Evidence answers the question “How do you know that this is true?” When evaluating evidence, consider the following questions: Is this true 100% of the time? Was there ever a time this wasn’t true? How could this view be proven false? Is there enough evidence to convict this thought in a court of law?
Since evidence is an important part of an argument, it is crucial to learn how to evaluate the evidence people give for their claims. There are four major types of evidence:
i.) Examples (Anecdotes): These are specific experiences that support the claim. Anecdotes can elicit strong emotions but can never prove cause and effect and you can generate anecdotes for almost any claim.
ii.) Statistics: Statistics are generalized conclusions about certain populations of people, situations or events drawn through the scientific method. They differ from anecdotes in that they have a large enough sample size that is selected randomly and in a controlled way.
iii.) Tangible objects: These are physical objects that act as proof. This includes documents, records and personal possessions.
iv.) Testimony: A testimony is a witness that something is true. We rely on the testimony of experts in areas that are outside our scope of expertise. You need to evaluate the credibility of a witness in terms of competence, trustworthiness, track record, consensus and good will.
v.) Social consensus: And the final type of evidence to consider is the opinion of society as a whole and includes shared values, assumptions and worldviews. The test here is whether a critical audience of informed peers would accept our conclusions and the evidence we give for them or not.
After you have brainstormed and listed all the evidence that supports your belief, you would then move on and do the same thing with evidence that contradicts your belief. This is usually the much more difficult as people tend to struggle with recognizing why their beliefs might not be true.
After brainstorming and listing all of the evidence that you can come up with you need to evaluate each piece of evidence and give it a rating on a scale from 1-10. Consider the limitations of anecdotes, check for statistical data, examine tangible objects, evaluate the credibility of testimonies and compare your belief to social consensus. Once each piece of evidence is evaluated, assign an overall score to each category. As a general guide, each category of evidence can be worth 20% and the total of all categories will form the final grade. A belief with strong evidence will be ranked from 7-10, ambiguous or moderate evidence will score between 5-6 and weak evidence will score between 0-4. Here are some other guidelines to consider when evaluating evidence:
- Weak Evidence – This kind of evidence is subjective, few people would believe it and a judge would throw it out in court. This includes rumors, opinions, hearsay and speculation.
- Moderate Evidence – This kind of evidence is somewhat objective but there isn’t enough of it or it is ambiguous and could mean many other things. An example might be: “It happened before” or “Other people agreed with me.”
- Strong Evidence – This kind of evidence relies on the facts and avoidance speculation, narrative and interpretation. Few people would contest this kind of evidence and it would achieve broad social consensus.
Another model suggests that we can use evidence to either describe, evaluate or prescribe a particular response.
i.) Descriptive evidence involves simplifying complicated phenomena. It simply describes something to an audience.
ii.) Evaluative evidence involves assessing instead of describing solutions. A case study is an example of this kind of evidence.
iii.) Prescriptive evidence argues in favor of a certain course of action. Instead of describing or assessing it prescribes a particular response.
Just like we can evaluate claims and evidence, we can evaluate the inferences that link a evidence to a claim. There are six common patters of inference that include:
i.) Inference From Example: This involves reasoning from parts to whole (generalizing) or whole to parts (Classifying).
a.) Generalization: We generalize when we say: “What is true of the parts is probably true of the whole.” Statistical generalizations are evaluated in terms of sample size and representativeness. Anecdotal generalizations are evaluated in terms of the number and range of examples given, presence of counter-examples and representativeness. We can make the fallacy of composition when we assume that what is true of the parts HAS to be true of the whole.
b.) Classification: Conversely we classify when we say “what is true of the whole must be true of the parts.” The test for this is: “Does the member really belong in the general class or is this member an exception?” We can make the fallacy of division when we assume that what is true of the whole HAS to be true of the parts.
ii.) Inference from Analogy: This involves reasoning by making comparisons as we assume that like things should be treated alike. We can therefore compare something new to something we know and treat them the same. There are literal and figurative analogies:
a.) Literal Analogy: This involves making a direct comparison between objects ,events and situations in the same sphere of reality. The assumption is that because something worked in a similar situation, it should work here. The judicial analogy is an argument that we should stick to past precedent when dealing with new situations. A super analogy occurs when we say that ‘because we acted this way in this lesser circumstance we must act this way in this greater circumstance.’
b.) Figurative Analogy: This involves comparing the relationships amongst things and not the things themselves as they are in different spheres of reality. For example, cutting social security because we can’t afford it is like burning the barn down to kill the rats. Both are examples of the cure being worst than the problem. Analogies are tested to see if the essential similarities outweigh the differences and if the answer is no then a false analogy has been presented.
iii.) Inference from Sign: This involves recognizing and predicting correlations. The assumption is that since one thing exists it usually predicts the other exists. Infallible signs are rare as most signs are fallible and rely on probability instead of certainty. For example, if you score well on tests it is assumed that you are smart. We use this inference in three ways: inferring the unknown from the known (we know your test score, not your IQ), predicting outcomes without explanations (studying correlates with higher test scores) and relying on the expertise of others (education is a sign one is informed).
