The Language Instinct: How the Mind Creates Language

by Steven Pinker

  Probably no “grammatical error” has received as much scorn as “misuse” of pronoun case inside conjunctions (phrases containing two elements joined by and or or). What teenager has not been corrected for saying Me and Jennifer are going to the mall? A colleague of mine recalls that when she was twelve, her mother would not allow her to have her ears pierced until she stopped saying it. The standard story is that the accusative pronoun me does not belong in subject position—no one would say Me is going to the mall—so it should be Jennifer and I. People tend to misremember the advice as “When in doubt, ‘say so-and-so and I,’ not ‘so-and-so and me,’” so they unthinkingly overapply it—a process linguists call hypercorrection—resulting in “mistakes” like Al Gore and I a chance and the even more despised between you and I.

  But if the person on the street is so good at avoiding Me is going and Give I a break, and if even Ivy League professors and former Rhodes Scholars can’t seem to avoid Me and Jennifer are going and Give Al and I a chance, might it not be the mavens that misunderstand English grammar, not the speakers? The mavens’ case about case rests on one assumption: if an entire conjunction phrase has a grammatical feature like subject case, every word inside that phrase has to have that grammatical feature, too. But that is just false.

  Jennifer is singular; you say Jennifer is, not Jennifer are. The pronoun She is singular; you say She is, not She are. But the conjunction She and Jennifer is not singular, it’s plural; you say She and Jennifer are, not She and Jennifer is. So if a conjunction can have a different grammatical number from the pronouns inside it (She and Jennifer are), why must it have the same grammatical case as the pronouns inside it (Give Al Gore and I a chance)? The answer is that it need not. A conjunction is an example of a “headless’ construction. Recall that the head of a phrase is the word that stands for the whole phrase. In the phrase the tall blond man with one black shoe, the head is the word man, because the entire phrase gets its properties from man—the phrase refers to a kind of man, and is third person singular, because that’s what man is. But a conjunction has no head; it is not the same as any of its parts. If John and Marsha met, it does not mean that John met and that Marsha met. If voters give Clinton and Gore a chance, they are not giving Gore his own chance, added on to the chance they are giving Clinton; they are giving the entire ticket a chance. So just because Me and Jennifer is a subject that requires subject case, it does not mean that Me is a subject that requires subject case, and just because Al Gore and I is an object that requires object case, it does not mean that I is an object that requires object case. On grammatical grounds, the pronoun is free to have any case it wants. The linguist Joseph Emonds has analyzed the Me and Jennifer/ Between you and I phenomenon in great technical detail. He concludes that the language that the mavens want us to speak is not only not English, it is not a possible human language!

  In the second story of his column, Safire replies to a diplomat who received a government warning about “crimes against tourists (primarily robberies, muggings, and pick-pocketings).” The diplomat writes,

  Note the State Department’s choice of pick-pocketings. Is the doer of such deeds a pickpocket or a pocket-picker?

  Safire replies, “The sentence should read ‘robberies, muggings and pocket-pickings.’ One picks pockets; no one pockets picks.”

  Significantly, Safire did not answer the question. If the perpetrator were called a pocket-picker, which is the most common kind of compound in English, then indeed the crime would be pocket-picking. But the name for the perpetrator is not really up for grabs; we all agree that he is called a pickpocket. And if he is called a pickpocket, not a pocket-picker, then what he does can perfectly well be called pick-pocketing, not pocket-picking, thanks to the ever-present English noun-to-verb conversion process, just as a cook cooks, a chair chairs, and a host hosts. The fact that no one pockets picks is a red herring—who said anything about a pick-pocketer?

