Back to Silas S. Brown's home page

The ``go and argue with them'' fallacy

``The accordion is an inferior instrument [1], a fact recorded in the world's most authoritative English dictionary. [5] Accordionists take people's money [3] and sell music for far more than its 4p-a-sheet value. [4] They even threaten national security by causing bomb scares [2] that waste police time. I propose all accordionists imprisoned and all humanity be denied that so-called `music' forever.''---Mad advisor to John Q. Dictator (attrib).
[1] The Oxford English Dictionary, 2nd edition, definition of ``Concertina''. ``A portable musical instrument ... Often improperly applied to inferior instruments ... [such] as the accordion...''
[2] Accordion Causes Bomb Scare. Music Trades September 2012, Vol. 160 Issue 8, page 32
[3] Fiddle and Accordion Society membership fee [4] Theo Wyatt's Oriel Library standard price (1990s) [5] OED advertisement

Once upon a time a cruel baron wrote a scholarly-looking article, complete with impressive-looking references that didn't fully support the claims, and perhaps even managed to trick a respected source with an otherwise-good track record into buying the argument. Soon, all fans of the baron's cruelty were citing it in public debates and using it to humiliate anybody they didn't like. If anyone dared question it, its supporters would simply point to all the impressive-looking sources that had run it and say, ``take it up with them''.

In today's well-connected world, where messages can be sent to practically anyone and online sources are continually revised and updated, it's too easy to rely on an information equivalent of the Efficient Market Theory (EMT) in Economics, and believe "any correction that needs to be done will have been done already". As with the joke about the EMT believer failing to pick up dropped money "because, if it were real, someone would have picked it up already", the problem here is confusing "high likelihood" with "absolute certainty"; if everyone behaved like this then no money would be picked up, no markets would operate and no corrections would be made. Of course there are situations where EMT-like theories work most of the time, and it is valid to reason that they are probably working, but we cannot ignore that there must be exceptions, which we might well encounter at least occasionally. Indeed, as people lose their skepticism of the system, it's likely that problems will go uncorrected for longer.

There are many reasons why sources might not accept good corrections:

  1. A source might have an ulterior motive (like the cruel baron) and give the illusion of being unassailable by deleting any counterargument that's too strong for them to answer.
  2. A source might be too frightened to correct itself, for example if it is a bureaucratic organisation whose workers are afraid of anything that might draw negative attention, and fears that the bad PR of having to make a correction would outweigh that of continuing to go uncorrected.
    • Another possible fear is "if we update that, we might have to update everything else, and that's too much work." Putting a new date on something might be taken to imply you've re-checked all of it (unless you want to say otherwise), and it can be easier to prefer letting the previous date speak for itself.
  3. A source might not have the necessary resources to make the correction, for example because its internal publications-handling system is very complex, or it lacks sufficient resources to separate good feedback from bad.
    • Most sources that accept feedback make at least some attempt to filter it (whether by their staff or by the community), but all filtering can be flawed and something brilliant might have been thrown out by mistake.
  4. The best possible corrections might not be sent at all, because people with the best counterarguments might consider it a waste of time to participate in such a volatile environment.
    • Search engines are also getting less and less good at finding the best counterarguments to popular misconceptions. Those who rely on search to find everything might be misleading themselves.

"Appeal to authority" has been around for a long time, and in some cases it does have some merit: if an authority is known for having been right before then it is reasonable to assign it higher odds, as long as you remember that this is only a first approximation and can be overridden if any more concrete reasoning exists. Additionally, if someone says "I like person P's argument" this might not even mean they treat P as an authority; it might simply be a time-saving "re-use" operation (like a #include directive)---if the argument would work just as well with P's name removed and the text plagiarised, then P is not really being used as an authority here. However, people who take either approach cannot expect a guaranteed "win", because even the best authorities can and do go wrong sometimes.

What seems to be new in the modern connected world is the idea of suppressing counterarguments by telling your opposers to go and argue directly with the authorities you are using. Use of this fallacy may have unintended consequences for both sides:

One mark of professionalism is knowing when to walk away from a degenerating discussion even though you could have "won" it. Another is to recognise when someone else does this. But making a "parting shot" by pointing to one's favourite authority and telling the opponent to ``take it up with them'' is usually fallacious and potentially dangerous.

(Related fallacy: "The critic must be right because the other side hasn't sued them for libel" or even "because the other side sued and lost". But lack of a libel judgement is not a proof beyond reasonable doubt that the "libel" was right---it only means they didn't disprove it beyond reasonable doubt, which is a very low standard of proof. Additionally, some courts acquit people from libel if they can show they were taken in by an impressive-looking reference, in which case the court won't even consider if it's true or not, only if it looks good enough to get the person who repeated it 'off the hook'. Knowing this, most organisations don't even bother to file libel cases, so we certainly shouldn't expect an "efficient courtroom theory" to ensure all statements are accurate.)

As for using references, the practice is not always bad as there are many respectable sources in the world that do handle matters professionally (whatever view they have), but there are also those that do not, and even those that do can slip up, so caution is always needed no matter how "good" the source is. Incidentally, original research by yourself is allowed in life (WP:NOR is not a life guide); if done honestly, it may offer some protection against being "tossed to and fro and carried about" by faulty sources. It might even help you get more out of works that are not currently faulting.

All material © Silas S. Brown unless otherwise stated.