To generalize from a small group to a larger group requires a representative sample. A specific or particular is, almost by definition, not such a sample. As they say, the plural of anecdote is not data. It may be that this this Steinberg would be uninterested in actual data. Perhaps he'd be unwilling to defend those who kill abortionists on the same terms. And perhaps truly representative samples would contradict his beliefs. I don't know the guy. But the argument as stated seems correct. It may be that someone from China is more likely to speak Mandarin. But I've met people from China who don't.

Granted, the author chose some truly lousy representatives for the counterargument, bordering on committing the falacy he accuses others of. But he doesn't explicitly make any generalizations about a defined group based on the arguments that his respondants put forward.

Just as many cardiac surgeons have higher-than-average rates of heart disease
Very interesting. Is it true? I've often wondered about pathogenically transmitted arterial calcification since there seems to be a fair bit of support for its liklihood.

I'd like to gently reiterate a point made not too long ago:

Quite often, when I say something controversial, I notice that the person to whom I am speaking suddenly starts acting as if I had made a completely different assertion. They place statements in my mouth that I haven't made; they respond to points no-one in the room is making, and which are completely different from what I have said.

The converse seems true as well: they seem to act as if I've disagreed with statements I just said were true. This is, of course, a tad frustrating. It's like living in some kind of bizarro-world where everyone seems to be locked in eternal "opposite day" mode.

Above, I wrote:

If I say the Dutch are unhygenic based on a sample of one individual, that is indeed fallacious reasoning. But if I survey a more reasonable-sized group of the Dutch and observed that most of my sample spoke, shall we imagine... DUTCH, it would not necessarily be fallacious to generalize about the larger from the smaller.

Ryan, here was your response:

To generalize from a small group to a larger group requires a representative sample. A specific or particular is, almost by definition, not such a sample.... the argument as stated seems correct. It may be that someone from China is more likely to speak Mandarin. But I've met people from China who don't.

So, I clearly pointed out, and agreed, that one couldn't generalize from a unrepresentative sample to a larger one — and then moved on to point out my core objections. And your refutation of this was... to claim he was "correct"? About the one minor point I already agreed he was correct about, though for entirely wrong reasons?

How is that even a refutation of what I wrote? What on earth value is that supposed to add to the conversation?

Further, if Mr. Steinberg really believed what he claims, he would be completely opposed to all polls, which, by their very nature, are examples of "extrapolating from the specific to the general."

Polls are not examples of extrapolating from the specific to the general, but from a staistically representative population to a larger population. Your argument that Steinberg should, if he is being consistent, not accept polls seems to rest on this point.

But if I survey a more reasonable-sized group of the Dutch

This only works if the sample is a random, representative sample of the larger population. This is a sometimes-crucial, difficult step that people easily skip. You could sample 10,000 Dutch college students and infer from that that the Dutch spend most of their time studying textbooks. Similarly, people can hear dozens of reports of school shootings and infer that shootings are on the rise without asking whether the increase in reports are representative of actual incidence, say.

Polls are not examples of extrapolating from the specific to the general...

Um, huh? Now you're getting completely silly. Polls ask things of specific (as you put it) "representative" individuals, who are thus taken, in aggregate, to "represent" (that is, stand in for or typify) a larger population. Polls certainly don't put questions to vague or abstract prototypes. Consider the various meanings of "represent" — your own choice of word. There is no such thing as a non-specific "representation" or "representative" of something else.

This only works if the sample is a random, representative sample of the larger population.

Um, yes, of course. Again, where did I assert otherwise? Please, again, read my complaint regarding putting arguments into my mouth which I haven't actually offered. And read more carefully.

And if that still wasn't clear enough: I didn't say having a larger sample was a SUFFICIENT condition, as you seem to be implying. I simply pointed out, again, that a larger sample "would not necessarily be fallacious" to extrapolate from. (Which, again, is precisely what Mr. Steinberg's assertion implied.)

Ponder a hypothetical:

R: "You implied all dogs are purebreds."

T: "No, I said some dogs were not necessarily mutts."

