Obstacles to Good Reasoning 4: Burden of Proof and Premise Acceptability

Burden of Proof

To argue or not to argue, that is the question. When someone makes a claim, burden of proof determines whether the arguer or the opponent must provide an argument. Burden of proof refers to the obligation of either arguer or challenger to provide further support for their primary claim (or its negation). How do we figure out who bears the burden of proof?

To determine who bears the burden of proof we'll first need to take a step back. Recall that arguments can be decomposed into premises and conclusion(s). A burden of proof can concern any of the premises or the conclusion, however, let's first focus on the concept as it applies to conclusions; i.e., main claims.

Two Heuristics for Determining Burden of Proof
When we evaluate who bears burden of proof, we can apply two heuristics:

Heuristic 1: Will the intended (specific) audience accept the claim without further support? If yes, then the burden falls on the opponents to the claim. If no, then the burden falls on the arguer.

Heuristic 2 (better): Would a universal audience (i.e., a group of reasonable people) accept the claim without further support? This is a much better standard of burden of proof for constructing your own arguments because usually we aren't trying to convince people who already agree with us on issues, we're trying to convince a general audience.

When we evaluate an argument for burden of proof with the second heuristic we are essentially asking if its conclusion is something a reasonable person would accept (i.e., universal audience). If the assertion is reasonable, then the opponent bears the burden of proof to show that we should not accept the assertion. If the assertion is unreasonable, then the arguer bears the burden of proof to show (with further supporting premises) why we ought to accept the particular assertion.

When the arguer's conclusion isn't reasonable (i.e., when the burden of proof falls upon the arguer's conclusion), an argument must be made! That is, he's now going to have to support his conclusion with premises (i.e., reasons and evidence). If, in turn, any of the premises are considered unreasonable, then they too will have to be supported with further premises. That is, he will also bear the burden of proof to support those premises until he arrives at support that the audience does accept. This should remind us of Sidgwick's insight: Begin your arguments with premises your opponent/audience accepts (if you want to lead them to your conclusion).

Now we can return to the formal definition. A burden of proof refers to which party (arguer or challenger) has the obligation to defend their position. It speaks to reasonableness of an assertion (be it a conclusion or a premise). The person who opposes whatever is considered reasonable bears the burden of proof—that is, it's up to them to convince the audience (through argument) that the default position is unreasonable or incorrect. Without a supporting argument, we have no good reason to take their point of view seriously.

Let's look at a few examples:
When people deny that the moon landing happened, the burden of proof is on them. They are taking a position against all experts and mountains of testimonial and physical evidence. The burden falls upon them to show why we should reject the reasonable position of thinking people landed on the moon. The reasonable position is that people landed on the moon; to assume otherwise would require (significant) further argument.

When people say that the earth is only 6 000 years old, the burden of proof falls upon them. It's up to them to show why multiple converging lines of evidence are mistaken in their implications and why the theory upon which modern geology and biology are founded is incorrect. It's reasonable to think that virtually all geologists are well qualified to determine what theories and evidence do or do not apply to to the age of the earth. To assume a claim that implies that virtually all geologists are wrong requires further argument than mere assertion.

One last note on burdens of proof:
Historically, burdens of proof can shift. So, what was a reasonable assumption a few hundred years ago might be unreasonable today. We see this with social assumptions. For example, it wasn't too long ago that it was reasonable (for men) to assume that women weren't capable of math and science. Someone (back then) assuming the opposite would bear the burden of proof. Today, that burden of proof has shifted.

Economics is one area where the burden of proof is shifting. It used to be the common assumption that humans are (classically) rational—always seeking to maximize personal interest along the lines of classical mathematical rules. Behavioural economics, interdisciplinary psycho-economics, and socio-economic theory are starting to show these assumptions are wrong. Giving this mounting empirical evidence, the burden of proof is shifting concerning economic models built upon the assumption of (classically) rational agents.

It's important to keep in mind that simply because a position bears the burden of proof it doesn't follow that the claim is false. Burden of proof just tells us who has to provide evidence for their position. Historically, marginalized groups have had to bear the burden of proof when they argue they are treated unfairly. Since they are almost by definition minorities their point of view will not be the mainstream view. Also, new scientific theories always bear the burden of proof when they seek to overturn older ones (as they should. If the evidence for the new theory is superior it will eventually win out. Burden of proof only refers to who has to provide the argument) it doesn't tell us that a view is true or false. So, maybe in the future we will discover mountains of evidence that the moon landing was a hoax and that the earth is 6000 years old. If this happens the burden of proof will again shift…

Inappropriate Burden Shifting:

burden%20of%20proof.png

One argument tactic is to inappropriately shift the burden of proof. This often occurs when the burden of proof is shifted onto the person with the reasonable claims. Generally, if your view is a minority/outlier view, the burden of proof is on you. Also, the person making the existential claim (i.e., that something exists) typically must bear the burden of proof because it's impossible to prove definitively that something doesn't exist.

