Post Hoc Reasoning, Special Pleading, and Ad Hoc Hypotheses

The powers of the paranormal, if they exist, cannot be very great if they are so easily thwarted by mere doubt. It seems as though, in the world of supernatural claims, doubt is the strongest magic of all. It can cancel anything, except science, which actually needs it to work. At least, this is the impression I get from the claims of paranormal believers when attempts to replicate initially successful parapsychology studies fail. And fail they have once the controls of the initial study are improved, reducing statistical significance closer to chance levels and shrinking effect-size to zero.

It seems to me that even with perfect methodology there would still be a chance for false-positive results, and that what these studies show is not what they are claimed to — only that something other than chance may be at work, and giving no indication of what that may actually be. It could be due to poor experimental design, inappropriate use of statistics, errors in reasoning, bad data collection, and rarely, but often enough to taint the entire field of study, fraud.

One thing never fails, though, and that’s the rationalizations offered for this failure to replicate. This post deals with a species of error in reasoning: Special Pleading, the Post Hoc [after this, or “after the fact”]fallacy, or Ad Hoc [for this (only)]hypothesis, and sometimes just “covering your ass by making shit up.” I also aim to show that it is not always a fallacy under the right circumstances.

This fallacy, regardless of its name, is an attempt to rescue a claim from disproof by inventing special reasons why it should be exempt from the usual standards of evidence, to deflect criticism without demonstrating that these alleged reasons are in fact true or actually exist apart from the claim they attempt to defend. Every attempt has been made to boil the following examples of its use down to their essence and to avoid committing straw-persons:

Psi phenomena are shy, or jealous. They do not work in the presence of skeptics. Skeptical doubt cancels them.

What about this one?

Successful replications do occur, but the doubt of skeptics reading the journals they are published in reaches back through time, retroactively changing the experiment and causing it to fail.


Psi is elusive and a delicate phenomenon. Imposing excessively strict controls (read: adequate ones) in a study impedes Psi’s natural functioning in a sterile laboratory setting.

What I find interesting about this sort of reasoning in its fallacious form is that it is considered acceptable in some circles.

Never mind that many of the replications are attempted by other believers and by those without an apparent bias against the paranormal, and another such rationalization goes something like:

They (believers or neutral parties who don’t get results) are burdened with a repressed skepticism that causes their replication attempts to fail, no matter what belief or neutrality they claim to have. These hidden attitudes unconsciously sabotage their efforts.

Never mind the fact that this argument is made on the basis of mere supposition and absent the use of a valid psychological test. Those who reason thus are essentially claiming to be able to read minds, the very thing that some of these replication attempts have failed to demonstrate.

This phenomenon, the success of some to get positive results in their studies and others to get negative results based on their belief systems, is in parapsychology known variously as the Shyness Effect, the Wiseman Effect, or, in a form broadly applying to any field of science where attitudes may unconsciously influence results, the Experimenter Effect, or Observer-expectancy effect, and this is one of the reasons for double-blinding studies and other forms of experimental controls.

A good example was a series of medical studies for a procedure known as the portacaval shunt, and in the analysis of these studies, it was discovered that those who were more enthusiastic about the procedure tended to get false-positive results more often than those not so inclined. And this was from a study assessing an experimental surgical method, not magic mind-powers.

Above were some examples of this form of argument used as a fallacy, but are Ad Hoc hypotheses always and everywhere bad reasoning?

Fortunately, no.

This can be a perfectly good way of reasoning, as long as at least one of the following conditions is met:

  • The reason for failure to demonstrate something has already been shown, or can be, to exist independently of the hypothesis it is used to support. There must be valid evidence that it is true and relevant as a viable supporting reason.
  • The Ad Hoc hypothesis is both interesting and fruitful in predicting new phenomena that could in principle be tested even without being true or existing itself. The key point is that it must be testable, whether by verification or falsification if a general or a particular claim.
  • The Post Hoc reasoning is used to invent new and creative ways to test a claim, and as long as it is used to further inquiry and not merely to thwart the goal of critical reasoning by making up silly excuses as needed.

A good example of an Ad Hoc hypothesis that was both interesting and fruitful was Einstein’s addition to General relativity of the Cosmological Constant, which though he later rejected it and called it his “greatest blunder” has shown to be useful today in the concept of Dark Energy to explain the accelerating expansion of the universe. Another would be the the Lorentz contraction offered to explain the failure of the  Michelson-Morley experiment to detect the Earth’s motion through the Ether, later incorporated into Einstein’s Special relativity.

One thing to note about many forms of argument used as fallacies:

Philosophers and communications specialists may differ on this, but informal fallacies are not so much violations of argument form as they are violations of argument procedure, as attempts to subvert the rules and goals of constructive argument and critical discussion. In this sense, they are abused, often out of ignorance but sometimes out of intellectual dishonesty, as rhetorical devices masking themselves as cogent arguments when they are not. For ethical, productive argumentation, try to keep this in mind and avoid this yourself whenever possible. Happy debating.

5 thoughts on “Post Hoc Reasoning, Special Pleading, and Ad Hoc Hypotheses”

  1. There are things in this life that still need explanation. Paranormal events are not in that group. If it cannot be replicated or independently observed outside of the mind of the claimant, it does not exist. Any statement to the contrary is wrong. It’s that simple. For any given paranormal event there are dozens (if not more) of people or groups that have refuted them with science and logic/reason to make such events unbelievable and childish. There is not yet even one paranormal event which has not been shown to be a hoax or trick of the human mind.

    Any post hoc explanations are worthless. If you can’t repeat it under test then your first observation was a fluke or false. Edison is said to have tried 10,000+ times to get a workable light bulb. If your paranormal event is important enough to you, you still have 9,999 times to try it under test. What are you waiting for?

    Of course, I speak as though you or the reader are the claimants… it’s just easier to speak that way sometimes. Take no offense.


    1. No problem. I used psi-research as an example even though I think that the phenomenon probably doesn’t exist. Probably, since it’s not logically possible to definitively prove a universal negative using a finite data set. Parapsychology is the one area of research where that kind of argument is considered acceptable, though not to its critics and for good reasons, so was perfect to use for the sample arguments given. It’s no coincidence that parapsychology’s most vocal critics are those who understand the mind best — most psychologists and neuroscientists — so it’s unlikely in the extreme the critics will ever be proven wrong.


      1. This is something that I ‘discuss’ with my partner. They have a strong desire to believe in something, a strong deist tendency. I have to show them repeatedly why a given tactic or phenomenon is false. I’ve had to explain to a coworker why martial arts fighters who claim to use energy to knock out their opponents is fake. Simply enough, if it were real, the US Army would be using it in training. They aren’t. It almost always dumbfounds me what people are willing to believe in.


  2. i see the ‘good’ uses of post hoc and ad hoc as meaning ‘this experimental result is not strong enough bayesian evidence to outwieigh the prior, but it raises the likelihood of this modified version of the hypothesis.’ i also think that using that procedure if you’re not einstein to be like adding an improv session to a rachmaninoff piece.


Comments are closed.