A Review of Stefan Molyneux’s The Art of the Argument

The Art of the Argument: Western Civilization’s Last Stand is meant to provide an effective method of argument (hereon named The Argument). But in doing so, it purports to do much more than this. It claims to have the means to perpetual peace and the good life.

The Argument is everything. The Argument is civilization; The Argument is peace; The Argument is love; The Argument is truth and beauty; The Argument is, in fact, life itself.

It provides a division of society into reasonable people, who observe the teachings of The Argument, and dishonest “sophists”. It even provides a pseudo-historical narrative of how The Argument wins out against those who would undermine it.

Let me begin with a disclaimer: I’m a professional philosophical logician. My PhD research focuses on theories of truth and paradoxes relating to them. I have a BA in mathematics and a MSc in logic. I don’t have a background in economics, politics, or biology, so I won’t talk very much about the parts of The Art of the Argument that cover these subjects. Nonetheless, the purpose of Molyneux’s book is to provide the fundamentals of a good argument, which is a subject near and dear to my heart, so I care very much what a popular writer has to say on the matter. I take it that the treatments of economics and biology are primarily meant to serve as instructive examples, and leave the criticism of his treatment of these fields to the respective professionals. As for the political philosophy, a helpful review has already been written by Joshua Stein.

My hope is that readers of this review come away with the following: The Art of the Argument won’t be much help for anyone developing their skills in debate, and it certainly isn’t helpful as an introduction to logic or argumentation. As a foundation for epistemology it does not withstand scrutiny. Molyneux’s philosophical teachings about truth and science are incoherent. The Argument itself, such as it is, is at best very limited in its scope of application, and in practice the book’s doctrine encourages bad faith and further isolation rather than dialogue. I’ve put this much effort into reading the book and writing this review because, frankly, I don’t look forward to attempting to reason with anyone who believes they’ve learned valuable lessons from this book, and if Amazon’s current sales rankings are any guide (#1 in political philosophy for the past week) it should outsell most of its competition.

The Argument and Terminology

Much of the criticism of The Art of the Argument that I’ve seen so far has focused on Molyneux’s defective use of terminology. Primarily, that he misuses terms like “valid” and “sound” that are generally taught in a first-year undergraduate course in logic. Freelance writer Sam Kriss was among the first to point this out:

Kriss’ criticism is fair, but it doesn’t get to the bottom of just what’s wrong with the book. Many newcomers to logic reading the book will not know the terminology being confused, and might not care that it’s being confused. Molyneux demonstrates little respect for virtually any academic philosophy written in the past 50 years, and neither do most of his avid readers. Why should they care about the right terms to use? Does it matter what you call a tool, like a valid argument, as long as it does the job?

In the teaching of philosophy, it matters. Ideally, if you want to teach people to reason with others in good faith, you should teach them in the same standard terminology that the rest of the English-speaking field uses, or at least mention where you significantly differ. Otherwise, you’re training people to confuse others.

Moreover, in the case of The Art of the Argument, the terminology is also suggestive of the limits of The Argument. It has its apparent roots in Molyneux’s notion of truth, which is where his entire project falls apart. To see this, you unfortunately have to more or less read the whole thing, and I’ve obtained 30 free days of Kindle Unlimited in order to allow me to do this and to discourage you from doing the same.

First, I should sketch some basic notions as they are likely to appear in a course on logic. Deductive reasoning is the form of all arguments that are meant to necessarily entail a conclusion from premises, or deductive arguments. If the conclusion is necessarily entailed by the premises, the argument is said to be valid, otherwise it is invalid. Let’s call this notion of validity standardly-valid. If the premises of a deductive argument are true, and that argument is also standardly-valid, then that argument is said to be standardly-sound.

Inductive reasoning is the form of all arguments that are meant to entail a conclusion from strong evidence given for it by premises, or inductive arguments. Abductive reasoning can be taken to be the form of arguments that start with a set of evidence, and then select a hypothesis that is meant to best fit them all.

