44

I have seen a number of instances of the following pattern.

  • Someone says something that absolutely explicitly states X and also clearly implies Y.
    • Y, of course, is more outrageous than X. In typical cases, X is true but not terribly surprising or interesting, whereas Y is false but would be surprising and/or shocking if true.
    • Y is plainly the point the person is hoping their readers/hearers will take away.
  • A reader of skeptics.se sees this and posts a question asking "Is this really true?".
    • In some cases, the ambiguity is noticed and dealt with by editing the question title to be about X, while the question body still contains material that implies, insinuates or presupposes Y.
  • Someone posts an answer saying "Yes, X is true." which addresses Y not at all or only very briefly.
  • This gets accepted (because X is in fact true and is the only perfectly explicit claim in the question).
  • But it is still terribly easy for a reader to think that the good people of skeptics.se have endorsed not only X but Y.

Here are a couple of examples.

Was a class of Dutch school children required to learn Muslim prayer? -- X is "A class of Dutch pupils were taught about how Muslims pray" and Y is something like "A class of Dutch pupils were indoctrinated with Islamic ideas and induced to do actual reverence to Allah". There is exactly one answer, just saying that X is true.

Did the US national debt fall by $100 billion in the first two months of Trump's office? -- X is "The US national debt did such-and-such in the first two months of Obama's presidency and such-and-such in the first two months of Trump's" and Y (as, actually, stated explicitly in the material quoted in the question) is "Obama caused the national debt to do this, while Trump caused it to do that"); perhaps there's also a Z which is something like "... and this shows that Trump is better at handling the economy". The question title was originally about Y but it was edited to say X instead. There are two answers, the higher-voted of which gives a nice clear endorsement of X; it briefly mentions that Y is different, but briefly enough that (as you can see from the comments) at least one reader didn't see it.

Is President Trump right that there was violence on “both sides” in Charlottesville? -- X is "on both sides, at least some people committed at least some violent acts", Y is "both sides have comparable liability".

In both of these cases, the only answer (first case) or highest-voted answer (second case) addresses X nicely but neglects Y.

For the avoidance of doubt, this doesn't always happen: for instance, see this question about Obama "admitting training IS" both of whose answers make it very clear that (1) yes, he said what he's claimed to have said (X) and (2) no, he didn't mean what that sounds like (Y).


When addressing a claim that implies much more than it undeniably states, is dealing clearly with the implications an important part of a good answer?

I think it is. More specifically, I suggest that

  • a question about such a claim should make it clear and explicit in both question title and question body whether or not it's concerned with the implications;
  • in either case it should make the difference between the implicit and explicit claims as clear as possible;
  • usually the more-dramatic less-explicit claim is what's actually of interest;
  • an answer that isn't absolutely clear about which it's addressing is not doing its job;
  • a really good answer will deal with both implicit and explicit claims and contrast them as appropriate.

This seems like something that would have come up before in meta, but I haven't found previous examples. What there are plenty of is earlier questions about whether it's OK to ask about implicit claims: one, two, three, four, but that's not the same thing.

