Law of Similarity

Two Versions of the Law of Similarity

Like causes Like

As described by Tylor (1879), Frazer (1895), and Mauss (1902), the law of similarity describes a belief that causes resemble their effects. Applied to treatment, this reasoning suggests that a remedy should resemble a disease or its presumed cause. For instance, the Zande people of Africa treat ringworm with fowl excrement because the two are similar in appearance. Homeopathy’s principle of like cures like is another use of resemblance as a guide to treatment.

Shweder (1977) argued that the conflation of likeness with co-occurrence likelihood is common not only in magical thought but also in mundane intellectual activities, concluding that “magical thinking is an expression of a universal disinclination of normal adults to draw correlational lessons from their experience, coupled with a universal inclination to seek symbolic and meaningful connections (likenesses) among objects and events.” This inclination is closely related to the representativeness heuristic (Kahneman & Tversky, 1972), whereby an entity is assigned to a category based on how closely its features match those of typical category members. When the category is cause–effect pairs, representativeness becomes “like causes like”: people expect a cause to mirror its effects. For example, because AIDS is lethal and extremely difficult to cure, many assume the causative virus (HIV) must itself be potent and indestructible. The severity of a disease does not, by itself, establish how readily its causative agent survives or spreads.

Similar thinking underlies voodoo practices, where harming an effigy is believed to harm the person it represents. Laboratory studies have examined reactions to such symbolic acts. Undergraduates throwing darts at a picture of John F. Kennedy’s face (aiming between the eyes) showed poorer accuracy than when throwing at a plain target or at a picture of Adolf Hitler. Moreover, they sometimes visibly cringed when a dart struck near the Kennedy target. Likewise, people commonly report reluctance to throw out or tear up duplicate photographs of loved ones—a phenomenon confirmed in unpublished studies. These reactions can feel as though an action on an image affects the person represented, even when participants do not literally believe that it does.

Appearance Equals Reality

A second, even more basic realization of the law of similarity is the assumption that appearance equals reality: if something looks like a tiger, it is a tiger. Here, surface features are used to infer deeper properties, and similarity drives categorization rather than direct causal inference. In voodoo, the action on the image is taken as acting on the real object because, at a fundamental level, the two are seen as one. Piaget (1983) described a related phenomenon in children as nominal realism—the tendency to believe that words or names embody the properties of the things they refer to. The label “poison” is not itself poisonous, yet even adults often act as if it were.

Extensive laboratory experiments with educated American undergraduates illustrate the power of appearance equals reality. Most students preferred a round piece of chocolate fudge to a piece with identical ingredients shaped realistically to resemble dog feces. Similarly, they were much less willing to place a clearly rubber toy vomit in their mouths than a plain rubber sink stopper of comparable size. In a striking demonstration of nominal realism, subjects watched sugar being poured from a commercial container into two bottles; they then randomly placed a “sugar” label on one and a “sodium cyanide, poison” label on the other. Despite knowing they themselves had attached the labels arbitrarily, participants were reluctant to consume sugar from the bottle marked “poison.” Remarkably, this reluctance persisted even when the label read “Not Sodium Cyanide,” a result compatible with Freud’s (1966) discussion of unconscious responses to negation, though it does not establish that the unconscious generally fails to process negatives. In these studies, subjects explicitly acknowledged that their negative feelings were unfounded, but they reported experiencing them anyway. The findings show a gap between participants’ explicit knowledge and their feelings about visual forms or labels.

---

Similarity-based thinking can be understood in relation to generalization: a response learned for one stimulus transfers to similar stimuli. That can be useful, but resemblance alone does not establish a shared cause, a shared property or a physical connection.