So let’s go over Molyneux’s examples of “bad deductive reasoning”. His first example is:

1. All plumbers can swim.
2. Bob knows how to swim.
3. Therefore Bob is a plumber.

Many readers would understand this as the formal fallacy of affirming the consequent (though perhaps “knows how to swim” does not mean “can swim”). But confusingly, he does not point this out. Instead, he remarks, “Of course, not all plumbers can swim, so why bother reading further?” This remark might at first glance seem irrelevant. An even more baffling example of “bad deductive reasoning” comes next:

1. Kind people are socialists.
2. Bob is a kind person.
3. Therefore Bob is a socialist.

This is a standardly-valid argument. Yet, Molyneux claims it is not valid. So what’s going on here?

The examples are confusing because Molyneux’s notion of validity is based on a notion of deductive reasoning that is also non-standard. In short, Molyneux mixes up soundness and validity.

Molyneux remarks shortly before he starts to introduce his examples of bad deductive reasoning that:

Deductive reasoning is the process of extracting general rules from specific observations, and then applying them to new empirical information.

This definition appears to be extrapolated from a “general to specific” definition of deductive reasoning, the kind that might appear in a dictionary, combined with Molyneux’s view of what truth is (more on that later), but it does not conform to the spirit of deductive reasoning. For logicians and philosophers, standard definitions of terms like “deductive reasoning” and “validity” are found in textbooks or classic texts, not dictionaries.

Molyneux’s corresponding idea of logical validity is very strange and has caught some early readers off guard. It does, however, shed some light on his deviant idea of what deductive reasoning is. First, there is a notion of what he calls a “valid premise” and what I’ll call a Molyneux-valid premise, which is that in line with “the scientific method” valid premises must be “logically consistent”, and “conform with — and predict — empirical data”. Presumably, then, a deductive argument is said to be Molyneux-valid when all of its premises are Molyneux-valid, and when the conclusion is necessarily entailed by the premises. All Molyneux-valid deductive arguments are standardly-valid, but not vice versa.

Incidentally, the idea of a Molyneux-valid premise and argument sounds a lot like a rough definition for a true premise and a sound argument. But it is specifically limited to propositions about empirical data. I’ll say more about that later.

OK, with that in mind let’s return to the two examples.

1. All plumbers can swim.
2. Bob knows how to swim.
3. Therefore Bob is a plumber.
1. Kind people are socialists.
2. Bob is a kind person.
3. Therefore Bob is a socialist.

The first example is not a Molyneux-valid argument because it is not standardly-valid and its first premise is not Molyneux-valid (i.e. not true), assuming there exists a plumber who can’t swim. The second example is standardly-valid but its first premise is not Molyneux-valid (i.e. not true), assuming there exists a kind person who is not a socialist, therefore the second example is not a Molyneux-valid argument either.

Curiously, Molyneux soon goes on to describe the following argument as being “logical” or “logically sound”:

1. All men are immortal
2. Socrates is a man
3. Therefore Socrates is immortal

This is a standardly-valid argument which is not standardly-sound. But being Molyneux-sound is the same as being standardly-valid. The meaning of “sound” and “valid” has been entirely reversed from that seen in undergraduate textbooks the world over.

Again, with inductive reasoning, Molyneux seems intent on a defective dictionary-style definition:

Inductive reasoning attempts to draw general rules from specific instances.

This is not exactly inductive reasoning either. The Internet Encyclopedia of Philosophy provides one example of an inductive argument among many that does not involve a specific-to-general inference: “I saw her kiss him, really kiss him, so I’m sure she’s having an affair.”

Molyneux goes on to make a long-winded analogy of how deductive and inductive reasoning can be compared to predator and prey, alpha and beta males, and proactive and reactive behaviour. For brevity’s sake, I won’t provide an account of all of this, other than that the analogies Molyneux makes of arguments are confusing rather than instructive. After all this effort at clarification, he adds:

Inductive reasoning moves from general observations to specific probabilities, rather than from general rules to specific instances.
Inductive reasoning takes a gathered set of general observations and attempts to apply a rational theory to explain them.

These characterise two different phenomena, neither of which is inductive reasoning as Molyneux previously defined it. The first paragraph describes the assignment of a probability to a hypothesis. And if anything, what the second paragraph describes is abductive reasoning. The fact that he mentions abductive reasoning a couple of pages later, seemingly as an afterthought, just adds to our confusion.

