Anecdotes and Arguments.
To move from one to many is to move too quickly: thus the fallacy of hasty generalization would be involved. If the subject of the anecdote were a well-known person who offers it as a testimonial, these logical problems would not disappear, even if her story were believed. The particularity of the story, though a basis for its interest and memorability, would make it inadequate evidence for a general claim.
Of course anecdotes are not only used in arguments. Many are recounted just because they are interesting or amusing. A favourite in my husband's family is the tale of a Dutch cousin who went to England to work as an au pair for a wealthy family. They gave her the use of a horse and taught her to ride. She recounted their generosity, enhanced in her view by the youth of the horse which, she proudly said, was only twenty years of age. No general point is suggested by this story as people tell it: it would be silly to find an argument in the case and criticize it, alleging logical defects. Anecdotes can serve to introduce discussions, to illustrate claims already established, and to suggest hypotheses for exploration. There are many possibilities of using anecdotes without committing logical errors. If a specific narrative is well documented and of medical interest, it may be recorded as a case and studied as a possible counterexample to established theories. But the standard logical point remains: by itself, an anecdote about a particular claim cannot provide good evidence for a general one.
Voltaire wrote about anecdotes in his Philosophical Dictionary (1764). He regarded them as typically unreliable in their details--careless talk of the town, gossip, stories full of unexamined falsehoods all too often repeated by historians or raconteurs who had not bothered to check the details. In his discussion, Voltaire offered many examples of narratives that he discounted as mere gossip. He criticized anecdotes for their lack of credibility, not on the grounds that inferences made from them were incorrect. Contemporary anecdotes may also contain false claims carelessly disseminated. For example, a person recounting her own experiences might believe that herbal treatment X relieved her knee pain, misinterpreting a daily variation that was actually due to weather or fluctuations of stress. But even if an anecdotal narrative were to be correct in all its details, the logical problems involved in inferring the general from the particular would remain.
Philosopher Louis Groarke (2001) once appealed to Aristotle's theory of knowledge to question contemporary logicians' dismissal of anecdotal arguments. Groarke referred to Aristotle's notion that if one's observation is "penetrating enough" one can grasp a universal through our perception of a particular; the mind will have intuitive powers making that knowledge possible (Posterior Analytics, 350 BCE). Without delving into Aristotle's theory of knowledge, it is rather hard to see how such intuitive powers might work. Presumably there would be a kind of insight provided by intuition. We may understand this idea for geometric objects: one circle is just like another in its essential features, so if a person grasps the relationship between radius and circumference in circle #1, he would grasp it for all circles. Between the particular premise and the general conclusion here, we may posit an unstated additional premise to the effect that all circles are the same with regard to their geometric features, and that premise will be true. Adding such a premise to the argument and applying generous background knowledge and contextual information to interpret perception of a single case, we could sympathize with a neo-Aristotelian view. But with background knowledge applied, considerable information would be added to the concrete anecdote.
If we could add similar additional premises to anecdotal arguments about human health, historical events or even physical circumstances such as the weather, and have good reason to think those added premises true, what started out as anecdotal arguments would become stronger. But at this point such arguments would no longer be anecdotal. Another problem is that in many cases the additional premises would be implausible at best. Even if a herbal treatment alleviates the pain of one person, it may not work for others, because human bodies and their sensitivities and circumstances vary considerably.
Recently several communication theorists have questioned logicians' negativity about anecdotal arguments, maintaining that such arguments have significant and legitimate rhetorical functions. Anecdotes are often vivid, interesting and memorable. They can display character; they are less 'dry' than statistics. Christopher Oldenburg and Michael Leff (2009) understand an anecdote as an individual narrative that may be used to invite the hearer or reader to grasp a whole. They support the idea that an anecdote can provide general insight, suggesting that it can occupy a space between narration and empirical generalization, or induction. (It is not clear where that space is.) They compare anecdotes to the device of synecdoche, a figure of speech in which a part represents the whole--as when we refer to 'bums in seats,' with bums representing persons or 'paying with plastic,' with plastic representing credit cards. But the reference to synecdoche is not helpful from the point of view of logic or evidence. A figure of speech is not an argument.
Can anecdotes make a general conclusion credible? Contrary to prevailing logical opinion, Oldenburg and Leff suggest that they can. But at this point we need to distinguish two meanings of "credibility." On one meaning a claim is credible if it is persuasive and likely to be believed. In that sense, Oldenburg and Leff are probably correct. Anecdotes are often persuasive; they do often serve to make general claims believable. In another sense, "credibility" means worthiness to be believed. This is the sense that's relevant to the merits of anecdotal arguments, and in this sense anecdotes do not make general claims credible.
The rhetorician may say to the logician that anecdotes can be highly persuasive; the logician will reply that that's precisely the problem.
Trudy Govier is a Canadian philosopher and Professor Emerita of the University of Lethbridge. Her many books and articles include A Practical Study of Argument, Forgiveness and Revenge, and Taking Wrongs Seriously.