a.) Tests: Evaluate how strong the correlation is? Are there countersigns that also exist that may signify something else? Does the sign also correlate with other outcomes? Is the person assuming causation instead of correlation?
iv.) Cause: This involves inferring that one thing causes another. A prediction involves reasoning that a cause will lead to an effect while an explanation involves trying to explain an effect by identifying a potential cause. Causal inferences rely on probability and not certainty, since we can’t directly observe what happens. There are three types of causes: A sufficient cause means a single factor can predict the outcomes. A necessary cause means that the factor must be present for the outcome to occur. A contributory cause means that a factor increases the probability of a cause but isn’t enough in itself.
a.) Tests: We use scientific experiments in which two things are alike in every way except one and then attribute differences in outcome to the difference we controlled for. Quantitative approaches involve doing a regression analysis to determine how much of the variance is attributable to each variable. We also userhetorical approaches to identify causes and providing arguments for why they should be considered causes. Look for the post-hoc fallacy which involves: “Since y happened after x, x must have caused y.” Look for other causes or other effects from the claimed cause. Look for counteracting causes which reverse the effects of something.
b.) Three Criteria: Temporal Precedence (cause occurs first), Co-variation of cause and effect (when cause doesn’t occur, neither does the effect) and elimination of alternative explanations. Does the identified cause meet these three criteria?
v.) Commonplaces: This involves making inferences based on social knowledge our audience accepts as true. The inference is that because it is accepted as true by the audience, it’s okay to state it as true.
a.) Pragmatism vs Principle: Pragmatism involves making choices based on consequences and principle is making choices based on enduring principles.
b.) Quantity vs Quality: Quantity means we focus on the greatest good for the most people at the least cost. Quality means we focus on achieving the best outcome for specific people in specific situations.
vi.) Form: Making inferences from what looks like deduction but it is not. These include dilemmas, arguments from hypothesis and reasoning from comparisons.
a.) Dilemmas: A situation in which two unattractive choices are given as the only options. Determine if this is afalse dilemma in that other options actually exist that aren’t being named.
b.) The argument from hypothesis: resembles the conditional syllogism (if-then) but actually isn’t. Testing your hypothesis and getting the result isn’t enough to determine cause as you need to rule out alternative explanations. This would be affirming the consequent which is a fallacy.
c.) The third common form is reasoning from comparisons which involves making imprecise comparisons such as: “Are you better off today than last year?” such things are hard to measure.
vii.) Hybrid Patterns: This involves reasoning by rules, values and dissociations.
a.) Reasoning with Rules: “If condition x occurs then y is required, permitted or forbidden.” This involves use of the judicial analogy (past precedent) and classifying a current situation as fitting into a category of situations. The tests include: Do the facts of the case justify activating the rule? Have all aspects of the case been considered? Is the rule being misapplied? Is the rule consistent with past precedent? Is the rule realistic and appropriate?
b.) Reasoning with Values: Most arguments about values are about where in the hierarchy a value should be rather than whether something is inherently good or bad. You can argue that one value subsumes another and allows accepting both. You can argue that a value leads to better consequences or promotes even more important values to both sides. You can also argue that one value is more supported by the experts.
c.) Reasoning with Dissociation’s: This means taking a single concept and dividing it into two concepts, one of which has more value. You then suggest your position is associated with the greater value and your opponents with the lesser. You then associate it with a philosophical pair which may include appearance/reality, letter/spirit and opinion/truth. Tests include: Are the two parts of the concept really distinct or are they the same? Is the philosophical pair correct in which part should be more valued? Does the dissociation really reframe the controversy?
Validity and Fallacies
A strong argument is one that would convince an audience exercising critical judgment. We should also ensure our arguments are valid in that they follow patterns that experience has shown lead to good results. The following is a list of fallacies to avoid:
1.) For inferences from example, valid arguments avoid these fallacies:
a.) They avoid hasty generalizations.
b.) They avoid unrepresentative samples.
c.) They avoid the fallacy of composition.
d.) They avoid the fallacy of division.
2.) For inferences from analogy, an argument is valid if the similarities of the analogy outweigh the differences.
3.) For inferences from sign, valid arguments satisfy the following tests:
a.) The sign and the thing signified generally occur together.
b.) The sign does not appear by itself, without the thing signified.
c.) There are not obvious countersigns.
d.) The same sign does not herald opposite things.
e.) The relationship is not mere coincidence.
f.) Sign is not confused with cause.
4.) For inferences from cause, valid arguments will avoid the following errors:
a.) They avoid confusing sign with cause.
b.) They avoid failing to identify a common cause.
c.) They avoid confusing temporality with causality.
d.) They avoid confusing cause with effect.
e.) They do not ignore multiple causes or multiple effects.
f.) They do not ignore intervening or counteracting causes.
5.) Valid inferences from commonplaces will show that the commonplace applies to the situation at hand more than do any competing commonplaces.
6.) For inferences from form, valid arguments are those that recognize the conclusion is based on probability and not certainty.
