The accuracy clients report is produced by well-understood features of the reading itself

The argument

Those who hold that fortune-telling does not work argue that the reported accuracy has an explanation which does not require the future to be legible. A reading is built largely from statements that fit almost anyone: that the client is more sensitive than others realise, that there is an unresolved matter with an older relative, that a change is coming which they have been putting off. Listeners supply the particulars themselves, matching the statement to their own circumstances, and remember the fit as the reader's achievement. Misses disappear in the same movement, because a statement that fails to land is simply passed over and is not what anyone recalls afterwards. The second mechanism is information the reader actually has. Appearance, clothing, accent, age, a wedding ring, the state of the hands, whatever the client volunteers while answering, and which guesses produce a flicker of recognition all narrow the field quickly, and skilled readers phrase guesses as questions and keep what works. Proponents of this position argue that this is a genuine and difficult skill, which is why good readers seem uncanny, and that it is a skill in reading people rather than in reading the future. The decisive point, on this view, is what happens when those channels are closed. Where readings are produced without the client present and later matched to clients by independent judges, or where feedback during the reading is withheld, the accuracy clients report in ordinary conditions does not appear. Proponents argue that a result surviving only in the conditions that permit the alternative explanations is not evidence for the claim being made.

Premises

Counter-arguments

Those who defend the practice answer that the explanation is incomplete. Cold reading accounts for the general statements, they argue, but not for the specific and apparently unguessable details clients consistently report, which the mechanism described has no way of producing. They also question the tests: a reading is an exchange in which the reader works from the person in front of them, so removing the client removes the practice rather than testing it, and judging transcripts matched blind measures something nobody claims to do. On this account the controlled setting does not reproduce the conditions under which readings are said to work.

Rejecting the premises

[Rejecting P1] General statements account for part of a reading but not for the specific and unguessable details clients report, which this mechanism has no way of generating. [Rejecting P2] Treating skilled attention to a person as mere guesswork begs the question, since reading someone accurately may be how the information arrives rather than a substitute for it. [Rejecting P3] Removing the client, or withholding their responses, removes the exchange the practice consists of, so the test may be measuring the absence of the practice rather than its failure. [Rejecting C] The premises show that much reported accuracy can be explained without foresight, which is weaker than the conclusion that the practice does not work.