Abductive reasoning

Is abductive reasoning fallacious? Why or why not? And how do you use it properly?

Perhaps it might help to compare and contrast the deductive, inductive, and abductive processes.

Deductive reasoning works from a general rule or a statement about all things in general and then works its way to a conclusion about a specific case.  For example, (1) the prize goes to the one who gets the most "likes"; (2) Mary got more "likes" than anyone else; (C) therefore, Mary gets the prize.   If all of the premises are true, the conclusion must be true .  If the prize really does go to the one with the most "likes" and if Mary really did get the most "likes", then Mary must get the prize.  

Inductive reasoning starts from observations of specific examples.  Based on the available examples, one reaches a conclusion about things in general.  It is often described as a sequence involving: evidence gathering, looking for patterns, and then coming up with a theory or hypothesis that explains what has been observed.  The more observations available, the more likely it is that the conclusion is correct so the inductive process usually seeks to have as many observations as possible.  For example, (1) a new railway track has been built through an agricultural area and a large team of observers is stationed for miles and miles along the new track, (2) when a train passes by, sheep in the fields run from the side of the field where the railway tracks are, (3) when a train passes by, cattle in the fields run from the side of the field where the tracks are, (4) when a train passes by, horses in the fields run from the side of the field where the tracks are.  One observer concludes that animals are fearful of a fast moving objects; another observer concludes that animals are fearful of unfamiliar noises; another observer concludes that animals are fearful of ground vibrations.  Any of those conclusions (or other ones, too) is certainly possible, but no individual conclusion is necessarily the correct explanation.  Inductive reasoning or arguments can be reasonable or cogent, but they don't absolutely prove a point.

Abductive reasoning shares some commonalities with inductive reasoning but often tends to be more limited in terms of the number of observations.  Rather than starting from a comprehensive set of observations, often abductive reasoning involves a less complete set of observations.  This approach can lead to a reasonably probably conclusion, but the best we can say is that the conclusion is the most likely one.

A licence issuing officer uses deductive reasoning to decide whether to issue a licence.  The officer works from the general statement of what is required for a licence, then looks at a specific application to match the applicant's qualifications with the requirements, and decide if a licence can be issued.  The officer either issues the licence (because all of the conditions have been met, i.e., all are true) or refuses to issue (because not all of the conditions are met, i.e., some are false).  This process is fairly cut and dried and leads to a consistent (valid and true) conclusion.

If we were to believe what some of the TV medical dramas portray, we see large medical teams of many doctors with seemingly unlimited medical resources using something that approaches inductive reasoning when they conduct test after test after test after test (for the whole hour of the show!) so that they have many, many observations [not particularly accurate in the real world, but it seems to work magic on TV].  Then, based on the massive number of findings, they make the magic diagnosis and save the patient.  While it's not reality, this fictional TV approach (with its nearly unlimited observations) does share characteristics with inductive reasoning ... although I suspect most medical diagnoses in the real world are based on far fewer tests and observations so they actually end up being more like the next example.

By contrast, a first responder who comes upon an unconscious person at an accident scene, is likely to make some quick observations and decide on a course of action based on the most likely explanation for the accident and the most likely reason the person is unconscious.  (There's fresh paint on a high part of a wall, a damaged ladder on the floor, a paint pail on its side with a pool of paint beside it, and the unconscious victim on the floor near by.) Here we have abductive reasoning at play whenin the first responder probably treats for person for head, neck and back injuries associated with a likely fall.  (While it is certainly possible that a deranged madman across the street with blowgun and a poison dart decided to murder our victim, the fall probably seems more likely!)

So, back to your question ... is abductive reasoning fallacious?  The short answer is, "No ... the reasoning isn't fallacious."  However, we need to keep in mind that the concept of logical fallacy speaks to the correctness of the reasoning process and not the the truth of the conclusion.  As we've seen, valid arguments can lead to false conclusions (when one or more of the premises is not true).  If you're asking if it make sense to use abductive reasoning in certain situations, the answer is "Certainly ... when you have limited information or observations available and a decision needs to be made quickly" (as with the paramedic example above).  If you're asking if conclusions from abductive reasoning are guaranteed to be true, the answer is "No".  Of the three approaches (deductive, inductive, and abductive), the only one that guarantees a true conclusion is the deductive approach, and then only when all of the premises are true.

In abduction, one makes a set of observations and then infers simplest (most likely) explanation from these.

For instance, if I observe B, I might abduce A as a cause, knowing that A can cause B. Formally that's written as 'A implies B. B, therefore A.'

Unfortunately, this is a form of affirming the consequent.

However, abduction is useful in that it can help us disregard some explanations for an event that are not likely. One just needs to bear in mind that even if A is the simplest or most likely explanation for B, it is not certain that it is the correct one.

Abductive reasoning is reasoning we all do all the time.  We look at a situation and hypothesize about the most likely explanation.   If there's mud on the floor we assume that someone with mud on their feet had just walked there.   In developing scientific theories, we ask "what is the most probable explanation for what we observe?"   This is abduction.  Scientists then go on to try to confirm or disconfirm the theory using inductive (probabilistic) reasoning. 

Abductive reasoning is not fallacious, unless one is claiming that the conclusion "must" be true.   In good Abductive reasoning, the conclusion is "that's our best theory based on what we know."    In a strong inductive argument, the conclusion is probably true.   In a valid deductive argument (and only deductive arguments), the conclusion necessarily follows.    To claim that an abductive hypothesis must be true is to commit what might be called the Abductive Fallacy, which is a version of the Affirming the Consequent Fallacy.  

By the way, in Sherlock Holmes novels it was said the Sherlock was a master at deduction.  In fact, it was abduction he was a master at.