Molyneux’s sloppiness in the discussion of fallacies comes out in a discussion of “synonym logic”, which appears to be a neologism of his.

“Synonym logic” occurs when two terms are introduced that seem different but turn out to be the same — you may also refer to this as a “tautology.”

I should mention that I have never seen such a thing referred to as a “tautology”. I’ll grant this is a relatively minor point in the grand scheme of things (in any case, tautologies are never mentioned in the book again). Going by this definition, and the first example Molyneux gives, I take it that by synonym logic he refers to some sort of rhetorical trick making two terms appear to have different meanings while they are the same:

In my university debating days, the example given was “Coke is it” — where the word “it” turned out to be defined as “coke.” This less-than-stellar intellectual achievement of proving that “Coke is coke” is a stirring reminder of Aristotle’s first law of logic, but nevertheless remains unimpressive to anyone over the age of four.

And yet, his third example of synonym logic is a completely different phenomenon:

In the realm of political debate, the word “education” generally turns out to mean, “coercively funded government schools,” which is not the same thing at all, since education takes many forms. However, by creating this false equivalency, opponents of coercively funded government schools can be easily painted as opposing education as a whole.

This is not synonym logic as I have by now come to understand it. Instead of being two terms presented as having different meanings when they are the same, these are two terms being presented as if they had the same meaning when they are different.

Molyneux’s sloppiness here, and the many digressions he takes, suggests that despite appearances he is not really interested in cataloguing the formal or informal fallacies that people make. Instead he is invested in what he claims is a long-standing ideological conflict between the forces of reason, as embodied in The Argument, and sophistry. And whatever The Argument is, going by his idea of truth and deductive reasoning, it must concern itself with empirical data alone. So, we continue to ask, what is The Argument?

The Premises of The Argument

The path to The Argument unfolds dramatically, claiming to draw from thousands of years of intellectual thought. When one understands it and can employ it one is enlightened, ready to live the good life and teach others. A good deal is written about how you might then suffer personal hardship for using The Argument, but as the proverb goes, virtue is its own reward. In this way The Art of the Argument could be viewed as more of a self-help book for those invested in Molyneux’s brand of right-wing politics.

Molyneux distinguishes between “value arguments”, to do with ethics and aesthetics, and “truth arguments”, to do with reality. As his own past book Universally Preferable Behaviour (2007) is referred to as still being indicative of his own views on the right approach to reasoning about ethics and aesthetics, and it has been widely reviewed, from here on in I will talk about truth arguments.

An important characteristic of The Argument, which is why it’s meant to play such an important role in the life of anyone who understands it, is that it is universally applicable to any nonviolent disagreement:

The Argument occurs every time you try to convince someone else of your position or preference without using threats or force. The Argument lives in language every time you negotiate. The Argument comes to life every time you accept being rationally opposed without resorting to punishment, either directly or indirectly.

The Argument is nonetheless governed by “essential premises”, under which it must be held:

1. Arguments have the potential to be objectively better or worse.
2. A better argument is one that more accurately reflects objective reality.
3. Objective reality is both rational and empirical.
4. Therefore, a better argument is one that is more rational and/or supported by more empirical data.

I will argue from here that these premises already leave us lost, without a good sense of how we have a better argument or a worse one.

As Molyneux repeatedly asserts, it is important in any argument that we first agree on definitions for all our terms. Wanting to be agreeable myself, I looked for Molyneux’s definitions for “rational”, “empirical”, and “objective reality”. Coming up short, I dug a little deeper, into Universally Preferable Behaviour, and had no luck there either, though there is some suggestion as to why:

I respect your intelligence enough to refrain from defining words like “reality,” “reason,” “integrity” and so on. We have enough work to do without having to reinvent the wheel.

These terms appear to be taken to not require a definition, which is fine, as I doubt the helpfulness of Molyneux’s contention that all our terms have to be defined.

In any case, the fourth premise doesn’t provide much help for us to understand what makes an argument better. What appears more rational might not be supported by more empirical data, and vice versa. Molyneux acknowledges this, but doesn’t tell us which of the two gets preference and when.

Molyneux starts to outline a methodology in “The Argument and Reality”:

We gain reason and science by using the direct evidence of our senses to extrapolate and discover general rules, which we can then apply to contradict the direct evidence of our senses.

and:

The Argument first requires logical consistency, and only then empirical verification.

One notion of the logical consistency of a theory is the requirement that it does not entail a proposition and its negation. There might be some explanatory value for a logical theory that is not consistent in this sense, as long as there are sufficient restrictions on what it can entail. Without digressing further, I invite the reader to look up Graham Priest’s Beyond the Limits of Thought (1995).

The earlier point about “reason and science” is more pertinent. Molyneux notes that reason can contradict the direct evidence of our senses, which is fine when your senses suggest the earth is flat. But he doesn’t provide an account of how and when to revise our general rules when they contradict empirical evidence.

The problem of how and when to revise these general rules is not obscure. It is not a bit of intellectual self-indulgence. It has been a central theme of the philosophy of science since antiquity and there is still no universal agreement about the standards of revising a scientific theory. Here, we are not just limited by what we know, but by uncertainty around what constitutes good evidence.

Molyneux does speak a lot of concepts. Here, concepts are mental representations of objects or categories, which are “derived from entities in objective reality”. They “describe reality” and are said to be “absolute”, “valid”, or “objective” if they accurately reflect reality. In that sense, he argues, they become part of objective reality.

Inductive reasoning recognizes that patterns are not absolutes unless they can be moved into categories. Concepts can be absolutes, instances cannot be. Moving the aggregated information of sense data and experience into absolute concepts — moving from data to conjecture to absolute, or from sense impression to inductive reasoning to deductive reasoning — is the true path of philosophy.

The goalposts have now shifted, but the problem is essentially the same: how and when do we know our concepts are absolute? This remark suggests that Molyneux doesn’t see where the problem is:

In general, anyone who claims that an absolute can result from inductive reasoning is wrong. You may arrive at absolutes from deductive reasoning, but rarely if ever from inductive reasoning.

Valid deductive arguments — either in the standard sense or Molyneux’s sense — do not provide further empirical evidence by themselves. If you couldn’t arrive at Molyneux’s absolutes from inductive reasoning, you’re certainly not going to from deductive reasoning either.

Molyneux acknowledges that reason can lead us into superstition and error, but he does not explain how we are meant to know that we’ve gone astray. When he goes on to define truth he may be intending to clarify things, but as we will see, his notion of truth is incoherent.

The Argument and Truth

The conditions below are taken by Molyneux as being immediate consequences of his four essential premises of The Argument:

[5.] Objective truth exists.
[6.] Objective truth requires conformity to reason and evidence.
[7.] Objective truth is universally preferable to falsehood and/or incomprehension.
[8.] Language can effectively and accurately interpret and describe objective truth.

I don’t see how points 5–8 are consequences of 1–4, though Molyneux tries to motivate them later on. But even if we take 5–8 as premises instead, as we have seen, the standard of conformity to reason and evidence is a big ask. If we are not careful, there would be a conflict over what constitutes a true proposition.

In the vast literature on what truth is, there are many ideas of how to define truth. The most well-known, which Molyneux clumsily attempts to adopt, is the correspondence theory of truth, often attributed to Aristotle, that the truth of a proposition is determined by a relation between that proposition and some portion of reality. He presents the following idea of what that relation is:

Truth is in general a word that means “a rational statement about empirical reality that can be objectively verified”.

This is affirmed by premise 6. So we see that a necessary condition for a statement to be true is that it is empirically verifiable. However, such notions of truth are self-defeating. After all, the statement “all truths are empirically verifiable” cannot itself be empirically verified, but presumably we are meant to assume it is true.

The self-defeating character of proclamations like “all meaningful statements are verifiable” was a serious issue with much of the doctrine of logical positivism, which reached its height in the late 1920s to early 1930s, defended with a great deal of care and deliberation, but nonetheless long dead by now. Those curious about logical positivism should read A. J. Ayer’s Language, Truth, and Logic (1936). A seminal paper in the criticism of logical positivism is Willard Van Orman Quine’s “Two Dogmas of Empiricism” (1951).