Conspiracy theorists love this one trick! They'll say stuff like "Well, you can't prove that the government DIDN'T know" or "You can't prove the government ISN'T in on it." Since it's impossible to prove a negative, this is inappropriate burden shifting. They're making a positive claim about what X knew or did. They need to provide evidence. The same goes for the anti-GMO crowd. They'll often claim that GMOs cause obesity or other health problems by pointing to increasing rates. Of course, the rise could be caused by any number of things (over-eating? lack of exercise?) but they will often inappropriately shift the burden of proof making the challenger prove there is NO causal link: i.e., "You can't prove that GMOs aren't causing the rise in X."

Argument Jiu Jitsu
When constructing a strong argument, whenever possible, try keep the burden of proof on your opponent. Hai-ya!

Premise/Evidence Acceptability

Premise Acceptability:
When it comes to value-based arguments (e.g., political, moral, and religious) we will often not be able to achieve 100% certainty of truth of the premises. Even something as intuitively obvious like "it's wrong to kill innocent people" will have counter-examples. However, simply because a premise cannot be 100% true in all cases, it doesn't necessarily mean that we should reject it or that it's part of a weak argument. Instead, it means we need to be subtle in our evaluations and consider carefully what the logical consequences are of accepting certain premises. We need to evaluate claims probabilistically and relative to the reasonableness of competing claims.

Recall from our previous discussion of premise acceptability that premises/evidence need to meet general two standards:

(1) Would the intended audience accept the claims without further support? (i.e., do the premises take into account the audiences "cluster" of beliefs and values) and
(2) Are the claims acceptable to a "reasonable" universal audience.

Acceptability is closely related to burden of proof. From the perspectives of both (1) and (2) we need to ask:

(3) Would these claims be accepted without anyone asking for more evidence/support?

And

(4) Is there no available information or evidence that we know of that contradicts the claims? (Think back to confirmation bias and falsificationism.)

If the answer is "yes" to both, then the premise is acceptable. If not, then it isn't, and the burden to provide further support for that premise rests upon the person making the argument. An acceptable premise is one that passes both the intended audience and the reasonable universal audience test.

Questionable vs Unacceptable
Going back to the reasonable universal audience test, if the answer is "no", then we have two possible ways to proceed: Depending on the next steps in our evaluation the premise will be"questionable" or "unacceptable."

Evidence/premises are questionable when

(a) they fail the "reasonableness" test but we still haven't come up with any direct contradicting claims or
(b) we just don't have any knowledge either way on the topic in regards to what a general audience would accept or
(c) the wording of a premise is too vague to evaluate. Very few of us are experts on everything (despite what some people might think!).

When you say a premise is "questionable" you need to also include in your evaluation what sort of evidence/additional support the arguer would have to support for their premise to be true. Then google it!

Let's look at some examples: Suppose I want to claim that it's an affront to the Olympic spirit that wrestling be removed from the Olympics. We might point out that the phrase "affront the the Olympic spirit" is quite vague and could be variously interpreted. So, in our evaluation of the acceptability we might say that the claim is questionable because its meaning is vague.

However, at this point we wouldn't yet be able to say that claim is unacceptable. An unacceptable claim is one for which there is known evidence that directly contradicts the claim. Since, in the wrestling case there isn't obvious evidence against the claim (because we're not quite sure exactly what it means), it is not unacceptable, merely questionable. From the context of the argument we might apply the principle of charity and try to ascribe a reasonable meaning to the phrase.

The evaluation of "questionable" could also arise where we just don't know much about a topic. Consider an argument that was commonly heard in the Food Stamp funding debate:

(P1) Giving people food stamps removes the incentive to find work and leads to chronic unemployment.
(C) Therefore, we should restrict or defund the Food Stamp program.

(P1) on its face seems intuitively plausible. If people are just going to be handed something, why work? The fact is, however, we aren't sure, so we say the premise is questionable.

