Surfaces and Essences: Analogy as the Fuel and Fire of Thinking (56 page)

Such intuitions about the “correctness” of people’s first names are simply the result of long-time associations between the name and the person’s entire self. Of course when a child is born, parents often waver between two or more names, and later they thank heaven that they made the “right” choice among all those on their short list, since it fits the child so well. But unless one unshakably believes in an amazing power of parental foresight, one has to admit that had any other choice from the list been made, it would, in the long run, have seemed equally predestined, and in that make-believe world, people would chuckle at the thought of any
other
name (including the actual one, in
this
world) having been chosen. And even so, despite all these careful conscious reflections about the arbitrary nature of a first name, one still can’t totally squelch a little inner voice from saying, “Emmas look like that girl down the block.”

There is a reason for all this: people whom we know well have become stable categories in our minds, and we cannot imagine changing their name any more than we can imagine changing the names that our language gives us for the mundane phenomena that surround us. How absurd it would be if tables were called “chairs”, chairs were called “boats”, boats were called “cars”, and cars were called “tables”. How absurd it would be if the labels of salt and pepper, dog and cat, night and day were all swapped, and if laughing were called “flying”, flying were called “eating”, eating were called “dying”, and dying were called “laughing”.

We even manage to deal quite handily with what might at first seem a paradoxical situation — namely, the fact that we know several extremely different people, all of whom share the first name “Ian”. We like some of them, dislike others of them, and yet it seems to fit them all perfectly. This is very similar to what happens with certain common words such as the noun “story”, which designates not only the recounting of an event but also each specific level of a building, or the verb “to file”, which means both “to classify” and “to grind down”.

Thus a name can act much the same way as other words that designate two or more very different categories. The several Ians we know can coexist in our memory without resembling each other any more than a half-kilogram weight resembles the local animal shelter (even if both are
pounds
of different sorts). This doesn’t mean, however, that we lack an overarching category called “Ian”, to which are attached all sorts of general traits, such as the most likely age and the most likely cultural background, etc. The existence of such general categories makes us susceptible to thinking that categories based on first names are somehow “universal” or “objective”. But this overarching generic
Ian
category coexists with more restricted
Ian
categories that are centered on particular individuals, and these individuals can differ greatly among one another and also from the stereotype of the overarching category, each of them having a first name that seems just right, since after all it is the lexical label that is welded to this specific person.

To sum up the last few sections, we can say that the type of abstraction involved in categorizations resulting in proper nouns leads us to seeing popes, Einsteins, and Jessicas in people who are neither the Pope, nor Einstein, nor Jessica. The same process can also play an essential role in the comprehension of metaphorical statements.

If someone says “Andrew is an ass”, “Fred is a fox”, “Molly is a monkey”, “Sid is a snail”, or “Sue is a snake”, by what right are they assigning these animal traits to humans? How do we know that Sue
isn’t
really a slithering reptile? How do we avoid thinking of an actual animal when we hear certain animal characteristics attributed to someone? Psycholinguists Sam Glucksberg and Boaz Keysar have studied the way in which we understand such sentences, which are usually said to be metaphorical. Other related examples would be “Patsy is a pig” or “My job is a prison”.

Glucksberg and Keysar suggest that we understand sentences like this by constructing, in real time, “ad-hoc” categories (a notion introduced in the preceding chapter). For example, to understand an utterance like “My job is a prison”, we would create on the fly a more abstract category that combines several qualities that constitute the “essence” of the original concept (
prison
), and that shares its lexical label. In this case, thus, the new on-the-fly category is named “prison”, sharing the name of the buildings in which convicts serve out their sentences, and the essential property connected with this new category could be verbalized as “an unpleasant place in which one is confined against one’s will”. A standard prison — a penal institution — is (by construction) a prototypical member of this new category, and the speaker’s job is stipulated to be another of its members. The listener thus realizes that the job has the quality of being
an unpleasant place in which the speaker is unwillingly confined
.