If in the last section it could be said that we were merely left unsure of what to do with The Argument, by now we’re in serious trouble. There is a contradiction in the heart of the doctrine of The Argument: the proposition that all truths are empirically verifiable cannot be true.

Even if we somehow dig ourselves out of this contradictory morass, perhaps by a stratification of truths into “verifiable” and “given”, with only the verifiable truths in the scope of The Argument, we have failed to define truth itself. Furthermore, we have fatally undermined Molyneux’s claim that The Argument is universally applicable to all disagreements.

A final remark should be made about point 8, that “Language can effectively and accurately interpret and describe objective truth.” Going back to my discussion of the essential premises of The Argument, I do not believe that the deficient account of how to decide between general rules and empirical evidence provides us with an effective way to do this.

The Argument and Its Form

Having seen the premises of The Argument, we can finally discuss its form. According to Molyneux these are the “essential elements” of The Argument:

[A] It [an instance of The Argument] must be a rational proposition, consistent with itself and empirical reality.
[B] It must make claims about empirical, testable reality.
[C] It must contain a null hypothesis… some standard by which it could potentially be disproven.
[D] Its terms must be objectively definable, to the satisfaction of both parties.
[E] Both parties must agree to abide by reason and evidence, at the expense of their own propositions.

Regarding point A, the question of the extent to which The Argument should be consistent with itself and empirical reality is one that has already been discussed. It is not clarified in The Art of the Argument or Universally Preferable Behaviour what a proposition is. In the spirit of this book, and in line with the examples Molyneux gives, I’ll follow Ayer and take a “rational proposition” to be a well-formed statement in English that is either true or false. When Molyneux speaks of a term, I take it to be the name of an object, or of some property it has, or of some category it falls under.

But without digressing too much, I should emphasise that there are good reasons for why there is no widespread agreement in contemporary philosophy on what a proposition is. Sometimes, for example, a proposition is thought of as an abstract entity that is expressed by a sentence. This helps us to investigate the meaning of a sentence with greater clarity, and perhaps to include information about the context of the sentence’s utterance that the sentence itself might not have. However, I’m not sure how such a proposition could be said to have terms, and “make claims”, as points B and D would have it, and it does not fit in with the rest of Molyneux’s discussion.

What of point B, that any statement in The Argument “must make claims about empirical, testable reality”? Then, given the universal character of The Argument, there is nothing else worthy of discussion. Thus, discussions of mathematics and logic are only virtuous when they have a particular application in mind. He doubles down on this here:

…are there arguments unrelated to empiricism? In other words, are there abstract arguments that never overlap or impact on objective, empirical reality?
I would say that there are, in the same way that there can be Klingon or wood elf biology — such abstract “disciplines” can exist, but so what?

Unfortunately, this is not a helpful restriction. Much of pure mathematics has had an intriguing record of gaining applications after its development. Number theory was a fertile area of mathematical research for hundreds of years before many of its results were applied in computer science and cryptology. When theoretical physicists and mathematicians started to develop the foundations of quantum mechanics, many of the mathematical ideas and results that they needed in linear algebra and functional analysis were already there.

A question this leaves us with is just what it is that makes a mathematical theory potentially applicable, but this discussion would take us too far from the subject of The Argument.

Points D and E present the pragmatic conditions of The Argument. Two people have a disagreement, and decide to put their propositions to the test. They do this by first reaching an agreement on exactly what they mean by everything that they say, then they work out how compatible what they’re saying is with reality.

Molyneux repeats on several occasions that arguments are, in the vast majority of cases, usually very easy to resolve as long as two parties have agreed on the definitions of their terms beforehand. And being precise about what one means, he argues, is a virtue. Thus, he implements the demand that two parties agree on the definitions of what they say beforehand to each-other’s satisfaction as condition D of The Argument and here:

For all comparison debates, all terms need to be defined to the satisfaction of all parties. (Once this is achieved, the debate is usually concluded very easily.)

I agree that clarity is often a good thing; there are many misunderstandings that could be resolved with a little time taken beforehand to make sure of what we mean. But this strategy only goes so far. If two parties agreed to only argue whenever they could agree on the meanings of their terms, that would eliminate a broad scope of substantial disagreements. This is not a good thing for The Argument, as The Argument is meant to be universally applicable, yet many (if not most) disagreements of substance ultimately hinge on what a term means. A few relatively explicit and well-known examples include:

  1. What is, and what constitutes, dark matter.
  2. The fundamental or central characteristics of autism. Autism is typically characterised by a list of symptoms but not by a particular cause, unlike Fragile X syndrome. There is widespread disagreement in the psychological literature on what features of autism are more fundamental, thus on what autism is.
  3. The meaning of human intelligence.
  4. The characterisation of race (as an anthropological classification).

The meanings of the terms we use are typically underdetermined by our knowledge of what they refer to. Often where there is underdetermination of meaning, there is also scope for reasonable disagreement on definitions, and thus an impasse before The Argument in its present form can even begin.

Universality

One demand that Molyneux makes of any argument, not discussed among the conditions of The Argument, is that it is universal. One sense of a universal proposition is that it purports to be part of some universal theory that concerns all objects, settings, and times. This is embedded in his idea of what philosophy is:

Philosophy is the process of moving patterns of experience into universal absolutes, which apply in all places, for all time.

The example he gives is “a theory of physics”:

A theory of physics cannot apply merely to Melbourne Australia, or the dark side of the moon, from 2 to 4 AM Eastern Standard Time.

The curious thing about this is that, in the sense of a theory being universal that we’ve used here, there is no extant and generally agreed-upon universal theory of physics. The “mathematical treatment of the axioms of physics” is David Hilbert’s sixth problem, presented among Hilbert’s famous list of twenty-three problems in 1900. The problem remains open, particularly as there has not been widespread agreement on a unified field theory.

Given what I’ve said about The Argument so far, a demand that propositions should be statements that form part of a single physical theory of all objects in the universe is just a non-starter. For we have seen that the scope of The Argument is limited to those topics for which the meanings of terms are not subject to reasonable disagreement. A single physical theory of all objects in the universe, assuming it could even be coherently formulated at this point, must employ terms whose meanings are subject to disagreement (it may include a treatment of dark matter, for instance). So instead of The Argument being limited in its scope, we would not be able to use The Argument at all.

Yet Molyneux provides a completely different notion of a universal proposition in a subsequent example:

If it is not universal, it is not an argument; it is merely a statement of subjective preference. “I like ice cream,” is not an argument. “Ice cream contains dairy,” is an argument, since it claims to describe a property objectively measurable and testable within empirical reality (not merely subjective perspective or preference).

That is to say, a proposition is said to be universal if it is a claim about objective reality, as Molyneux would have it. This is a much weaker, less controversial demand, already provided in the “essential elements” of The Argument, but Molyneux does not distinguish between these two different demands of universality.

Molyneux has said a lot more about the demand that propositions that make moral or aesthetic judgments are universal. As we’ve mentioned before, these concern the kinds of “value arguments” that aren’t being discussed here, as they’ve already been discussed elsewhere. The relevance of value to truth for Molyneux is briefly explained:

These two kinds of arguments are not completely independent: for a truth argument to have value, we must value the truth; for a value argument to have meaning, it must be true.

Taking for granted that engaging in The Argument is meant to be universally preferable behaviour only amplifies the concerns made in the previous sections. For we have seen that the scope of The Argument is limited to empirical topics for which the meanings of terms are not subject to disagreement among competent practitioners, so engaging in it would not be universally preferable.

Practical Considerations

I have spoken about the form and setting of The Argument and shown that the scope of The Argument, far from being universal, is limited to topics for which two competent practitioners would agree on the precise meanings of everything they say, and in practice this often cannot happen.

I wish to conclude from this that The Argument is of little use to someone wanting to learn how to argue effectively. I might expect a challenge to this along the following lines: there is a great deal of denial of facts that, on careful examination, are already clearly evident. The Argument is merely a practical doctrine: that in a dispute two people should first agree to the same definitions and then to discuss the matter of their disagreement only in propositions that concern objects in reality rather than abstractions. And in many cases, someone in the dispute has simply been in denial of the truth, which should come to them through The Argument.

My response to this is that the limitations of The Argument show how destructive it would be as a practical doctrine. Sometimes when two people disagree it is easy to check who is right and who is wrong. But often people disagree on something with reasonable but mutually intractable grounds, and the best course of action is to just move on to another topic.

Contrary to this, Molyneux’s ethical teachings dictate that someone who does not agree to the terms of The Argument should be ostracised:

Philosophical standards must be defended through ostracism. Intellectual ostracism, or refusing to engage with people who cannot rise to minimum standards of rationality, is an essential and often overlooked aspect of our society, and it pretty much runs everything, once you recognize it.
Societies that encourage ostracism flourish, because their social rules contain rewards and punishments. Societies that disallow or punish ostracism lower — and eventually destroy — the value of The Argument, because failure to conform to The Argument no longer carries penalties.

He even provides a utopian scenario of a stratification of society: the first tier of society being comprised of those who consistently agree on the same terms and standards of verifying reality and thus can employ The Argument, and the second tier being comprised of those “sophists” who do not.

Ostracizing the incompetent from The Argument also serves to create two sub-societies within the larger society: the society of those who respect The Argument, and the society of those who do not. Given that The Argument as a whole rests upon empiricism, having these two opposing epistemological methodologies in society can be enormously instructive. Place a conceptual wall between these two opposing camps, and see which one does better over time, and which does worse.

The practice of this, if it were widely observed, would be unlikely to lead to a division of society into two parts, but rather to a fragmentation of society into individuals and cults of dogma, a new fragmentation appearing (at least) with each substantive disagreement on the meanings of terms. There will be little hope of surviving any future challenges that humanity as a whole might face without a basis for collective agreement. So much for perpetual peace.

Conclusions

There is a great deal of the book that I thought objectionable and still did not comment on, either because I felt it would be too much of a digression or I didn’t feel qualified to speak on the matter. For instance, Molyneux argues that women are typically more prone to “hysteria” in the face of a well-reasoned argument. There is also much that I’m sure could be said about his observations on evolutionary psychology, child psychology, and the history of philosophy and science. I am also leaving out a discussion of his questionable scholarship. Particularly dubious is his reading of Plato, where the “sophists” are identified with the gods of the city, and Socrates’ last words are a statement of gratitude that he finally rid himself of his irrational fellow Athenians. I particularly welcome other, more qualified writers to comment on these topics as they are treated in the book.

Molyneux appears to embrace a doctrine of avoiding accusations of making the correlation-causation fallacy by carefully avoiding the stipulation of any cause whatsoever. He also speaks of confirmation bias as only being a problem if you’re wrong, but a virtue if you’re right. Both of these views gravely undermine his own project.

The Art of the Argument purports to have a means of resolving disputes by reference to facts. But if The Argument were widely adopted, society would likely fragment into cults of dogma. The search for the good life continues.

For newcomers to logic and argumentation, I recommend the following books:

Introductions to logic:

Irving M. Copi and Carl Cohen — Introduction to Logic, Fourteenth Edition (Pearson Education Ltd., 2014) [For those without a mathematical background]

Elliott Mendelson — Introduction to Mathematical Logic, Fourth Edition (Chapman and Hall, 1997) [My own introduction, but only for those who already have a mathematical background, i.e. are experienced with the practice of mathematical proof]

L.T.F. Gamut — Logic, Language and Meaning, Vols. 1 & 2 (University of Chicago Press, 1991) [Provides an introduction to logic from the perspective of its applications to formal semantics, culminating with possible worlds semantics and Montague grammar. For those familiar with the content of a mathematical logic textbook like Mendelson’s only Vol. 2 is necessary.]

Introductions to argumentation and rhetoric:

Anthony Weston — A Rulebook for Arguments, Fourth Edition (Hackett Publishing Co., 2009) [A concise instruction for undergraduate students across disciplines on how to write persuasively]

Stephen Toulmin, Richard Rieke, and Allan Janik — An Introduction to Reasoning, Second Edition (Macmillan Publishing Co., 1984) [A wide-ranging introduction to the study of argumentation]