  The thing that is confusing Safire is that pickpocket is a special kind of compound, because it is headless—it is not a kind of pocket, as one would expect, but a kind of person. And though it is exceptional, it is not unique; there is a whole family of such exceptions. One of the delights of English is its colorful cast of characters denoted by headless compounds, compounds that describe a person by what he does or has rather than by what he is:

  bird-brain

  blockhead

  boot-black

  butterfingers

  cut-throat

  dead-eye

  egghead

  fathead

  flatfoot

  four-eyes

  goof-off

  hard-hat

  heart-throb

  heavyweight

  high-brow

  hunchback

  killjoy

  know-nothing

  lazy-bones

  loudmouth

  low-life

  ne’er-do-well

  pip-squeak

  redneck

  scarecrow

  scofflaw

  wetback

  This list (sounding vaguely like a dramatis personae from Damon Runyon) shows that virtually everything in language falls into systematic patterns, even the seeming exceptions, if only you bother to look for them.

  The third story deconstructs a breathless quote from Barbra Streisand, describing tennis star Andre Agassi:

  He’s very, very intelligent; very, very, sensitive, very evolved; more than his linear years…. He plays like a Zen master. It’s very in the moment.

  Safire first speculates on the origin of Streisand’s use of evolved: “It’s change from the active to passive voice—from ‘he evolved from the Missing Link’ to ‘He is evolved’—was probably influenced by the adoption of involved as a compliment.”

  These kinds of derivations have been studied intensively in linguistics, but Safire shows here that he does not understand how they work. He seems to think that people change words by being vaguely reminded of rhyming ones—evolved from involved, a kind of malapropism. But in fact people are not that sloppy and literal-minded. The lexical creations we have looked at—Let me caveat that; They deteriorated the health care system; Boggs flied out to center field—are based not on rhymes but on abstract rules that change a word’s part-of-speech category and its cast of role-players, in the same precise ways across dozens or hundreds of words. For example, the transitive to deteriorate the health care system comes from the intransive the health care system deteriorated in the same way that the transitive to break the glass comes from the intransitive the glass broke. Let’s see, then, where evolved might have come from.

  Safire’s suggestion that it is an active-to-passive switch based on involved does not work at all. For involved, we can perhaps imagine a derivation from the active voice:

  Raising the child involved John. (active)

  John was involved in raising his child. (passive)

  John is very involved.

  But for evolved, the parallel derivation would require a passive sentence, and before that an active sentence, that do not exist (I have marked them with asterisks):

  * Many experiences evolved John.

  * John was evolved by many experiences. (or) * John was evolved in many experiences.

  John is very evolved.

  Also, if you’re involved, it means that something involves you (you’re the object), whereas if you’re evolved, it means that you have been doing some evolving (you’re the subject).

  The problem is that the conversion of evolved from to very evolved is not a switch from the active voice of a verb to the passive voice, as in Andre beat Boris Boris was beaten by Andre. The source Safire mentions, evolved from, is intransitive in modern English, with no direct object. To passivize a verb in English you convert the direct object into a subject, so is evolved could only have been passivized from Something evolved Andre, which does not exist. Safire’s explanation is like saying you can take Bill bicycled from Lexington and change it to Bill is bicycl
ed and then to Bill is very bicycled.

  This breakdown is a good illustration of one of the main scandals of the language mavens: they show lapses in the most elementary problems of grammatical analysis, like figuring out the part-of-speech category of a word. Safire refers to the active and passive voice, two forms of a verb. But is Barbra using evolved as a verb? One of the major discoveries of modern generative grammar is that the part of speech of a word—noun, verb, adjective—is not a label assigned by convenience but an actual mental category that can be verified by experimental assays, just as a chemist can verify whether a gem is a diamond or zirconium. These tests are a standard homework problem in the introductory course that linguists everywhere call Baby Syntax. The method is to find as many constructions as you can in which words that are clear-cut examples of a category, and no other kind of word, can appear. Then when you are faced with a word whose category you do not know, you can see whether it can appear in that set of constructions with some natural interpretation. By these tests we can determine, for example, that the language maven Jacques Barzun earned an “F” when he called a possessive noun like Wellington’s an adjective (as before, I have placed asterisks beside the phrases that sound wrong):

  : 1. very X:

  REAL ADJECTIVE: very intelligent

  IMPOSTER: * very Wellington’s

  : 2. seems X:

  REAL ADJECTIVE: He seems intelligent

  IMPOSTER: * This seems Wellington’s

  : 3. How X:

  REAL ADJECTIVE: How intelligent is he?

  IMPOSTER: * How Wellington’s is this ring?

  : 4. more X than:

  REAL ADJECTIVE: more intelligent than

  IMPOSTER: * more Wellington’s than

  : 5. a Adj X Adj N:

  REAL ADJECTIVE: A funny, intelligent old friend

  IMPOSTER: * a funny, Wellington’s old friend

  : 6. un-X:

  REAL ADJECTIVE: unintelligent

  IMPOSTER: * un-Wellington’s

  Now let’s apply this kind of test to Barbra’s evolved, comparing it to a clear-cut verb in the passive voice like was kissed by a passionate lover (odd-sounding constructions are marked with an asterisk):

  very evolved / * very kissed

  He seems evolved/ * He seems kissed

  How evolved is he? / * How kissed is he?

  He is more evolved now than he was last year / * He is more kissed now than he was yesterday

  A thoughtful, evolved, sweet friend / * a tall, kissed, thoughtful man

  He was unevolved / * He was unkissed by a passionate lover

  Obviously, evolved does not behave like the passive voice of a verb; it behaves like an adjective. Safire was misled because adjectives can look like verbs in the passive voice and are clearly related to them, but they are not the same thing. This is the source of the running joke in the Bob Dylan song “Rainy Day Women #12 & 35”:

  They’ll stone you when you’re riding in your car.

  They’ll stone you when you’re playing your guitar.

  But I would not feel so all alone.

  Everybody must get stoned.

  This discovery steers us toward the real source of evolved. Since it is an adjective, not a verb in the passive voice, we no longer have to worry about the absence of the corresponding active voice sentence. To trace its roots, we must find a rule in English that creates adjectives from intransitive verbs. There is such a rule. It applies to the principle form of a certain class of intransitive verbs that refer to a change of state (what linguists call “unaccusative” verbs), and creates a corresponding adjective:

  time that has elapsed elapsed time

  a leaf that has fallen a fallen leaf

  a man who has traveled widely a widely traveled man

  a testicle that has not descended into the scrotum an undescended testicle

  a Christ that has risen from the dead a risen Christ

  a window that has stuck a stuck window

  the snow which has drifted the drifted snow

  a Catholic who has lapsed a lapsed Catholic

  a lung that has collapsed a collapsed lung

  a writer who has failed a failed writer

  Take this rule and apply it to a tennis player who has evolved, and you get an evolved player. This solution also allows us to make sense of Streisand’s meaning. When a verb is converted from the active to the passive voice, the verb’s meaning is conserved. Dog bites man = Man is bitten by dog. But when a verb is converted to an adjective, the adjective can acquire idiosyncratic nuances. Not every woman who has fallen is a fallen woman, and if someone stones you you are not necessarily stoned. We all evolved from a missing link, but not all of us are evolved in the sense of being more spiritually sophisticated than our contemporaries.

  Safire then rebukes Steisand for more than his linear years. He says.

  Linear means “direct, uninterrupted”; it has gained a pejorative vogue sense of “unimaginative,” as in linear thinking, in contrast to insightful, inspired leaps of genius. I think what Ms. Streisand had in mind was “beyond his chronological years,” which is better expressed as simply “beyond his years.” You can see what she was getting at—the years lined up in an orderly fashion—but even in the anything-goes world of show-biz lingo, not everything goes. Strike the set on linear.

