allowing for a finer-grained process of belief revision. In particular, if a highly trusted individual tells us something that is incoherent with our previous beliefs, some revision is unavoidable:  We must revise either our confidence of the source or our previous beliefs.
  • We are likely to choose the revision that reestablishes coherence at the lesser cost, and this will often consist in accepting the information communicated and revising our beliefs.


1. Recently, reasoning has been used simply as a synonym of inference and is then unproblematically attributed to infants or to nonhuman  animals.  In this article, however, we use “reasoning” in its  more common and narrower sense.  The content of the article should make it clear  why we see this as a principled terminological choice.
2.  In the psychology of reasoning, some tasks can be described as production tasks because participants have to produce a logically valid conclusion from a set of  premises. However, these tasks are very different from the production of arguments in  a debate. In a dialogic context, one starts from the conclusion and tries to find  premises that will convince one’s interlocutor. It is this meaning of production that is  relevant here.  Namely, contexts in which real or anticipated argumentation takes place.
3. It may be worth mentioning that what general motivation fails to bring  about is efficient or unbiased reasoning rather than reasoning per se.  If you pay people to get the right answer, they may reason more but will still be as biased, and their answer will still be wrong.
4. Note that motivated, or motivation, as used here do not refer to conscious  motivation based on reasons, as in “I’m going to think of arguments supporting this  opinion of mine in case someone questions me later.”  Instead, it refers to processes  that influence either the direction or the triggering of reasoning in a mostly  unconscious manner.  Even though a lawyer, for instance, can consciously trigger  reasoning and influence its direction, this is the exception and not the rule.  Generally,  people (including lawyers) have limited control over the triggering of reasoning or the direction it takes.
5. Attitude polarization is most likely to occur in individuals who hold a very strong attitude with a high degree of confidence.  The problem is, then, that these individuals will tend to fall at one end of the attitude scale before reading the arguments, which makes it close to impossible to detect any movement towards a more extreme attitude.


One thought on “Using Reasoning to Persuade: Making Sense of “Irrationality” + How Reasoning Helps Groups But Also Drives Bad Behavior

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s