This roughly illustrates Glucksberg and Keysar’s theory of how we understand sentences of this type — namely, by constructing new ad-hoc abstract categories. Their theory would also explain why we don’t drown in a sea of ridiculously literal and confusing interpretations, such as the belief that at the speaker’s workplace there are armed guards in uniform, cells with steel bars at the windows, designated exercise hours, and a private little room where you can speak with visiting family members or lawyers. The theory would also explain why we hear “My surgeon is a butcher” and “My butcher is a surgeon” in very different fashions — namely, the abstract ad-hoc categories that we create in these cases are very different from each other. For the first sentence, the new category derives from a fresh and context-dependent essence of
butcher
, while for the second sentence it derives from an on-the-fly essence of
surgeon
.

As for “Patsy is a pig”, Glucksberg and Keysar’s theory would also explain why no one imagines the poor woman as having a curly tail, wallowing in mud, and oinking — namely, because Patsy is seen as a member of a higher-level category that has been created on the fly by the listener, and which, despite sharing the label “pig”, has been stripped of those superficial features of the barnyard animal. We certainly agree that categorization is central here, but we don’t see the category to which Patsy has been assigned as fresh, ad-hoc, or creative; we have all heard people called “pigs” hundreds of times. What matters about this category is not its
newness
but its level of
abstractness
.

If we possess an abstract category that shares the same name as a concrete category (
i.e.
, if we are in a situation of marking), then we will refrain from transferring
all
the properties of the more concrete category to instances of the more abstract category.
For instance, we don’t need to think hard to realize that it would be folly to imagine
the Mozart of mushrooms
as a wunderkind fungus that composes Viennese music, or
the mecca of wind-surfing
as a sacrosanct spot to which Muslims clad in swimsuits traditionally make pilgrimages atop sailboards, or
the bible of Thai cooking
as central to masses held at ethnic eateries baptized “First Siamese House of Christ”, or
the Rolls-Royce of dishwashers
as an ultra-quiet British kitchen appliance parked by valets in fancy garages. We know this because the categories
Mozart, Mecca, Bible
, and
Rolls-Royce
, having long ago been endowed with a higher level of abstraction by the process of marking, have two types of instances — those that belong to the specific category and those that belong to the more general one. And so we see that the requirement that the category should be
ad-hoc
— that is, freshly minted in real time — is not the crux of the matter. What matters more is that it should possess (at least) two levels of abstraction.

Understanding utterances like “Patsy is a pig” or “Patsy is a powerhouse” is not at all like understanding “Patsy is a pamphlet”, “Patsy is a prawn”, or “Patsy is a palace”. In the latter case, for the sentence to make sense, the category
palace
will have to be creatively extended in real time, without advance warning. Depending on the context, the essential property of the category
palace
might be that it is a rich storehouse of art, or that it is a glittering but nearly unapproachable entity, and so on. If Patsy is seen as a member of this last abstract category, then she will be imagined as being a glamorous or glittery person who exudes an air of aloofness. However, the possible interpretations for “Patsy is a palace” are extremely diverse, and there’s no guarantee that we will correctly divine the speaker’s intention, nor that all different listeners will arrive at the same interpretation of the remark. By contrast, saying about someone that they are a pig is not a clever new invention, but merely a re-use of a stock phrase that has been part of our language for centuries. In other words, the unmarked sense of “pig” — the word’s more abstract sense — has been around for a very long time, and thus using it is neither creative nor ad-hoc.

An ad-hoc category will be created only when the original category — for example,
palace
— doesn’t yet have a standard abstract category based on it. The abstract categories involved in the comprehension of conventional metaphors such as “Patsy is a pig” do not have to be constructed on the fly. Understanding such a sentence can be seen as a perfectly standard, run-of-the-mill act of categorization — an act that places Patsy in the abstract category
pig
₂
as opposed to the concrete category
pig
₁
.

Oddly enough, though, the fact that there exists a concrete category
pig
₁
makes it seem, at least to some, as if the statement “Patsy is a pig” is not only failing to tell the truth, but even that a surrealistic image is being suggested. This is why such a sentence has traditionally been labeled “metaphorical”, as if to suggest that we have no prior abstract category associated with the word “pig” and that we freshly and creatively interpret such sentences each time we hear them. But the fact is simply that the more abstract sense of the word “pig” — namely, “dirty and sloppy” — is an old, familiar category that has the same lexical label as does the curly-tailed animal. Thus, the category
pig
provides a canonical case study of marking. Some people might claim that the sentence “Patsy is a pig” is, in some sense, “less true” than sentences like “Patsy is a
partygoer” or “Patsy is a pouter”. But if one realizes that for the word “pig” there exists not only a concrete, marked category called “pig” but also an abstract, unmarked category bearing that same name, then a more accurate way of describing the situation is to say that while Patsy is a false member of
pig
₁
, she is a true member of
pig
₂
.