  • 1
    What is wrong about finding a better version of the claim which is more explicit? – Sklivvz Mar 22 '17 at 16:49
  • 1
    That would be great, where it's possible. But of course what often actually happens is that a claim is (or is known to be) notable because some particular person made it on some particular occasion, and questioners here have to work with what they have. – Gareth McCaughan Mar 22 '17 at 16:50
  • Unless there's some evidence that a claim is notable, we don't allow it. If a notable person stops short of saying something, then they have not said it, thus there's no notable implicit claim. – Sklivvz Mar 22 '17 at 16:51
  • 1
    What is wrong about finding a better version of the claim @Sklivvz I think this meta-topic is about how to answer claims, not about how to question them. The users who answer aren't able to choose the claim. For example, [this topic] has an implicit claim, about the link with the euro (though in this example the OP made it clear they weren't interested in that implicit claim). There are other "better versions" of the claim where Le Pen does talk explicitly about the euro, but they're not the OP question's and it's not for us to substitute the claims that we prefer instead of the OP's choice. – ChrisW Mar 23 '17 at 13:30
  • @ChrisW "The users who answer aren't able to choose the claim" -> therefore, unless we change the question, they should not answer about an implicit claim, since questions are always about explicit claims. – Sklivvz Mar 23 '17 at 13:32
  • "one" and "two" link to the same post. – Andrew Grimm Mar 25 '17 at 0:55
  • So they do. They weren't meant to, of course but I unfortunately no longer have the list of questions I was intending to link to. – Gareth McCaughan Mar 25 '17 at 1:12
  • 4
    The claims are often intentionally structured the way they are to specifically imply what they don't say. It's a form of lying. What's wrong with dealing with what the claim is making? The Trump/Obama question clearly was making an attribution, so what's wrong with an answer that says, "yes, X is true, and this is why it does not logically follow that Y is true."? It deals with the explicit claim, and then goes further. I don't think "skeptic" = "obtuse," though the moderators here seem to demand that we behave as if it does. – PoloHoleSet Apr 3 '17 at 13:42
  • 1
    Just make sure in the Y claim you say something discouraging about the president, that will surely push your answer over the line. – daniel Aug 18 '17 at 21:16
  • 1
  • I think the "violence on both sides" question is a bit of a shit storm. Not a good example. – fredsbend Sep 11 '17 at 16:33
  • The "violence on both sides" question, at least as it stands right now, is not an example of the failure mode I was asking about, because the answers generally address both X and Y. It is an example of a question where X and Y are different and where addressing only X could be misleading. Whether it's good to have it here, I don't know. (It wasn't part of what I originally wrote here but was added later.) – Gareth McCaughan Sep 11 '17 at 16:42
  • Don't forget to add Are women physically weaker?, where the subtle claim is "women are inferior" and the answer explains clearly that "many women are actually stronger than some men". It is very improtant to help people understand that we are all pretty much the same, men women, muslims christians jews arabs, europeans and US citizens. – Maria check profile Sep 12 '17 at 16:15
  • 2
    @Maria We don't care about your political opinions. This site is not meant to support anything other than factual analysis and verification. "It is very improtant to help people understand ..." Not on this site. "... isolating right idiologies in a stealthy way." That's way outside this site's purview. – fredsbend Sep 14 '17 at 12:45
  • 1
    @Sklivvz - This -- therefore, unless we change the question, they should not answer about an implicit claim, since questions are always about explicit claims is why I participate so rarely at this corner of the SE network. Fox News and MSNBC oftentimes are more open and truthful than is skeptics.SE. – David Hammen May 19 '18 at 1:53
29

"Is dealing clearly with the implications an important part of a good answer?"

I think it is. Another example is, some question are of the form "Did so-and-so really say this?" It's on topic to clarify the context in which it was said, i.e. what the person meant when they said that.


Also I think that the current answer to Was a class of Dutch school children required to learn Muslim prayer? could be improved at least slightly.

The title says, explicitly, "required" and "learn Muslim prayer": and I think that makes at least those aspects of "Y" specifically on-topic. Given that I might be inclined to clarify that:

  • It wasn't required
  • They were learning something slightly other than prayer: e.g. that this was an opportunity to learn Muslim "culture" and "language" and so on (including "how Muslims pray"); an opportunity to meet someone (an Imam) in person; to see inside a mosque, and so on.
  • And (at risk of inviting political controversy), given that this visit happened in Holland, they were learning more about (present-day) Dutch culture and society.
  • 1
    Define an objectively clear implicit claim. Unless it's a logical necessity, I don't see how that is possible. – Sklivvz Mar 23 '17 at 22:39
  • 2
    I doubt I understand your question. But for example in the quote about Horace, the implicit claim is that Horace wrote that making money is a good thing; similarly in the claim about Pascal. In the story about Dutch school children there is IMO an implicit claim about coercive Muslims and/or left-wing do-gooders, forcing children, forced conversion, and (more implicitly) never forget the newspaper horror stories about ISIS, this could begin to happen here. So better to put the story in context. – ChrisW Mar 24 '17 at 7:20
  • 1
    The first two examples are explicit claims and not implicit, "Did Horace write X?" is explicitly "Did Horace write X?". I just don't understand your example. In the story about Dutch children, you have literally just repeated the factual question, but loaded with political innuendo. It's not examining an "implicit" claim, it's taking a factual question and adding political partisanship, prejudice and non factual opinions. It's actually my job as a moderator and your job as a high-rep user to defuse such behavior as you see it happen, so why would you think it's a good thing? – Sklivvz Mar 24 '17 at 8:48
  • 1
    taking a factual question and adding political partisanship I thought you asked me to, and so I gave an example (exaggerating the political partisanship for the sake of clarity or demonstration). I don't see it as me adding the partisanship, I see it as inherent in the claim (and part of the intent of whoever published it); and saying that prayer wasn't "required" is explicitly on-topic as well as defusing the implicit claim. This (Dutch example) is a mild example, and boundary case -- the answer is OK without it, but it could add that, and I might have mentioned it if it had been my answer. – ChrisW Mar 24 '17 at 8:59
  • 1
    I thought you asked me to. What I asked you was to define an objectively clear implicit claim. The example you provided was a subjective, politically loaded version of a claim that added nothing of value (IMO) to the original, which is exactly the point of why we don't allow implicit claims. – Sklivvz Mar 24 '17 at 9:03
  • 2
    I just don't understand your example Yes the question was, explicitly, "Did Horace write that?". It's implicitly, "Did Horace believe what he said", did Horace say that making money is good, if we want to be like the virtuous Romans (like Horace) should we too proclaim this as virtuous. To which my answer to the explicit claim is Yes and my answer to the implicit claim was No. – ChrisW Mar 24 '17 at 9:03
  • 1
    Surely you are not suggesting we allow implicit questions on beliefs and motivations since we don't allow questions on motivations in the first place. – Sklivvz Mar 24 '17 at 9:07
  • 8
    I agree there's good reason to forbid questions which are explicitly about motive and intent. On the subject of propaganda, I think that propaganda sometimes states untrue facts/claims in order to promote a lie: in which case it's sufficient to "debunk" the lie (the claim). Sometimes though I think that propaganda publishes a true fact in order to imply a falsehood. I think you're saying it's sufficient to confirm the true fact, and the OP thinks we should (or can) also show that the implication isn't sound. I think it's permitted to touch on the implied claim in an answer, iff your answer ... – ChrisW Mar 24 '17 at 9:20
  • ... addresses the explicit claim. – ChrisW Mar 24 '17 at 9:20
  • Let us continue this discussion in chat. – Sklivvz Mar 24 '17 at 11:39
2

Implicit claims are not allowed here so it's important to look at what is actually being claimed.

It's OK to add a commentary as to what one perceives to be implied as long as

  1. It's a minor part of the answer
  2. It's clearly stating that this is the interpretation of the answerer, not the claim itself
  3. It's not answering an obvious straw man
  4. It's not non-answering by nitpicking on a definition
  5. It's not a discussion about intent, motivation or beliefs of the claimant

But please, don't simply answer to your perceived notion of what's being implied and stick to the actual claim being asked about.


Longer explanation of the points

It's a minor part of the answer

Questions and answers should be about explicit claims and explicit evidence. We have allowed a little bit of leeway in both to allow some opinion through, and an implicit claim is always the poster's opinion of what is implied.

It's clearly stating that this is the interpretation of the answerer, not the claim itself

There's nothing wrong with being imprecise or having opinions—as long as we are intellectually honest about them. If you have a personal opinion, or no evidence, say so. Don't pass it as a fact.