Since we've said it's questionable we need to also specify what sort of evidence or support would be needed for it to be acceptable. This might be government data on un/employment rates of people on food stamps that shows people on food stamps have high unemployment rates and continue to be unemployed while on food stamps.

Let's Google and find out:
It turns out that children, the elderly and disabled—i.e., people who, normatively speaking, we don’t expect to fully participate in the workforce—account for 84% of SNAP recipients. Also, contrary to what the premise needs for support, 58% and 62% of SNAP households with an able-bodied adult work (no children/children) and 82% and 87% are employed within a year.

This data doesn't give the premise the required support and there's evidence that contracts it so we say it is unacceptable. Since the premise is unacceptable this particular line of argument against the Food Stamp program fails. If I were to claim that the moon is made of cheese, this claim would be unacceptable since there is known evidence that directly contradicts the claim.

To summarize: The difference between a questionable and an unacceptable claim is that with the former, it fails the reasonableness test because we don't have enough information (maybe it's too vague or we don't know enough about the topic); and with the latter there is known evidence that contracts the claim that a reasonable audience would be aware of.

The Nitty Gritty: How Do We Know What a Reasonable Audience Will Accept?

For the most part this is an element of subjectivity here but there are main two guidelines we can begin with to determine if a reasonable audience would accept a claim: (A) The claim doesn't contradict any of the other claims made in the argument and (B) the claim could be defended in front of an audience comprised of a broad cross-section of society. The second reason can be broken down into five specific reasons to accept a claim.

1. Acceptable By Definition or Self-Evidently Acceptable
Some premises are definitions. Often, (but not always) definitions are considered self-evident or true by definition. For example,

A triangle is a three-sided figure or
a bachelor is an unmarried male.

You can't argue with that. For definitions, the litmus test is what a community of language-users would accept as a definition for a term. If I use a word completely differently than everyone else, my usage is unacceptable because it doesn't conform with standard use.

We can also have claims that are self-evident because they are logical truths, like "a thing cannot both exist and not exist at the same time" or "a statement can't both be true and false at the same time."

Self-evident claims can also pertain to the moral realm (unless your audience is philosophers!). Statements like, "causing unnecessary suffering is bad" or "killing innocent people is bad" are considered self-evident to a reasonable general audience.

2. Acceptable as a Factual Statement Reporting an Observation or as a Statement of Eye-Witness Testimony
Unless we are provided with some reason not to, we accept people's testimony about what happened to them or what they observed. If someone said "it's sunny outside yesterday," you have no reason not to accept the claim. Also, if I told you that I was going to the store to get myself a cold pop, you should accept my claim at face value.

Of course, if I have a reputation of being a big fat liar, then you'd have reason to be skeptical of my claims. Also, if what I said happened to contradict something else I'd said, then you would also have reason to question the acceptability of my claim. But barring such situations, we take speaker observations and testimony at face value until or unless we have reason not to.

3. Acceptable by Common Knowledge or Assent
We have to be careful with this one because it can lead us to accept things uncritically that we probably shouldn't. "Common knowledge" can be divided into two categories: factual claims and value claims. Factual claims describe the world. Value claims suggest how things should be.A factual claim would be something like, "the earth is a spheroid" or "Obama is the current president of the USA." A value claim or value judgment would be something like, "it's wrong to hit children" or "we shouldn't allow people to starve."

So, where does acceptability fit into this? Although we said previously that acceptability depends largely on what a reasonable member of a universal audience would accept, there are exceptions. Concerning factual claims, if the intended audience has specialized knowledge (doctors, scientists, etc…) then it's OK to evaluate the claim in relation to the knowledge base we'd expect that group to have.

We might reply, but wait! What happens if the knowledge isn't possessed by every member of that audience? In such cases, we can make assumptions about what factual knowledge we'd reasonably expect the members of a specific expert audience to have.

In all this talk of specific audiences, let's not lose sight of the "common knowledge by a universal audience" aspect of this criteria. Just because some knowledge might be particular to a field of study or expertise, doesn't mean that there isn't knowledge that we can reasonably expect Joe Schmo to know. Stuff like, "the third Batman movie was awful" or "grass is green" or "Las Vegas is the entertainment capitol of the world" are all things that we'd expect a general (North American) audience to know and so we can accept them at face value. Similarly "Uzbekistan was part of the former USSR" is something we'd reasonably expect every general audience (outside of the USA) to know.

