The ongoing (and coming?) health care debate will no doubt be a gold mine of sloppy and dishonest reasoning. We've already noticed some examples of this already. Just as the debate over gay marriage seems to inspire certain particular patterns of fallacious reasoning (the equivocation on "marriage" and the slippery slope), I think the health care debate will have its own definitive fallacies. At the moment, I'm thinking that we'll see a lot of red herring–changing the subject from the less appealing facts of the matter (for instance, the fact that Americans pay more for health care and get less than other developed nations) to tangentially related, yet incendiary, notions such as "socialism."
But I think we'll also see a whole lot of weak analogy–in particular comparisons of health insurance to any other complex consumer product. Here's one from George Will yesterday:
Some advocates of a public option say health coverage is so complex that consumers will be befuddled by choices. But consumers of many complicated products, from auto insurance to computers, have navigated the competition among providers, who have increased quality while lowering prices.
Those things are different in that they are largely optional purchases. Sure, you "need" them, but you don't need them. I might mention, by the way, that auto insurance is legally mandated for all drivers (yet another difference from health insurance–and I doubt, by the way, that Will would advocate such a mandate). In any case, before one starts comparing health insurance to any other consumer product, one ought to take note of the vast differences. Few products typical consumers (i.e., anyone of any income level) would absolutely have to buy involve possible outlays of hundreds of thousands of dollars. And few of those products carry with them (often in their fine print) the real possibility of physical and financial ruin.
In the interest of fairness, I should point out that this entire piece, however bad, does not argue against the feasibility or desirability of single-payer health coverage. In fact, it does a lot to make the case for it (though not on purpose). Will's purpose is merely to argue against the "public option." I think his argument is bad (citing as it does Mort Kondrake and a health insurance industry funded study), but I think such an option is a bad one (for other reasons).
John, I have more of a general question: when is it ever acceptable to use the word “some” in an argument. It seems that whenever one uses it, one is asking for trouble. The word “some” seems to be the door to quite a few fallacies.
Webster defines some as “ being an unknown, undetermined, or unspecified unit or thing“. So, in theory, at least, some could mean none 🙂
I think I’m in need of some explanation.
“Some” could mean “none”? (Since I’m at the keyboard, I thought I might take a shot at this.)
Not by itself. “Some” is a quantifier, equivalent to “There is.” The ony way for a statement of the form “Some x is P” to be true is if there is some thing “a” such that Pa is true.
Depending on one’s quantifier theory, where the “none” might slip in is if there is a negation hiding in the statement. Thus, if “P” is a short-hand for “~Q” (which is arguably cheating) then the above becomes “Some x is ~Q” which is true even if a much stronger claim is true, namely “No x s Q” (or “For all x, there is no x that is Q”). “Some” does not mean “none” in this instance, it only allows it to be the case because the existential quantifier (“some”) is true as long as something (which could include everything) is not Q.
This does not work in Aristotelian logic which (if memory serves) treats “some” as (more or less? strictly?) convertible with “some not”. So one cannot slip and slide on the negation and say “some” when really “none” is the case.
Good explanation, Gary!
I guess John is right, never look at the dictionary to understand the meaning of the words. To be honest Webster has another definition for the word that serves us better: “ being at least one —used to indicate that a logical proposition is asserted only of a subclass or certain members of the class denoted by the term which it modifies”
Let’s go to the other extreme: Could some mean all?
@GeorgeW: Auto insurance quality has increased, while prices have been lowered? It seems pretty scam-ridden and deceitful to me.
@JohnC: I’d be curious to hear your reasons against the public option some time–particularly whether they’re theoretical or practical. I’ve resigned myself to the fact that congress is too timid and impotent to pass single payer this time. The way things have been going this last week, we’ll be lucky to even get the public option.
@BN: In most logics, if it’s true to say all P are X, then it’s likewise true to say “some P are X.” But more pertinent to this blog, which analyzes real discourse, and maybe more to your point, is the question of whether it’s misleading or deceiptful to use “some …”, when it’s true that “all …”. The answer is, it depends.
If I know that all P are X, but I say only some P are X, then, according to Grice’s maxim’s, I’ve failed to be appropriately informative. So while what I said was true, it was “unsatisfactory,” because it “implicates” that some P are X, while some P are not-X. I’ve spoken a truth, strictly interpreted, but misled you, because you would rightly assume that, were it the case that all P are X, I would have said that instead; “all” is always more informative than “some.”
That’s not the whole story, though. Sometimes it’s the case that all P are X, yet “some P are X” is both true and wholly appropriate/satisfactory/felicitous. Cases like expressions of subjective epistemic states. Say a bag is filled will balls–this much I know. But all of them are red, and that I don’t know. I draw two balls, which are both red, and you ask me what I know about the balls in the bag. The most I am in a position to truthfull assert is that some of the balls are red, even though all of them are in fact red. So once again, the “some” assertion is true, but this time, it’s felicitous as well, and so our theory about some can’t frown on all such uses.
Could some mean all?
I would resist the terminology of “means” either all or none. In most orthodox systems of discourse, ‘Some x is P’ follows from the fact that “All x is P’, but it does not mean the same thing. That is, if P is true of all x, then one can certainly find an x for which P is true — provided there are any “x’s” at all! (There are subtleties here that are typically glossed over for purposes of simplicity and exposition. For example, if there is no such thing as a ‘Gleeb,’ then is it true or false that ‘all Gleebs are red’? If there are no Gleebs, then one cannot find a single instance that disconfirms the claim that they are all red. Various Meinong inspired area’s of logical theory study quantifier relations in the absence of any concrete commitments to the existence of the things being quantified over.)
Again if memory serves, the Aristotelian rules keep “all” and “some” more strictly separated so that “all” does not automatically include the case “some”, since for the syllogistic rules “some” also implies “not all.” (Presumably an obvious fact by now, but my background is a bit uneven and definitely weighted more toward contemporary logic.)
great points, Jeremy!
Maybe the word “some” should be banned from public discourse. Instead, one should just be specific. To use your example: Don’t tell me that some balls are red, tell me that the 2 you saw were red.
Again, maybe it’s just me, but I don’t find the word “some” very useful; it’s rather confusing; it’s a quantifier that does not do a great job of quantifying 🙂