If your psychology instructor held in her hand a thin rectangular metal object with a video screen and it suddenly started to ring, you wouldn’t think, “Hey, look at that, a ringing piece of metal with a video screen.” You’d think it’s her “smartphone.” If she was carrying 500 sheets of paper bound together with some cardboard on the top, bottom, and one side, you wouldn’t think, “Look at all that paper and cardboard.” You’d think it’s “a book.” If she was accompanied by a short-legged, midsize, mixed-breed Labrador/basset hound, your first thought probably would not be “It’s a mammal” or “It’s a short-legged, midsize, mixed-breed Labrador/basset hound.” You’d think it is “a dog.”

All this might sound kind of obvious. Yet these simple examples illustrate a critical fact about thinking: People categorize objects and events. To categorize something is to see it as a type of thing, that is, as a member of a group. Almost every time you see anything—a person, plant, animal, object, or event—you recognize it as a member of a category. Even in those rare cases when you can’t immediately categorize an object (see photo), the rarity tells you that 9999 times out of 10,000, categorization is easy.

309

Let’s explore two ways in which categories differ: (1) category level and (2) category structure.

Consider three categories we just mentioned: (1) mammal, (2) dog, and (3) mixed-breed Labrador/basset hound. “Dog” is a higher-level category than “mixed-breed Labrador/basset hound” because all mixed-breed Labrador/basset hounds are dogs. Similarly, “mammal” is a higher-level category than “dog” because all dogs are mammals. Categories 1–3 are ordered from highest to lowest.

Category level thus refers to a relation between categories in which low-level categories are contained within (i.e., are subsets of) higher-level ones (Rosch, 1978). Compared to higher-level categories, lower-level categories are relatively narrow and specific, and they contain fewer members. Among the animals of the world, far fewer are “mixed-breed Labrador/basset hounds” than “dogs” or “mammals.”

The psychologist Eleanor Rosch (1978) explained that some category levels seem more natural to us than others. Some, in other words, come to mind spontaneously, whereas others are rarely used. Suppose a friend purchases a new item—the one in the photo. You probably would not say, “Hey, you got a new vehicle.” The category “vehicle” is unnaturally high. You also are unlikely to say, “Hey, you got an all new 2015 black Honda Fit with alloy wheels.” That category is unnaturally low. You would probably use a middle-level category, saying, “Hey, you got a new car.”

Middle-level categories that we use naturally are called basic-level categories. Basic-level categories are moderately abstract categories that combine two qualities: informativeness and efficiency (Rosch, 1978). You call your friend’s vehicle a “car” and (returning to an earlier example) your professor’s animal a “dog” because the basic-level categories “car” and “dog” efficiently provide enough information for most purposes.

310

How do we know if an item fits into a category? Category structure refers to the rules that determine category membership. Different categories have different structures, that is, different types of rules. You can see this with a simple exercise. Perform these two tasks:

Task #1: Categorize the following numbers into the categories “odd” or “even”:

17

44

100

53

9

Task #2: Categorize the following into the categories “educational” or “entertaining”:

a calculus lecture

The Simpsons

a documentary movie

a trip to an art museum

a Rachael Ray cooking show

The instructions were the same, yet the tasks differed. In #1, categorizations were clear-cut; every item fit unambiguously into one category or the other. In #2, some items were clear-cut—calculus is educational; The Simpsons is pure entertainment—but others were ambiguous. Rachael Ray is entertaining, but you learn a little something, too.

The categories “odd” and “educational” (or “even” and “entertaining”) have different types of structure. “Odd” has a clear-cut boundary; there are no ambiguous cases. “Educational” does not have a clear-cut boundary; rather, experiences vary in the degree to which they are educational, and sometimes it is hard to tell whether an example belongs in the category or not. Let’s look at three types of category structure: classical, fuzzy, and ad hoc.

CLASSICAL CATEGORIES. Classical categories have rules that determine membership unambiguously. There are no “in-between” cases; items clearly fit into a category or not. This category structure is called “classical” because it has been discussed since the classical era of ancient Greece by writers such as Aristotle (1963 CE/350 BCE).

“Odd” and “even” are classical categories. The defining feature for “even,” for example, is “a whole number that, when divided by 2, results in a whole number.” If a number has that feature, it is even; if not, it is not. Another example is “bachelor.” If you are a man and not married, you are a bachelor; otherwise, you’re not. There are no in-between cases. Examples discussed in the classical era by Aristotle include location (if you are “in the market,” you cannot be anywhere else) and time (e.g., “yesterday” identifies a clear-cut time period).

