'Plipping' through language
WORDS AND RULES: The Ingredients of Language
Here are three riddles you should try before deciding whether you want
to read Steven Pinker's new linguistics
book. If the past tense of "walk" is "walked," what is the past tense
of "balk"? If the past tense of "sing" is "sang," what is the past tense
of "bring"? And if the past tense of "go" is "went," what is the past
tense of "do"?
The answers to those riddles are obvious to anyone who knows enough English to read a comic book, but what makes them interesting to linguists like Pinker is that they are not equally obvious. The past tense of "balk" can be inferred directly from the pattern of "walk": "walked," and exactly the same inference gives you the right past tense for the vast majority of English verbs, like "rub," "rap," "retract," "reimburse," "contravene" and "discombobulate." It is more than an inference, because it is a pattern that imposes itself on the minds of English speakers even in the absence of an overt stimulus like "walk": "walked." You can guess what those sophomores replied when they were told to make the unreal verbs into past tenses: "slaced" and "plipped," they said, with near-perfect unanimity.
Regular past tenses are the product of a rule, a minuscule piece of the computational equipment in the mind that comprises the language faculty and, not incidentally, distinguishes us from beasts. Because of this rule, we don't have to convene international congresses to decide on the past tense every time a new verb enters the language. So when people engage in brand-new activities such as faxing and telebanking, they can report that they "faxed" a document and "telebanked" a deposit, literally without giving it a thought. If people ever feel the need to contraceive something (the heretofore-missing verb implied by the noun contraception), putting the verb into the past tense will be the least of their worries.
These may be simple facts, but their implications are enormous. Our ability to compute past tenses and other grammatical niceties with subconscious, automatic rules is what makes language a powerful, creative human attribute. Because of it, we can, for instance, understand statements we have never encountered before -- the sentences in this paragraph among them -- and we can reply with collocations of words we have never used before. And we can do it endlessly, without limit, as long as we have breath.
Our language ability requires so little care and maintenance that we seldom think about it. It comes to us free, built into our human nature, just as puberty and left- or right-handedness are, and so deeply imprinted that we can talk long before we can skip a rope or subtract or play a piano.
When we do think about language, it is often because something goes awry, as in those other riddles. If the "sing": "sang" pairing momentarily beguiled you into offering "brang" for "bring," instead of "brought," then it might have crossed your mind that something was out of whack -- not in you, necessarily, but in the language. And the pairing of "go": "went," common as it is, offers no model for the past tense of "do" or, come to think of it, for anything else.
These are irregular verbs, a few of the 164 in the language today (by Pinker's count). They are irregular precisely because we cannot compute their past tenses using our innate equipment. "Go" and its past tense truly belong to the lunatic fringe of the language, because they merge parts of two unrelated verbs. As Pinker puts it, "went originally went with wend, and now goes with go." Only one other English verb does that, and that is "was"/"were" for "am"/"is"/"are," the special lunacy of copula "be" (with its infinitive from yet a third verb). We learn these forms by rote, not by rule.
Though the irregular verbs appear odd, Pinker takes great pains to point out that they are just like thousands of ordinary words that form the base material for speech. There are no rules that can predict what something like dog might mean, because its meaning might as readily be assigned to sounds such as chien or Hund or shunka (as it is in French and German and Stony). The meaningful parts of words (their roots) have to be learned one by one. We start picking them up in infancy from our mums and other caregivers, and we keep on learning new ones all our lives. The word-roots become the input to the innate computational processes, but once in a while those processes get pre-empted, as when we need to summon the past tense of bring and go from memory.
Pinker shows that the interplay between memory and process (the "words and rules" of his title) is so intricate as to challenge thinkers from Hobbes to Chomsky. At his grandest, he touts the English past tense as the "perfect site" for resolving the 350-year-old empiricist-rationalist debate, and to that end he devotes an inordinate amount of space to a current academic skirmish between genera- tivists and connectionists, construed by Pinker as a battle between rule-mongers and memory-zealots, which will probably please neither side.
Pinker makes a better impression when
he plays to the bleachers, and he does that often, peppering his pages
with cartoons and smart-aleck terminology and digressions on child-rearing,
word frequencies, baseball talk, brain diseases and much more. "Ultimately
all knowledge is connected," he says, "and insight into a phenomenon can
come from any direction, from the outcome of the Battle of Hastings to
the sequence of a kinase gene." Right or wrong, that notion allows Pinker
to take his readers on an intellectual joyride. At some point, they might
wonder if they have "joyridden" or "joyrided" or (yikes!) "joyrode." That
too is in the book.