Chapter 5 CHAPTER 4 SYNTAX - THE FOUNDATIONS OF INTELLIGENCE

It is hard to imagine how creatures without language would think, but one might suspect that a world without language would be in some sense similar to a world without money—in a world without money, actual things would be exchanged, and Not the metal or paper talismans that represent their value.In this case, the simplest transaction is still slow and cumbersome, and the slightly more complicated transaction is easier said than done! Compared with our "closest relatives" among the living apes (who also have a considerable part of our abilities, such as social intelligence, soothing touch and deception), humans have certain outstanding abilities.We have syntactic language through which we can reason with metaphors and analogies.We always plan ahead, conceive a vision for the future, and then put contingencies together to make choices.We even have music and dancing.What are the stages in the evolution from chimpanzee-like organisms to humans?That is a central question for us humans. Undoubtedly, syntax is the main hallmark of human intelligence, and without syntax, humans would not be much better than chimpanzees.Using the example of an 11-year-old deaf boy who spent 10 years without sign language, Oliver Sack describes what life would be like without syntax: Joseph can see, recognize, categorize, and apply; he has absolutely no problems perceptually categorizing and generalizing, but his abilities seem to be limited there, he has no grasp of abstract concepts, no reflexes , nor play and plan.He seemed to have no imagination--couldn't play with images, didn't know what a hypothesis was, couldn't speculate on possibilities, couldn't enter the realm of imagination or metaphor... He seemed like an animal, or a baby, confined to the present, limited to the immediate and the immediate. Stereotyped perception, although he uses consciousness beyond the reach of an infant. Cases like this show that any innate disposition to language must have been developed by practice in early childhood.Joseph lost the opportunity to see how syntax worked during critical years of his infancy: he couldn't hear people speaking, and he didn't know what the syntax of sign language was. There may be a biological program, sometimes called a universal grammar.It is not the grammar of thinking itself (after all, each dialect has a different grammar of thinking), but the nature of discovering specific grammatical rules from the many possible grammatical rules around it.In order to understand why human beings are so smart, we have to understand how our ancestors reconstructed the entire set of symbols of apes and improved them through the creation of syntax. Sadly, all that remains of our ancestors over the past 4 million years is stone and bones, not their advanced intelligence.Other species branched off along this pathway, but they no longer exist for us to test.We have to go back to 6 million years ago, when no species with our common ancestor existed: the non-human branch itself split into two—chimpanzees and, to a lesser extent, bonobos—3 million years ago.If we want to learn something about the behavior of our ancestors, bonobos are our best choice. They have more behavioral similarities with humans. They are also much better than chimpanzees for studying language; for chimpanzee language Research in the 1960s and 1970s was star-studded. Linguists have a bad habit of thinking that anything that lacks syntax is not language.That is to say that Gregorian chant is not music because it lacks the technique of close and corresponding counterpoint, parallel voice progression, and thematic mirror inversion that Bach used.Because linguistics confines itself to "Bach and after Bach," it leaves dealing with pre-syntax problems largely on the head of "musicologists" such as anthropologists, behaviorists, and comparative psychologists. superior.It is a curious category error that linguists have traditionally dismissed all such research ("You know, that's not actually language!"), since the object of such research is to recognize the language provided by syntax. Antecedent events for powerful build features. One can sometimes get some help from ontogenetic reproduction of the phylogenetic process, but human language is acquired so rapidly in infancy that I doubt that integrating it would completely obscure any initial stages, much like It is like completely forgetting the old post roads because of the road.With the development of phoneme boundaries, language learning begins to go on a fast track in infancy: the original phoneme becomes a "magnet" that captures many variants.Afterwards, significantly more new words are absorbed in the second year, and the ability to introduce isotypes is acquired in the third year (the child suddenly starts using the past tense (-ed) and plurals (-s)), without much practice Can make sentences consistent, and acquire narrative and reverie by the fifth year.Fortunately for us, chimpanzees and bonobos lack this fast-track development, so it gives us an opportunity to study those intermediate stages in their development that precede our strong syntactic abilities. Vervet monkeys in the wild can use four different calls to alert them to different major predators.They also use other calls to call a group together or to warn of the approach of another group of monkeys.Wild chimpanzees have about thirty or forty different calls to express about thirty or forty different meanings, and each call has its own meaning like that of the vervet monkey.The chimpanzee roars loudly "Wow" to express defiance and anger, and the soft cough-like cry, surprisingly, expresses a threat. "Ooo" is a mixture of fear and curiosity ("This weird thing!"), and a soft "Whoo" is strange but not hostile ("What is this thing?") If a "wa-mingla-hoo" call meant something other than "hoo-mingla-wah," the chimpanzee had to ignore the standard meaning of each call itself until the entire string of sounds was heard. A judgment is made only after complete analysis.This is not the case.The combination of these calls is not used to express a special meaning. Humans also have about 30-40 vocal units called phonemes, but they are all meaningless.Even most syllables, like "ba" and "ga," are meaningless unless combined with other phonemes to form meaningful words, like "bat" (bat) or "galaxy" (galaxy).At a certain stage of evolution, our ancestors made the meaning of these vocalizations themselves completely obliterated.Only their combinations make sense: we string meaningless sounds together to form meaningful words.This is unique in the animal kingdom. Furthermore, the sequence of tokens can be strung together—for example, to form a sentence with words and short explanations—as if the same principle is repeated at another level of organization.Monkeys and pigs may repeat vocalizations to reinforce their intended meaning (many human languages ​​such as "Polynesian" do this), but so far no wild nonhuman animal has been found to combine different vocal rates to establish Whole new meaning. No one has yet explained how our ancestors got over the barrier of replacing a single sound with an ordered combination of meaningless phonemes, but it may have been one of the most important events in the evolution from ape to man . At least in the sense of a simple coordination system, bees seem to have broken the one-sign-one-meaning pattern.The figure-of-eight "waggle dance" that a bee performs on its way back to the hive, relays information to fellow bees about the location of food it has found, among other things. The direction of the 8-shaped axis is the direction where the food is located.The duration of the dance is proportional to the distance from the hive, for example, following at least the traditional version of the story, the Italian bee doing three figure-eight dances means it finds food about 60 meters away from the hive (for the German bee it is about 60 meters away from the hive). Means 150 meters, which is largely determined by genes rather than the group in which it is bred. Human linguists do not think highly of this behavior of bees, Bickerton argues in his book Language and Species: All other animals can only communicate about things of evolutionary significance to them, but humans can communicate about anything...Animal cries and movements are structurally holistic and cannot be broken down into basic components like language Divided, ...Although the sounds that make up a language are meaningless in themselves, they can be combined in any number of ways to form thousands of words, each with its own meaning...Similarly, the words that make up a language—…can be composed in an infinite number sentence.Animal communication is not the same. With enough training, many animals can learn a large number of tones, symbols, or human gestures. Once you have to carefully separate understanding from the ability to communicate, they don't necessarily go hand in hand. As I mentioned before, a dog domesticated by a psychologist understands about 90 signals; the B signals it produces do not overlap in meaning with most of the ones it can receive.A sea lion can learn to understand 190 human postures, but it cannot express anything in its posture.Hou orangutans have even mastered more lexical symbols and can combine gestures to express their needs.Parrots can master 70 vocabularies in about 10 years, including 30 object names, 7 colors, 5 shapes, and various other "vocabularies", and can use some of them to express needs. All these smart animals can't tell you who did what to whom, and they can't even talk about the weather.But it is clear that our "closest relatives" chimpanzees and bonobos, with the help of experienced teachers who train them, can achieve a considerable level of language understanding.The most talented bonobo can understand sentences it has never heard under the training of Savage-Lenburg. When it hears "Kanzi, go to the office and fetch the red ball", it can do as well as a two-and-a-half-year-old child. Neither bonobos nor two-and-a-half-year-olds can make sentences like this, but their behavior shows that they can understand. Understanding comes first, expression second, just like a child's language development as well. I have often wondered how much of the limited success in the study of ape language is simply due to inadequate training; perhaps teachers have to be sufficiently disciplined to supersede the normal spontaneous nature of language acquisition in young children.If a bonobo in its first two years of life can be trained to understand new words at a rate close to that of children of the same age, is it possible that bonobo then discovers word patterns in the same way that presyntactic children learn? ?But did the process happen so quickly that we fail to see the distinct stages that preceded the mastery of formal syntax (obscured by the integrated "highway" currently provided by the human genome)? All of these animals have an amazing ability to communicate, but is it language?For most people, the usage of the word language is really loose.First of all, it refers to certain languages, such as English, Frisian", Dutch (and German thousands of years ago, from which these languages ​​​​have evolved, and further back, the original Indo-European languages). But language is also used to refer to some general category of elaborate communication systems. Bee researchers use the word language to describe the behavior they see in bees, as do chimpanzee researchers. So, at what stage of evolution did animals What about the ability of symbolic communication to become human language? There is no obvious answer.Webster's Collegiate Dictionary defines language as "a systematic means of communicating thought or feeling by conventional, well-defined actions, sounds, gestures, or symbols."That covers the examples given above.Savage-Lenburg believed that the essence of language was "the ability to tell another individual something he or she did not know before", which of course meant that the receiving individual had to use some sort of Piagetian method of guessing correctly intelligence to construct a meaning. And what about human language?Linguists will immediately say "Ah, there are rules!" They will start talking about the rules implied by the grammar of thought, and question whether non-human languages ​​have rules.Even if some well-trained animals, such as Counts, can combine words in order to express their needs, they don't care.Linguist Ray Jackendoff put it more euphemistically than most, but he sees it no differently: Everyone is arguing about whether apes have language, and they all cite both positive and negative definitions to support their views.point.I think this is a pointless debate, often due to the enthusiasm of some people to shorten the distance between man and animals, while others are trying to maintain this distance, both sides are at any cost.To avoid empty talk, ask: Can apes communicate successfully?No doubt, yes.They can even communicate in symbols, which is amazing today.But, beyond that, they seem to lack an ability (they may have a little of this ability, but not on a par with a human ability) to put these symbols together in an orderly way using the grammar of thought.In short, Universal Grammar, or even something close to it, seems to be peculiar to humans only. What does this debate about real language have to do with intelligence?There is indeed a relationship between the structures of mind discovered by linguists and the rules discovered by ape linguists about bonobo inventions.Let's start with a simple example. Some expressions are very simple and do not need complex rules to analyze their components. This is the case for most requests, such as "banana" and "give", no matter in what order they convey the same information.A simple association is enough.But if two nouns and a verb "dog, boy, bite" appear together in a sentence, how do you connect the words?It doesn't take much mental grammar either, since boys don't usually bite dogs.But "boy, girl, touch" is ambiguous, and in this case some rules help you distinguish which noun is the subject and which is the receptor. This can be determined with the help of a simple rule, such as the subject-verb-object (SVO) structure of most English declarative sentences (the dog bites the boy) or the subject-object-verb (SOV) structure in Japanese.In short sentences, this reduces to the fact that the first noun is the subject—a fact that Counts may have picked up from Savage-Lenburg's usual phrasing of demands (as in "hit the ball at that banana"). one rule. You can also divide words into subjects and objects to distinguish their role in the sentence, either by regular inflection, or by special forms (called case markers), like when we say he (him), to express It is the person who is the subject, and you say him (the accusative case of h) to show that he is the object of the verb or preposition.There used to be many case changes in English, such as using ye (you) to express the subject and you (you) to express the object, but most of them now only appear in self pronouns, as well as who/whom (nominative)/who (object) grid), etc.Special endings also help identify their role in a sentence, as the -ly-ending word softly (softly) is used to modify a verb rather than a noun.In highly inflected languages, such precedent markers are used a lot, making word order less important in determining the role a word plays in constructing a mental model of the connection between the parties involved. In order for us to speak and understand new sentences, our minds must store not only the vocabulary of our language, but also the possible sentence patterns of the language we use.What these sentence patterns describe is not only the combination of words, but also the combination of phrases.Linguists believe that these forms store the rules of language in memory.The combination of all these rules is called the mental grammar of language, or grammar for short. Aggregating words in a simpler way, such as pidgin English (or German when I travel), is called a primitive language by the linguist Bickerton.They do not use many rules of thought, word combinations (boy. dog, bite), by means of customary word order, such as SVO, are sufficient to convey information.A linguist might classify a child's linguistic achievements (understanding and expression) as protolanguage. Children learn grammar by listening (and deaf children by observing sign language).As they learn new words, they also learn word combinations, and a set of complex combinations make up the grammar of the Teu language.From about 18 months, children begin to learn local language rules and begin to use them in their own sentences.They may not be able to describe parts of speech, or diagram a sentence, but their "language machines" seem to know it pretty well after a year of experience. This biological instinct to find sequences and imitate them is so strong that deaf children create their own sign language with inflections ("house signs") when they have no suitable models to imitate.Bickerton has shown that children can invent new languages—Creole—from the original pidgin spoken by their immigrant parents.Pidgin was used to communicate by traders, tourists, "guest workers" (and old slaves) who did not know a true language.Because of all these troubles, there are usually many gestures, which take a lot of time and which convey only a little. In a real language with many rules (grammar), you can say many things with one short sentence.Creole blends are indeed real languages: children of pidgin speakers use the words they hear and create certain rules (grammar) for them.These rules are not necessarily the ones they learned at the same time from their parents' native language.In this way, a new language emerges from the children's mouths as they hurriedly describe who did what to whom. Which aspects of language are easy to acquire and which are difficult to acquire?General categories are perhaps easiest, as children call all four-legged animals "dogs," or all grown men "daddy."It is more difficult to go from general to special.But, as we have seen, some animals can eventually learn hundreds of symbolic representations. A more important point of contention may be whether new categories can emerge from old ones.Comparative psychologist Duane Rumbaugh has noted that prosimian monkeys (lorises, bush monkeys, etc.) and young monkeys are often confined by the discrimination rules with which they were originally taught, unlike rhesus monkeys and monkeys. Apes can learn new rules that contradict old ones.We can also superimpose a new category on top of an old one, but sometimes with difficulty; categorical perception (briefly mentioned earlier when talking about auditory hallucinations) is why some Japanese distinguish between the L sound in English and the R sound The reason why I find it so difficult to sing. In Japanese, there is a phoneme whose pronunciation is between English L and R, and these two English phonemes are mistakenly treated as just umlauts of the Japanese phoneme.Those Japanese who cannot hear the difference will not be able to articulate the two phonemes because they are "captured" by traditional categories. Combining words and gestures is more complicated than one word with one meaning, and it is much more difficult to link several words into a sequence of symbols with special meanings.Basic word order helps resolve ambiguities in meaning when you otherwise cannot tell which noun is subject and which is receptor.The SVO declarative sentence in English is just one of six arrangements, each of which is found in a language, suggesting that it is a cultural convention rather than a biological rule in the way that Universal Grammar assumes. Words indicating time ("tomorrow" or "before") require more advanced skills, words that question information ("what" or "is there") and words about possibility (" may" or "could") as well.It is worth noting the defect of the foreign original language: it does not use the inactive article (a) or the definite article (the) to distinguish whether a noun refers to a specific object or just a general reference.It does not use inflections (-S, such as, etc.) or clause constructions, and often omits verbs, but infers meaning from context. Although it takes a lot of time to learn vocabulary and basic word order, they are actually easier to master than other parts of the language that are bound by other rules.Indeed, research by Jacgueline S. Johnson and Elissa L. Newport has shown that Asian immigrants who start learning English as adults tend to succeed in mastering vocabulary and the basic word order of sentences, However, it encounters great difficulties in other aspects that are not a problem for minor immigrants.At least in English, the arrangement of words such as "who-what-where-when-why-how" deviates from the basic word order, "What did John give to betty?" (What did John give to Betty?) is Common usage.In addition to the basic word order and emphasis used in the questions asked in the quiz show: John gave "what" toBetty? (John gave Betty "what"?).The non-essential word order in English is difficult for immigrant adults as adults, as are other sentences with distant dependencies, such as plural nouns that must match plural verbs, regardless of How many adjectives are there.Not only do adult immigrants make these grammatical errors, but they also fail to detect them when they hear them.For example, the English inflection system: the plural of nouns changes the tail (The boy ate three cookies. That boy ate three cookies), is this an appropriate English expression?Or is the conjugation of the past tense of the verb (Yesterday the girl petted a dog. Yesterday the girl petted a dog) correct?Immigrants who arrived in the United States under the age of seven made fewer of these mistakes than did