FUZZY CATEGORIES. The boundaries of many everyday categories are “fuzzy” rather than clear-cut. Fuzzy categories have boundaries that are ambiguous; it can be difficult to tell whether an item is a member of the category or not.

You’ve already seen a fuzzy category: “educational.” Was the Rachael Ray show educational? Maybe no (it’s not as educational as calculus), but maybe yes (you do learn something). Another example is the category “athletes.” Soccer players are athletes. But what about race car drivers? Pool players?

311

A philosopher, Ludwig Wittgenstein (1953), first explained how categories violate the classical structure. When looking closely at how people actually use categories, he explained, you rarely see explicit rules and sharp boundaries like those that define even/odd or bachelor. Instead, rules are subtle, ambiguous, and used flexibly. He gave the example of the category “game.” What defines the category? It’s hard to say. Checkers and Monopoly are games—but so are the strategic military exercises known as “war games.” Politics, romance, and life itself could be said to be games. The category structure has no sharp boundaries.

Wittgenstein suggested that categories have a family resemblance structure: Category members share many features, but no single feature is absolutely necessary for membership in the category. Category members are like members of a family (Figure 8.1); they are similar without necessarily sharing any one feature. The category “game” is like this. There is no one thing an activity must have to be a member of the category—not pieces, or a board, or a ball, or competitors you play against (solitaire is a game). Yet “games” have enough shared features that we can immediately tell that some things are games (football, Monopoly) and others are not (coughing, dreaming).

Eleanor Rosch proposed a related idea: prototype structure. In a prototype structure, a category is defined by its most typical, central member. That member is the prototype, the “clearest case of category membership” (Rosch, 1978, p. 36). Individual items are closer or farther from the category’s center, depending on their resemblance to the prototype.

A category with prototype structure is “chair.” A wooden chair, of the sort you might find at a kitchen table, is a prototypical chair; it’s at the center of the category. A large vinyl bag filled with small pieces of foam also is a member of the category “chair”—it’s a “bean bag” chair—but it is far from the category’s center; it is low in prototypicality.

312

When you think about categories, your thoughts reflect the prototypicality of its members (Rosch, 1978). Think of a bird. Now think of another. And another. If you’re like most people, the first one that came to mind was a prototypical bird (e.g., a robin). If you thought of one low in prototypicality (penguin, ostrich), it probably wasn’t until the second or third bird. When listing examples of a category, people tend to mention prototypical members first.

Prototypicality also influences the speed of thinking. When judging whether items belong to a category, people make judgments about prototypical items more quickly. Consider the category “sex.” People more quickly judge that prototypical items (lust, arousal) are category members than less prototypical items (candles, bed). Response speed also reveals individual differences; people with a stronger heterosexual orientation more quickly judge that “reproduction” is a member of the category (Schwarz, Hassebrauck, & Dörfler, 2010).

Prototypes also affect emotions. People like prototypical members more. This is true even for abstract categories. When participants viewed random patterns of dots, they liked the patterns that were similar to the most typical, average dot pattern—the prototype (Winkielman et al., 2006).

Prototypes can vary across cultures. Consider, for example, the category “good person.” To find out if its prototype structure varies, researchers asked people from a variety of countries to list the qualities they associate with a “good person” (Smith, Smith, & Christopher, 2007). Ethical and moral qualities such as honesty, kindness, and caring for others were part of the prototype universally. Yet there were also cultural differences. For example, in Taiwan but not in the United States, being self-directed and achievement-oriented was central to the category “good person.”

AD HOC CATEGORIES. Not all categories have classical or fuzzy structures. Ad hoc categories are groupings of items that go together because they relate to a goal that people have in a specific situation (Barsalou, 2010). The phrase ad hoc means “for this”; ad hoc categories thus contain items that are all useful for some common purpose.

The following example (from Little, Lewandowsky, & Heit, 2006) shows why ad hoc categories need to be included among the types of category structures. Consider these items:

Pictures

Cats

Children

Jewelry

Money

Documents

313

Do these items strike you as members of any one category? They would if a house caught fire. They are members of the category “things to save from a burning house.” This ad hoc category has a meaningful structure, but it isn’t classical or fuzzy; there are no clear-cut classical boundaries and no prototypical item. Ad hoc categories, then, are unique—distinctive ways in which people group together people, places, and things on specific occasions.

You may have noticed that the categories we have discussed—“bachelors,” “birds,” “chairs,” “things to save from a burning house”—were identified using words; they were language-based. Not all categories use language. For instance, you can categorize the sounds of musical instruments without being able to describe them in words. (If asked what a guitar sounds like, you wouldn’t have much to say except, “Um, like a guitar.”) But most of the time, categorization and language are intertwined. Let’s turn, then, to the psychology of language.