Identifying Fallacies of Clarity
Many fallacies result from a deficiencies in clarity caused by the imprecision of language.
i.) Equivocation is using the same word to convey different meanings in the same argument.
ii.) Ambiguity results when we aren’t sure what specific meaning a term has.
iii.) Amphiboly results when we aren’t sure what specific meaning a phrase has.
iv.) Vagueness results when terms or phrases are used that do not have specific meanings.
v.) Heaps and slippery slopes result when boundary lines aren’t specified. With heaps you can’t tell when something has changed and with slippery slopes any change becomes a difference in kind. The Fallacy is that something can’t stop once it has started.
vi.) False Equivalence: Focusing on a few attributes between things and then pretending they are the same. For example, both cats and dugs are fluffy, so they are basically the same.
vii.) No True Scotsman: This is a form of circular reasoning and is illustrated with the following example: Someone says “All True Scotsmen are brave” but then an example of a cowardly Scotsman is pointed out and the person responds: “Then he is no true Scotsman.”
Identifying Errors of Vacuity
Errors of vacuity are empty arguments that do not provide sufficient proof for claims.
i.) Circular reasoning occurs when evidence for the claim is simply the claim in different language or the evidence presupposes the claim to be true.
ii.) Begging the question occurs when claims are made that depend upon other claims that have not yet been established as true. “Have you stopped beating your wife yet?”
iii.) Red Herring: Sometimes claims are simply ignored and irrelevant material becomes the focus of an argument. The name comes from fox hunts, when red herrings would lead dogs off fox trails.
iv.) A non sequitur is an argument where there is no connection between the claim and the evidence given for it. The term means “it does not follow” and simply means there is no inference linking evidence to claim.
v.) A “straw man” occurs when an opponent changes your argument to make it weaker and attacks the “straw man” argument instead.
vi.) A self-sealing argument is one that cannot be proven false because opposing results will still confirm it. It is an unfalsifiable claim.
Identifying Fallacies of Relevance
These fallacies point to something that has nothing to do with the relationship between claim and evidence.
a.) Ad hominem arguments involve attacking the person rather than the argument the person is making.
b.) Appeals to authority are fallacious when the authority has no basis to claim expertise in a given area.
c.) Appeals to popularity (“bandwagon effect”) occur when a claims popularity is used as evidence in place of real evidence that would support the claim.
d.) Appeals to tradition occur when a claim is taken as true just because it’s what traditionally has been true.
e.) Appeals to ignorance occur when you assert that a claim that cannot be disproven must therefore be true.
f.) Appeals to inappropriate emotion occur when strong emotions are elicited that interfere with arguments.
g.) Threats can also be used in place of arguments to coerce agreement.
h.) Appeal to Consequences: Believing or disbelieving something based on whether you like the consequences or not. “I don’t want to live in a world where children starve to death so it can’t be true.”
i.) Poisoning the well: You don’t address the argument directly but try to associate it with something unpopular.
j.) Moving the Goal Posts: When you change the criteria for accepting a claim after the criteria are met to avoiding having to accept it. “I will believe your theory if it is replicated three times.” After the experiment is replicated three times the person may move the goal posts: “Five more replications and I will believe.”
k.) False dichotomy: Oversimplifying things by presenting a false dichotomy and ignoring grades in the middle or other options. “You either agree with my policy or you are a racist.”
When Fallacies aren’t Fallacies
There are circumstances where many of these “fallacies” are actually appropriate arguments. For example, Ad Hominem attacks can be appropriate under certain circumstances:
a.) The bad character (or “abusive”) type of argument asserts that a person who makes a claim is untrustworthy.
b.) The circumstantial type asserts that the person making the claim does not act consistently with it.
c.) The bias type asserts that the person has a vested interest in a certain claim being true and therefore biased.
d.) My idea is sacrosanct: This argument states that an idea is too sacred and does not need to be defended.
There are also times where you can inappropriately accuse something of using a fallacious argument.
a.) The “fallacy fallacy:” This occurs when you believe that proving a certain argument is fallacious means what you are arguing for is false. An argument for something can be invalid and yet that invalid thing can still e true.
b.) The Nominal-Fallacy Fallacy: This occurs when you simply name a fallacy and believe it has invalidated an argument. For example an appeal to authority is sometimes valid and sometimes it isn’t yet you could always just say “appeal to authority” to try to shut someone down.
The best fallacies can do: “Your argument instantiates a form which may or may not be an error; therefore it is in error!“ “Arguments can be fallacious and still valid. An argument is valid if it can get from point A to point B by any possible route; it is invalid only if one cannot rationally get from A to B at all. If there is a valid route from A to B, it doesn’t matter whether there is an invalid route from A to B, or many, or an infinite number (which there are: there are always an infinite number of invalid routes from A to B you can propose).”
There are three main spheres within which arguments take place:
i.) The Personal Sphere: People argue through dialogue and are motivated to resolve personal disagreements relevant only to them. Only those involved evaluate each other’s arguments.
ii.) Technical Sphere: These arguments occur between experts in highly specific fields and their arguments are accessible by other experts.
iii.) The Public Sphere: These arguments occur over issues that affect everybody and ideally everybody would have access to these arguments.