  Like many language mavens, Safire underestimates the precision and aptness of slang, especially slang borrowed from technical fields. Streisand obviously is not using the sense of linear from Euclidean geometry, meaning “the shortest route between two points,” and the associated image of years lined up in an orderly fashion. She is using the sense taken from analytic geometry, meaning “proportional” or “additive.” If you take a piece of graph paper and plot the distance traveled at constant speed against the time that has elapsed, you get a straight line. This is called a linear relationship; for every hour that passes, you’ve traveled another 55 miles. In contrast, if you plot the amount of money in your compound-interest account, you get a nonlinear curve that swerves upward; as you leave your money in longer, the amount of interest you accrue in a year gets larger and larger. Streisand is implying that Agassi’s level of evolvedness is not proportional to his age: whereas most people fall on a straight line that assigns them X spiritual units of evolvedness for every year they have lived, this young man’s evolvedness has been compounding, and he floats above the line, with more units than his age would ordinarily entitle him to. Now, I cannot be sure that this is what Streisand had in mind (at the time of this writing, she has not replied to my inquiry), but this sense of linear is common in contemporary techno-pop cant (like feedback, systems, holism, interface, and synergistic), and it is unlikely that she blundered into a perfectly apt usage by accident, as Satire’s analysis would imply.

  Finally, Safire comments on very in the moment:

  This very calls attention to the use of a preposition or a noun as a modifier, as in “it’s very in” or “It’s very New York,” or the ultimate fashion compliment, “It’s very you.” To be very in the moment (perhaps a variation of of the moment or up to the minute) appears to be a loose translation of the French au courant, variously translated as “up to date, fashionable, with-it.”

  Once again, by patronizing Streisand’s language, Safire has misanalyzed both its form and its meaning. He has not noticed that: (1) The word very is not connected to the preposition in; it’s connected to the entire prepositional phrase in the moment. (2) Streisand is not using the intransitive in, with its special sense of “fashionable”; she is using the conventional transitive in with a noun phrase object, the moment. (3) Her use of a prepositional phrase as if it was an adjective to describe some mental or emotional state follows a common pattern in English: under the weather, out of character, off the wall, in the dumps, out to lunch, on the ball, in good spirits, on top of the world, out of his mind, and in love. (4) It’s unlikely that Streisand was trying to say that Agassi is au courant or fashionable; that wo
uld be a put-down implying shallowness, not a compliment. Her reference to Zen makes her meaning entirely clear: that Agassi is very good at shutting out distractions and concentrating on the game or person he is involved with at that moment.

  So these are the language mavens. Their foibles can be blamed on two blind spots. One is a gross underestimation of the linguistic wherewithal of the common person. I am not saying that everything that comes out of a person’s mouth or pen is perfectly rule-governed (remember Dan Quayle). But the language mavens would have a much better chance of not embarrassing themselves if they saved the verdict of linguistic incompetence for the last resort rather than jumping to it as a first conclusion. People come out with laughable verbiage when they feel they are in a forum demanding an elevated, formal style and know that their choice of words could have momentous consequences for them. That is why the fertile sources of howlers tend to be politicians’ speeches, welfare application letters, and student term papers (assuming there is some grain of truth in the reports). In less self-conscious settings, common people, no matter how poorly educated, obey sophisticated grammatical laws, and can express themselves with a vigor and grace that captivates those who listen seriously—linguists, journalists, oral historians, novelists with an ear for dialogue.

  The other blind spot of the language mavens is their complete ignorance of the modern science of language—and I don’t mean just the formal apparatus of Chomskyan theory, but basic knowledge of what kinds of constructions and idioms are found in English, and how people use them and pronounce them. In all fairness, much of the blame falls on members of my own profession for being so reluctant to apply our knowledge to the practical problems of style and usage and to everyone’s natural curiosity about why people talk the way they do. With a few exceptions like Joseph Emonds, Dwight Bolinger, Robin Lakoff, James McCawley, and Geoffrey Nunberg, mainstream American linguists have left the field entirely to the mavens—or, as Bolinger calls them, the shamans. He has summed up the situation: