Delahunty and Garvey
conditions (a) fail to apply to all members of a parts-of-speech class, and (b)
seem to apply to words outside the class that the conditions are chosen to
identify.
Two natural reactions to this situation are possible. One response would
simply be to ignore the anomalies and present the conditions as absolutes.
This approach requires that conditions be both necessary and sufficient.
(Conditions are necessary if they all have to apply; they are sufficient if no
other conditions are needed.) This approach has the serious disadvantage of
colliding head-on with reality.
A second response would give up the entire enterprise of defining parts
of speech as too haphazard to be worth doing. But if we adopt this course
of action, we will never learn anything about language and our students will
wrongly assume that it is utterly chaotic.
As you might suspect, we view both of these extreme positions as funda-
mentally wrong—wrong about the nature of language and wrong about the
way in which language should be studied. Let us now examine why.
Let’s begin with two exercises (or party games, if you prefer). First, ask
a group of your friends to make a list of ten birds. If you tally your lists of
birds, you will find that certain names appear early on many lists (e.g., eagle,
robin, sparrow) while others appear later (e.g., owl, crow). Also, some names
will appear on almost all lists while others (e.g., chicken, penguin, ostrich)
will appear less frequently or not at all.
The next step is to ask your subjects why they made their choices. They
will probably say that birds have feathers, lay eggs, and are able to fly. And
in the clearest cases, all these conditions apply. However, in less clear cases,
some but not all of the conditions apply or the conditions conflict. For ex-
ample, robins have feathers, fly, and lay eggs, while chickens typically do not
fly, or at least do not fly very far; penguins and ostriches do not fly at all.
While our experiment tells us about how we categorize actual birds, it
also tells us about how we use the word bird. When we use an expression
like bird, we group together objects not so much on the basis of a rigid set
of characteristics, but on a set of criteria or conditions that we use flexibly.
In this way we group together objects on the basis of what the philosopher
Ludwig Wittgenstein called “family resemblances.” You might envision,
then, a target, with some objects close to the bull’s eye—those entities that
are clearly birds. Those entities that fall closest to the center of the target are
called prototypes. They possess all of the features typical of the category.
Toward the periphery of the target lie entities that are less “birdy,” according
to how many of the conditions they meet. In language, also, the boundaries
between classes of words may be fuzzy. That is, just as we may occasionally