The members of higher-level, unmarked categories are not in any sense “less true” members of their categories than are the members of the lower-level, marked categories from which they have been derived. They are merely members of
other
categories, in other contexts. Thus a tea can perfectly well be a
coffee
at the close of a meal when the server asks, “Who’ll have coffee?” Likewise, trucks are perfectly genuine members of the category
car
when little Megan is watching out for
cars
as she walks to school. Boys are genuine
men
when they walk into a men’s room. Kittens are genuine
cats
, people are true
animals
, and so on. And thus Patsy is a
genuine
pig — every bit as much a pig as the quadrupeds that supply bacon — but she is simply a pig on a different level of abstraction — in an unmarked, higher-level sense — of the term.

Beyond the marked category
pig (e.g.
, that of curly-tailed oinkers) and the unmarked one (
e.g.
, that of dirty and sloppy creatures), it’s possible that other subcategories of the abstraction category
pig
could exist. And indeed, if one looks in a dictionary, one will find a subentry for a sense of the word that applies solely to humans. A similar situation holds for many other terms, where an abstract, unmarked category bears the same name as some of its subcategories. And so, when a term is very frequent in a language, its “metaphorical” meanings soon take on a life of their own, becoming autonomous new meanings. Understanding such terms is then much like understanding the word “animal”. In some contexts, “animal” will designate an unmarked category that
includes
humans; in other contexts, it will designate a marked category and will
exclude
humans. Neither of these usages, however, can or should be called “metaphorical”. Likewise, when we say “Patsy is a pig”, it is perfectly reasonable to think (and dictionaries will generally support this) that in addition to the abstract category
pig
₂
(of very messy, sloppy, animate beings) and the base category
pig
₁
contained inside it, there is also a different subcategory consisting of
people who are messy.
In fact, as a parallel to what we did for the family of categories sharing the label “desk”, it would be more sensible to label the most abstract category “pig
₃
” rather than “pig
₂
”, and instead to reserve the label “pig2” for human beings who are messy and sloppy, with the label “pig
₁
” applying to the omnivorous farm animal.

In summary, then, we can identify three different types of understanding of “metaphorical” sentences.

The first type involves situations where the metaphorical expression is completely standardized, and in order to understand the sentence one does not need to invoke an abstract category (that is, an unmarked term). The unmarked, abstract category can certainly exist, but it is simply not needed for the understanding process, and hence is bypassed. This can be illustrated in the case of fluent English speakers who hear the
sentence “Jack’s house is a dump”, dealing with it in much the same way as they would deal with “Patsy is a pig”. We assume that the abstract category
dump
— the unmarked sense — existed in their minds long before the sentence was heard, as did the more limited subcategory of
dump
that applies just to messy dwellings (hotels, houses, rooms, etc.). The understanding of “Jack’s house is a dump” is thus straightforward, simply taking advantage of this concrete sense of the word “dump”, which applies specifically to messy dwellings. Moreover, much as the farm-dwelling type of pig and the messy human type of pig share the property of being sloppy and making loud noises, garbage dumps and messy dwellings share the trait of being unpleasant to see and to be in. This is because there is a more abstract category — the
unmarked
sense of “dump” — that includes them both and binds them together. In the mind of a fluent English speaker who hears “Jack’s house is a dump”, this abstract category is present but only as a kind of silent backdrop, and it is not needed in order to understand the sentence, because the more specific subcategory was long ago constructed and this guarantees that the sentence will be easily understood. In other words, much as with
desk
and
pig
, we are here dealing with three distinct categories:
dump
₃
, an abstract category that includes dirty, unpleasant places of all sorts, and two subcategories that are more specific —
dump
₁
for garbage dumps, and
dump
₂
for messy dwellings.