Your seeming insistence on this point is even odder given that I just plainly stated, in my last comment, that I agreed with you on this very point (bold added):

I clearly pointed out, and agreed, that one couldn't generalize from a unrepresentative sample to a larger one...

How on earth are you reading that as meaning that I disagree about the importance of representative data? Again, please re-read my complaint about people who take agreement with them on some point and pretend I've just said the exact opposite, as you have now done a second time, on the same exact subject.

---

Finally, and for the record: Why yes, you're correct, I didn't write the word "representative" the first time around. Why? Um, because it I assumed it was obvious, given that his suggested sample size was ONE? But, I suppose, while I'm at it, and, again, for the record: Yes, I also didn't explicitly note that samples had to be, as you have also now noted, random (that word was also omitted!). Nor did I note they to be of a statistically significant size in proportion to the target population. Nor did I discuss margin of error. Nor, quite frankly, did I even mention anything about the possibility of systemic errors or bias.

Why, oh, why, did I fail to allude to such CRUCIAL details?

Oh, how about because it seemed so blindingly beside the point? This isn't a field guide on sampling. My goal, again, on that aside, was only to point out valid extrapolations exist, and move on. If someone had questions, they're welcomed to ask more. That might make my argument one you wouldn't, emotionally, prefer. (Undoubtedly, it would have been vastly better had I hyperlinked to a guide to sampling, or, even better, immediately diverted the discussion for several paragraphs, as we have here, now, to a more extensive (and yet still incomplete!) treatment.) But it doesn't make it factually incorrect.

Geez.

Ryan, if someone doesn't say something, it's really poor form to insist they've taken some incorrect position on the unspoken issue. That wastes everyone's time. If you want to chime in, and demonstrate your knowledge in that area, great, have at it. A bit of a tangent, but if it makes your day, wonderful. But PLEASE don't write it as though you assume I disagree, and, indeed, was trying to disagree with you on some painfully obvious point.

To say "most people cannot lift a car" is not to say that a group of people cannot lift a car. An individual is not a family or a nation, though both are composed of individuals. A specific example is, by definition, not a collection or representative sample since in that case it ceases to be specific.

I mean, if you produced an example of him actually arguing that extrapolating from poll data or similar really was a form of racism and that we could therefore not rely on it I'd readily believe your assertion that he means what you say he means. And for all I know, he may actually be that crazy. Otherwise, though, it seems reasonable, based only on the text given, to not twist Steinberg's words to mean something which they don't.

That might make my argument one you wouldn't, emotionally, prefer.

I didn't disagree with many of your conclusions, as I've noted. I do disagree with your assertion; "Further, if Mr. Steinberg really believed what he claims, he would be completely opposed to all polls" because I think such an assertion misrepresents what Steinberg is claiming.

Emotionally, the only statement of his that I'm really attached to here is that I think people across the political spectrum do too much argumentation based on the particular anecdotes that they're exposed to and Steinberg, his other arguments aside, was right to criticize it.

Quite often, when I say something controversial, I notice that the person to whom I am speaking suddenly starts acting as if I had made a completely different assertion.

This seems to be exactly what you're doing here, with the word 'speficic' where you interpret it as possibly meaning 'not specific.'

Incidentally, I seem to remember something else in your reply that isn't there now about people typically reasoning from anecdotes and the prices at grocery stores. Perhaps it was late and I imagined it? Or did you alter your response?

In any case, I won't post on this thread again since its probably not worthwhile for either of us.

Best,

Ryan

A bit out of order:

This seems to be exactly what you're doing here, with the word 'speficic' where you interpret it as possibly meaning 'not specific.'

What are you referring to? Where is "here"? (Quoting is so helpful!) What instance of the word "specific" am I interpreting as "non-specific"?

Incidentally, I seem to remember something else in your reply that isn't there now about people typically reasoning from anecdotes and the prices at grocery stores. Perhaps it was late and I imagined it? Or did you alter your response?

I DO sometimes alter my own comments (and articles), right after posting, but not after I see someone else respond -- and I get that the moment they post it. (I have trouble proofreading in the input form.) But I will never intentionally alter stuff retroactively to make a response look less plausible; I view that as grossly dishonest.