It's not answering an obvious straw man

The basic reason why we don't allow implicit claims is that often they are straw men, not-quite-correctly-reported claims that are easy to disprove. Don't do that on this site.

It's not non-answering by nitpicking on a definition

Similar to the straw man point above, answers need to try to be convincing to someone in doubt about the question, not simply "preach to the converted". Arguing a claim is false because it's only literally false, but not substantially false can be a case of an implied claim we don't allow.

It's not a discussion about intent, motivation or beliefs of the claimant

These discussions are never truly evidence based, because direct evidence of beliefs is impossible, you can only rely on hearsay or similar subjective evidence. We don't allow it in questions, and we equally don't allow it in answers.


Update: Here's an example of a question with an implicit claim that has been asked and answered according to good practices:

Have there been more than 27,000 Islamic terrorist attacks since 9/11?

The question sticks to the explicit claims, the answers analyze the sources, find the numbers plausible but also give context by saying it's misleading and providing some evidence of that. There are no answers merely focusing on the perceived claim, nor any straw man attacks. Reading the answers would plausibly convince an undecided person of the facts.

  • Comments are not for extended discussion; this conversation has been moved to chat. – Sklivvz Sep 15 '17 at 6:27
2

Note: This was originally an answer to this question.

I would suggest that, if there isn't an absolutely definitive interpretation of a claim, then some leeway should be given for answers that address the claim in ways that provide further information, rather than having strict adherence to the literal claim.

This shouldn't be a carte blanche to post any possible interpretation, but if an answer provides information in response to a common or likely interpretation of the claim (or an interpretation specifically given by the questioner), and makes clear that it is a matter of interpretation that distinguishes it from another answer, then this should be permitted.

Establishing such a policy would have two key benefits:

  • Where a questioner has demonstrated an interpretation that is at odds with the most obvious, intended meaning of the claim, it would allow the answers, in combination, to be of greater benefit to the questioner and others who have the same interpretation.

  • It would ensure that claims that fall into the "true but misleading" category (such as the Hillary Clinton flag-burning example above) receive not only the "why it is true", but also the "why the claim isn't what it seems" - an equally valid part of skepticism. This ensures that answers on this site do not result in propagation of falsehoods hiding within technicalities.

Also note that it would provide consistency, as the examples (in the question) demonstrate that such information, contained within the same answer as the literal interpretation, are not considered to be an issue, and can even be the critical part of an answer... and it is often seen as rude on stackexchange sites to post practically the same answer as someone else along with extra information rather than crediting the original answer.

This policy would cause some slight overlap, of course, with other SE sites - this should not be considered a weakness, as it is not unusual for there to be some crossover between sites (for example, one wouldn't be surprised to find a mathematics question and answer regarding equations found in physics on the physics site, or on the mathematics site).

Context matters - if the notable claim was "the sky is blue", then most would say "true". As a result, if someone asks a question about whether the sky is blue, an answer may show up providing references confirming that the sky is, indeed, blue (and likely explaining why).

But an observant person may notice that the question was asking about whether the sky was blue and noting that every time they look out of their window, the sky is red. In this situation, the question has a clearly different meaning - "I've always heard that the sky is blue, but my experience is otherwise" - and the correct answer is "it's blue most of the time, but under some situations it is red, such as during certain weather conditions and during a sunrise or sunset".

Similarly, someone might be questioning "1+1=2". The obvious answer is "yes, 1+1=2, and here's a source confirming it". But there are mathematical situation in which that equation is false, and this may be revealed by the wording of the question as the relevant context. This, again, changes the meaning.

An actual example has arisen here, where the questioner is asking about "profit per customer", and the claim provides a value that is clearly "net profit per customer". However, the question itself is worded in a way that indicates that the questioner has interpreted it as "profit gained per ticket sale", which is a different thing, and has a different value. Both interpretations are valid, with the more common interpretation being the one used in the claim to get its value, but the second interpretation being the more useful value, and the more likely one for a reader to have when reading the claim, as demonstrated by the wording of the question itself.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .