Weak Analogy

50 Weak Analogy

The eye appears to have been designed; no designer of telescopes could have done better. How could this marvelous instrument have evolved by chance, through a succession of random events? Many people in Darwin’s day agreed with theologian William Paley, who commented, “There cannot be a design without a designer.”

— Robert Jastrow (in Life—How Did It Get Here?)

An analogy compares things to explain, illustrate or support a claim. It is weak when the shared features do not provide enough support for the conclusion. The issue is the relevance of those features and the effect of important differences, rather than which list is longer. The fallacy is also called bad analogy, false analogy, faulty analogy or argument from spurious similarity.

Arguments from analogy typically have this structure:

1. Thing X is relevantly similar to thing Y.
2. X has a property P.
3. Therefore, Y probably also has P.

Evaluating premise 1 requires asking whether the shared properties support transferring P to Y. A difference can weaken that inference even when the two things resemble each other in many other respects. As Patrick Hurley writes, the weak analogy fallacy “occurs when the conditions of an argument depend on an analogy (or similarity) that is not strong enough to support the conclusion.”

Labels such as “correct” or “incorrect” can obscure differences in an analogy’s strength. Ask what conclusion the comparison supports and how much support it supplies. Persuasiveness alone does not settle either question.

In the opening quotation, Paley’s design argument (often expressed through the watchmaker analogy) contends that the complexity of the universe—like the complexity of a watch—implies a designer. A watch is a familiar human artifact, but inferring the origin of a universe from the manufacture of a watch requires a further argument. Biological evolution addresses the development of living organisms, rather than the origin of the universe as a whole. The analogy overstates the similarity: we have direct experience of watches being made, but no comparable experience of universes being made, and the very notion of “design” may not transfer. Hume noted that our limited exposure to an orderly corner of the cosmos does not license global conclusions about a designer. Another variant compares a rose to a watch, suggesting both “appear designed.” Yet a rose is a living organism that grows and reproduces through natural processes, while a watch is an artificial mechanical device with a known human creator. The analogy founders because of these fundamental differences.

Weak analogies can surface in many everyday arguments, often where the two things compared share only a superficial resemblance. For instance:

• “Renting a house, rather than buying, is like flushing your money down the toilet.” In reality, renting is paying for the temporary use of a property, much like renting a car or a movie. The money is exchanged for a lived-in home, not simply destroyed. The analogy ignores the essential nature of the transaction.
• “Denying the existence of God is like refusing to believe that the sky is blue.” The sky’s blueness is a publicly observable, empirically verifiable fact; belief in God involves questions of faith, revelation, and metaphysical interpretation. The difference in the type of claim makes the comparison unhelpful and misleading.
• “Recycling is like putting a band-aid on a broken arm.” Recycling is a partial solution to environmental degradation, while a band-aid on a broken arm is a completely inadequate response to a severe injury. The analogy is weak if it is used to dismiss recycling’s benefits without examining them. If the claim is merely that recycling alone is insufficient, that is a different conclusion.

Another common weak analogy appears in defenses of meat-eating: “Animals eat each other in nature, so it’s permissible for us to eat them as well.” The structure runs:

1. Animals eat each other in nature (X).
2. Humans eating animals (Y) is analogous.
3. It is permissible for animals to eat each other (P from X).
4. Therefore, it is permissible for humans to eat animals (P from Y).

Both acts involve eating other creatures, but a moral argument must explain why that similarity establishes permission. Human choices, available alternatives and the conditions of farming differ from predation. Animal behavior is not sufficient moral guidance: male lions kill a predecessor’s cubs, but that does not justify human stepfathers killing stepchildren. The comparison alone cannot settle the ethics of eating animals.

Not all analogies are weak, and many can be deployed successfully. Judith Jarvis Thomson’s famous violinist thought experiment illustrates this. Imagine you are kidnapped and a renowned violinist is plugged into your body to survive for nine months. Thomson compares this to an unwanted pregnancy. There are both parallels and divergences; the debate centers on whether the relevant moral features align closely enough to justify the conclusion that a woman is not obliged to sustain a fetus. The strength of the analogy depends on one’s assessment of those shared moral dimensions.

Similarly, Peter Singer uses analogy to argue that animals feel pain. He points out that nearly all external signs of pain in humans—writhing, moaning, avoidance behavior, fear—are present in other mammals and birds. Their nervous systems are structurally and functionally like ours, responding with the same physiological changes (blood pressure, dilated pupils, etc.). These nervous systems evolved along a shared ancestral path, not as artificial mimics. Since the features that justify our belief in other humans’ pain are also present in these animals, the analogy is strong—just as with human infants or severely cognitively impaired humans, the lack of sophisticated rationality is irrelevant to the capacity for pain.

Using analogies effectively. Identify the property being transferred and explain why the shared features support it. Then test differences that could defeat the inference. A comparison useful for illustration may still be too weak to establish a conclusion.