And yes, there was a blurb about grocery stores in prices, which I thought was badly done, overly confusing, probably too argumentative (I don't recall, it was late for me too), and unhelpful, and so I edited it out upon re-reading.

So no, you weren't losing your mind or overly tired. My apologies for any confusion.

To say "most people cannot lift a car" is not to say that a group of people cannot lift a car...

Um, yes? And so? Again, what does this have to do with anything? I see that you & I both used the word "most" at least once (I applied to an unspecified quantity of Dutch), but I fail to see how your analogy (?) maps to, refutes, or answers anything said previously.

I'd readily believe your assertion that he means what you say he means.

Which assertion is any way unbelievable? I have made several, and defended each. Each, as far as I can see, seems perfectly sensible and reasonable. Again a specific quote here would be really helpful. As it is, I am left to keep guessing as to whether this charge is the same as your next one, or is a separate objection.

And why do you think I'm saying I know what he means? One of my points has been, repeatedly, the exact opposite: I don't know what he means. So I'm not sure what view you think I'm saying he actually holds. (Nor will you deign to tell me, apparently.)

I do disagree with your assertion; "Further, if Mr. Steinberg really believed what he claims, he would be completely opposed to all polls" because I think such an assertion misrepresents what Steinberg is claiming.

First, THANK YOU for actually citing something specific. That makes responding infinitely easier. I still have to guess which half of the sentence you object to (you don't say, so I have to write two separate responses), but that's still better than in your other allegations against me.

(1) If the first half: you seem to be misreading it. You seem to be treating my statement an assertion when, to the contrary, it begins with a conditional: "IF [he] really believes what he claims..." — that is, his claim regarding the wrongness and/or immorality of going "from general to specific" and vise-versa. I didn't at all say he truly believed that, as you seem to be asserting. (Indeed, my point was quite the opposite...)

(2) On the other hand, if you're referring to the second half, and again implying you believe I'm being unfair by "misrepresenting" the nature of polling so that it fits within Mr. Steinberg's dictum:

Sure, you're welcomed to do so. You've given one objection (you apparently do not believe it in any way reasonable to treat polls as a form of extrapolating from specific instances to a more general class), and, in response, I've provided several reasons why that assertion is incorrect.

If you still wish to believe that, that's fine, but I'm certainly not arguing for my understanding of polling (and thus my understanding of it as covered by Mr. Steinberg's rule) in bad faith, or to "twist" anyone's views (if you mean this to be covered by your previous accusation against me).

I think people across the political spectrum do too much argumentation based on the particular anecdotes that they're exposed to and Steinberg, his other arguments aside, was right to criticize it.

Again, please note I how regularly beg and plead for specifics. (And receive almost none in response.) I have no idea what "it" is here. Criticize what "it"? What "argumentation"? He gives no examples which fit such a complaint.

If he had only said "People shouldn't judge all Muslims by one bad example." Then I wouldn't have written a thing. I agree. I would still wonder what on earth he's talking about (since I don't see anything like that happening, on any meaningful scale). But he said nothing of the sort. What he SAID was that making any kind of generalization at all (with specific allusions to nationality and religion) was characteristic of racism.

Perhaps you can read his mind, or are otherwise quite sure that he meant something else. If so, more power to you! I honestly have no idea what he specifically meant, and all possible permutations I can see strike me utterly indefensible:

* I have trouble thinking he meant what you said: too many generalizations "across the spectrum": If so, why didn't he mention any other varieties, bipartisan-style? Why is he specifically fixated on "racism"? What "generalizations" is he seeing regarding Hasan? And, speaking of "across the spectrum", did he object to the much more frequent and prominent generalizations in the past? (Say, the Holocaust museum shooting be rapidly blamed on "the right wing" by the media — and something similar with the census worker, where we don't even know the killer, such that we could generalize about him.) And why spent time arguing Hasan couldn't have religious motivations? That's not an attack on generalizations, in general (heh), that's a worry about one very specific characteristic of one specific individual. (As if he were disputing whether his dirty friend were Dutch at all, rather than admitting that, and only attacking the generalization itself.)

* If he meant people shouldn't judge all Muslims by one example, then he again needs to explain who's judging, and what specific judgments are objectionable. I don't think he'd be able to produce any. In addition, he'd be guilty of incredibly poor writing (given his vast overstatements, and resort to fallacy), and note the sole "example" given doesn't seem to match his complaint.

* If he really meant what he actually said, then I stand by all the implications I've already noted. In addition, I'd ask why he didn't write a similar column against the much more common and egregious past generalizations which I also noted.

* If he meant something in between (generalizations on "race" (spelled r-e-l-i-g-i-o-n for him, apparently) are always wrong, but generalization on other criteria are fine, then I'd ask, as noted above, why the rules of logic should change from subject to subject.

I'm open to other options too, but I can't see any way in which his column makes a lick of sense, as written (which, again, is nothing like my example at the top of this section).

If I had to guess, it seems that he's frustrated that people are thinking Islam had ANYTHING to do with Mr. Hasan's murders (he quite clearly says the idea is nearly impossible to believe) and is not too subtly calling all of the morally inferior people who notice this "racists."

And, if my life depended upon a guess, to apply a stereotype, I would then guess the conclusion he really objects to is any implication that we have more to worry about from (say) Islamists, than from some other group or thing he really fears. Frankly, I suspect that — or something close to it (see next paragraph) — motives a lot of the "you're a racist!" and "don't generalize!" talk we hear today. The real enemy is Glenn Beck, or the polluters (or whoever, I don't know his specifics), and the pesky Islamists keep screwing up the narrative.

(Or perhaps he's just offended by the idea there are real threats in the world, and that we might have real enemies who aren't just "misunderstood".)

But he doesn't want to SAY his real belief, whatever it is, directly, so he's written a jumble of vague, self-contradicting platitudes and allusions — which are supposed to "teach" his less enlightened readers this, indirectly. People on the same wavelength (with the same fears or prejudices) WILL read his column as making perfect sense, because they share the same unevaluated misconceptions; the same mental shibboleths — a kind of Rorschach test where they can lazily read into it whatever they'd like to see, without noticing the discontinuities.

So that's my #1 theory. But (a) it's only a theory, and (b) I don't like to tar people with such base motives (even though I believe them to be common) so I stick to pointing out that what he HAS WRITTEN makes no sense at all. Which still wouldn't have been so bad, Ryan, if he didn't present it as a lesson in logic.

Do I think he's opposed to polling? Of course not. (Nor did I say he was.) I also suspect he's isn't actually opposed to generalizations, even of exact sorts he's cited. (I believe I also said this in the article, actually.) What I wrote is called reductio ad absurdum, and it's a valid form of argument.

When it comes to certain crucial core ideas, most people have no idea what's in their own mind. They won't admit, to themselves, what they believe. Hence they can't say it plainly — if inspected, it would be revealed as intellectually and/or morally bankrupt. But they have a huge emotional stake in whatever it is.

Some thoughts upon re-reading Mr. Steinberg:

The killings at Fort Hood might say something about the strain that overtaxed U.S. soldiers are under. It might say something about security on Army bases. But if you think that it says something about religion, what you're really doing is saying something about yourself, and it isn't something good.

So let me get this straight: Based on this one example, we can generalize about the frame of mind of US soldiers everywhere. We can extrapolate about Army security, in general. (Apparently it would be fine with him to imply the Army has poor base security because of a fluke inside job? Or did he mean we should extrapolate positively?)

Those forms of going from the "from the specific to the general" are just peachy! But try to do the same regarding any religious ideology involved? No, those other generalizations were acceptable, but not this one.

He's not against extrapolating. Even doing something as absurd as using a US-based psychologist to deduce the feelings of troops in Iraq. Not at all wrong. Just don't draw any inferences about religion. Then you're like the racists he just finished describing.

Or am I somehow being grossly unfair again? >;-)

Point taken, the guy does not even follow his own advice.

Point taken, the guy does not even follow his own advice.

Not to beat a dead horse, but I couldn't resist... Hmmm... a few other non-extrapolations by Mr. Steinberg...

On rural people and their intelligence (re: conspiracy theories):