4. Acceptable Because it is Defended in a Reasonable Sub-Argument
In Mill's proof of utilitarianism he makes the sub-claim that the general happiness is the supreme good that all humans desire. This is on its face is not a claim that we'd expect a universal audience to accept. However, Mill, knowing what a general audience might not accept, provides a supporting argument working from our particular desire for our own happiness to the more general claim.

Since he supports his sub-claim with a reasonable argument we can now accept it. (And consequentially turn our critical thinking toward the supporting sub-premises.)

When evaluating for premise acceptability, we can do the same with any claim/premise. It it doesn't seem reasonable we can see if it is supported by a sub-argument/premise. If it is, and that sub-argument is reasonable, then we can accept the claim/premise.

5. Acceptable on the Authority of the Arguer or an Expert
We can broadly divide this criteria into two types: uncontroversial claims made by an arguer and claims made by an expert.

In the first class these are claims about relatively uncontroversial things that the arguer might know about. For example, I might say that the University of Houston has a good philosophy graduate program. Because I've been there and you haven't, you have no grounds to doubt my claim and since it isn't particularly contentious, it should be viewed as acceptable.

In the second class we have claims made by experts. This is known as an "appeal to authority". This is when the arguer supports a claim by appealing to the expert knowledge of a person, institution, or source. It's important that the expert is an expert in the topic of the claim. For example, arguing that climate change is false because of famous physicist said so is an illegitimate appeal to authority. If a claim is about climate change, a legitimate appeal to authority would appeal to the views of a climate expert. Just like we wouldn't ask a climate scientist about particle physics, it doesn't make sense to ask a physicist about the climate. A common tactic of science denialists, conspiracy theorists, and pseudoscientists is to appeal to experts that aren't experts in the topic at hand.

An additional note: the best appeals to authority are appeals to the consensus opinion of a community of experts. Appeal to a single expert doesn't carry much weight, especially in controversial topics. The opinion could easily be an outlier. Also, not all experts's view carry equal weight. The views of a community college instructor don't carry the same weight as those of a famous Harvard professor in the same field.

Conditions of Unacceptability

1. Unacceptable because of an Inconsistency:
We might label a claim as unacceptable because it is inconsistent with other claims the same arguer has made. For example, in alt-med we often see arguments stating that the flu vaccine shouldn't be used because its efficacy can sometimes be as low as 45%. However, the northern Andean magic rainbow-berry or acupuncture they propose instead has no reported efficacy.

So, if the argument is that we should reject a treatment because of low efficacy, the same should apply to their own conclusion.

We also hear that we shouldn't take manufactured (i.e., "unnatural) drugs because big-pharma's just trying to sell you stuff to make money, then in the same breath the same website or person will try to sell you the latest all-natural (!) miracle cure…and not for free either!

Here's another example of inconsistency:
internal%20inconsistency.jpg

And here's yet another example of inconsistency (from the beginning until about :45):

To be sure, inconsistency doesn't mean the conclusion is false, it only means that the argument for the conclusion is poor and we need to either reject the premise(s) or the conclusion (or both) because both can't be true. Sometimes people can hold the right views (conclusions) for the wrong reasons or as a matter of dumb luck, not because they arrived at them through good argument.

2. Unacceptable because of Begging the Question
The current use of "begging the question" meaning "raises the question" is something that irks philosophers to no end. The original (and TRUE!) meaning of the phrase is "circular reasoning." In other words, in your argument, you assume to be true the very thing you are trying to prove.

The classic example comes from an unsophisticated religious argument for the truth of the contents of the Bible. A caricature of the argument goes like this:

How do you know what's in the Bible is true? 'Cuz it's the word of God. How do you know it's the word of God? 'Cuz it says so in the Bible.

Notice that for the argument to work you have to assume the very thing the arguer tries to prove: that the contents of the Bible are true.

3. Unacceptable because of Language Problems
We can call a premise unacceptable if it has one or more of the language problems we encountered on the section on definitions. In other words, if the language of the premise is overly vague or suffers from semantic and/or syntactic ambiguity, we might say it's unacceptable (if context can't reasonably sort it out).

Suppose I claim that "I've never been seriously sick since I started taking Tibetan Booga-Booga Bush capsules. We can't accept this claim because "seriously sick" is too vague. For how long and for what intensity do I have to be sick to be "seriously sick". Do I need to be hospitalized or just miss work? Or maybe just miss my work out. It's not clear from the phrase so we'd say the premise is questionable or unacceptable depending on the severity of the vagueness.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License