My interest was recently piqued by an article in The Atlantic (October 23, 2018) claiming that “Americans over 50 are worse than younger people at telling facts from opinions, according to a new study by Pew Research Center.” Author Alexis Madrigal summarizes the results of the study: “Given 10 statements, five each of fact and opinion, younger Americans correctly identified both the facts and the opinions at higher rates than older Americans did.” But is the fact vs. opinion dichotomy a viable one? Philosopher John Corvino thinks not. In a hard-hitting article titled, The Fact/Opinion Distinction (The Philosophers’ Magazine, 4 March 2015), he surveys several attempts to elucidate the distinction, and concludes that they all fail.
For instance, some people claim that facts are true, but surely opinions can be true as well. Others assert that facts are directly observable, but many scientific facts would fail to satisfy this condition. (Seen a quark lately?) Still others propose that opinions express beliefs, but this makes no sense either: some of our beliefs are factual beliefs, while others are mistaken beliefs. Nor will it do to define facts are statements which are provably true: as Corvino points out, “proof” is to some degree audience-relative, and there is no absolute yardstick that measures whether a proof works or not. (I might add that few scientific statements are actually provable.) More promising is the proposal that facts have objective content, whereas opinions have subjective content, but many statements commonly classified as opinions (“God exists”) say nothing about subjective states of mind, while other statements which describe subjective states (“Being publicly humiliated is psychologically harmful”) are arguably factual. Lastly, the much-vaunted “fact-value distinction” does not coincide with the alleged distinction between facts and opinions, as some statements which are said to express opinions (e.g. A Democrat will win the White House in 2020) say nothing at all about values. In the end, Corvino tentatively puts forward a definition of his own: a statement of fact, he suggests, is “one that has objective content and is well-supported by the available evidence” whereas a statement of opinion is “is one whose content is either subjective or else not well supported by the available evidence” (italics mine – VJT). But this strikes me as a contrived definition: something is a fact if it is A + B, and an opinion if it is not-A or not-B. In the end, Corvino suggests “that we abandon the ambiguous fact/opinion distinction” and “focus instead on whether people can offer good reasons for the claims they make” – a sentiment which I would echo.
In order to see more clearly why the fact-opinion is philosophically flawed, let’s have a look at the five opinion-statements listed by Madrigal in his article in The Atlantic:
1. Democracy is the greatest form of government.
2. Immigrants who are in the U.S. illegally are a very big problem for the country today.
3. Government is almost always wasteful and inefficient.
4. Abortion should be legal in most cases.
5. Increasing the federal wage to $15 an hour is essential for the health of the U.S. economy.
But what about these statements, which are identical in form to the above statements, and differ only in content?
1′. Longer life is the greatest form of poverty reduction. [See here for why.]
2′. Climate change and pollution are a very big problem for the country today.
3′. Bottled water is almost always wasteful and inefficient.
4′. Self-defense should be legal in most cases.
5′. Maintaining our competitiveness is essential for the health of the U.S. economy.
While statements 1 to 5 above would generally be regarded as opinions, one could make a good case that statements 1′ to 5′ are factual: they are either axiomatic or well-supported by evidence.
I can only conclude that semantics fails dismally to capture the fact-opinion distinction. Pondering this, I had another idea: could we construct an algorithm which would distinguish reliably between facts and opinions? Obviously, we could, if we allowed it to learn from human evaluations of whether a statement expresses a fact or a mere opinion. And in time, I imagine that the algorithm would surpass the abilities of most humans, most of the time. But then I had another thought, after reflecting on Corvino’s final definition of a fact as a statement supported by good reasons. A generalized algorithm for distinguishing a good argument from a bad one sounds like an impossibility to me: the number of different kinds of arguments is potentially infinite, and one could never build an algorithm to evaluate them all. And for that matter, how would the algorithm evaluate arguments relating to its own abilities?
One thing I am certain of, however, and it is this: attempts to school children in the distinction between a fact and an opinion are, for the most part, pure and unmitigated rubbish.
Consider this video by David Snell, which blithely informs readers that facts can be proven, whereas opinions cannot. We have already observed that many scientific statements would fall foul of this criterion. To make matters worse, Snell claims that any statement containing the word “should” should be deemed an opinion. (What about that statement?) Snell maintains that “Pregnant women should not drink alcohol” expresses an opinion, even though he freely acknowledges that studies show drinking alcohol during pregnancy is harmful to the health of the unborn child. I can only reply: if proving that doing X is inherently harmful isn’t enough to show that you shouldn’t do it, then what is?
Or take this video by Imagine Easy Solutions which defines a fact as a statement which is accepted by the majority (!), verified by experts and proven information. Since the video fails to inform its readers what constitutes proof, it is utterly worthless. To make matters worse, the video claims that if the author of a statement is known for having extreme points of view, the statement should be rejected as an opinion. And how does it propose that viewers identify bias in an author? By Googling it. I kid you not. To compound its errors, the video classes all statements preceded by the words, “I think that…” as mere opinions. Finally, statements like “Cronuts are the greatest dessert pastry of all time” are uniformly classified as opinions, but it ain’t necessarily so. Food critics can agree, for instance, on what makes a great pizza (see here and here), and most of us know a good wine when we taste one.
Lastly, this popular video (if you can bear to watch it – honestly, it’s like watching paint dry) informs its viewers: “An opinion is a belief or feeling. It is a person’s thoughts. It cannot be proven.” There is so much philosophical nonsense contained in these three statements that I feel like throwing my hands up in the air. “Shopping is hard work” is classified as an opinion. Oh, really? What about “Trudging around a shopping mall for eight hours is hard work,” or “Carrying a 20-kilogram shopping bag for 2 kilometers is hard work”? Are these mere opinions?
I feel sorry for today’s schoolchildren, who are often forced to take lessons on distinguishing facts from opinions by well-meaning but misguided teachers. They deserve better than this. What say you? Do you think the distinction is a vital one?