immigrant adults.For people who started learning English at the age of 7-15, the adult errors they made increased with age at the start, reaching adult error levels by age 15 (I should stress that in all cases language (The immigrants tested by the researchers had all lived in the English-speaking world for 10 years; they usually scored normal on vocabulary and sentences explaining basic word order). By the age of 2-3, children learn the plural rule of nouns: add -s.Until then, they treat all nouns as irregular nouns.But even if they have learned to say 'mice', once they learn the plural rules, they will say the plural of mouse as 'mouses'.Finally, they learn to treat irregular nouns and verbs as special cases (exceptions to the rules).Children have a desire to learn these rules in their first year of life, and the window of opportunity seems to gradually close during school age.It's not impossible for adults to learn the rules, but being completely English-speaking works well for 2-7 year olds, but not so much for adults. Whether you call this biological programming or universal grammar, the hardest aspect of learning a language seems to be made easier by a childhood learnability, like learning to walk upright All the same, with a biological basis.Perhaps this acquisition is specific to language, or perhaps it simply seeks out complex sound-image patterns and imitates them.As far as we know, deaf children (such as Joseph) who often watch chess games, find chess patterns.In many ways, this pattern-seeking biological program seems It is an important foundation of human intelligence. Grammar is defined in the dictionary as (1) Morphology (word forms and endings); (2) Syntax (from the Greek word "arrange" meaning to arrange words into phrases and sentences); (3) Phonetics (voice and its arrangement).But just as we often use "grammar" loosely to refer to society's common, correct usage, linguists sometimes go to the other extreme, using overly narrow rather than overly loose definitions.They often use the term to refer to a small part of the grammar of thought.All those little words that help convey relative positional information, like near (near), above (above), into (inside), whatever you call them, they're important in our understanding of intellectual It is also very important in the analysis. First of all, the relationships that these grammatical items can express include relative positions Fabove (above…), below (below…), in (in…), on (above…), at (at…), not (adjacent to...), nextto (adjacent to...); relative direction (to (to...), from (from...), through (pass...), left (left), right (right), up (on), down (below)); In addition, there are relative time (befor (before...), after (after...), while (at one time), and reflect tense]; relative quantity cmany (many), few ( less), some (some), an s added after the plural]; the article expresses relative familiarity, the definite article is used for things that someone thinks the listener can understand, and the indefinite article "a" or "an" Used for things that are not understood by the listener; other grammatical items in Bickerton's checklist denote relative probabilities can (can), may (can), might (possible subjunctive voice)] and relative chance of unless (unless ), although (although), until (until), because (because) and various relations, such as owning (-of), possessive -s, have (have); medium out (by means of); purpose [for (for...) ; Necessity [must (must), haveto (must), Qing Ren [should (should), oughtto (must); exist [be (is)]; does not exist (n (not), none (no one), not (Not right; prefix bn-(not a knife, etc. Some languages ​​use conjugations to indicate whether a statement is based on personal experience or indirect information. Therefore, grammatical words help to outline the relative positions of objects and events on the mental map of mutual relations.Since correlations ("bigger", "faster", etc.) are what analogies are usually compared to (as in "bigger is faster"), this aspect of grammar that determines lexeme may also contribute to intellectual improve. Syntax is a tree structure of interrelationships in your mental model of things, which goes far beyond the usual word order or the aforementioned "positioning" of grammar.With syntax, the speaker is able to quickly convey to the listener a mental model of who did what to whom.These relationships are best represented by an inverted tree structure—not the "sentence diagram" of my high school days, but the modern phrase structure called an X-cup.Since there are already some excellent popular books on the subject, I will not repeat them here. The tree structure is most evident in clauses, as in the nursery rhyme about Jack building a house (This is the farmer so wing the corn/ That kept the cock crowed In the morn/ … That lay In the house that Jack built )". Bickerton explains that this nesting is possible because: ...phrases are not beaded together as they appear. Phrases are like Chinese boxes, nested within one another. The importance of this idea cannot be overestimated. Many who are concerned with the origin of human language, and with the supposed capacity for language in animals, make an overall simplistic view of how language arose on the basis of a false assumption. Hypothesis. They think that tones are connected into phrases in sequence, and phrases are connected into sentences in sequence, just like steps are connected into walking... This is far from the truth... The following phrase clearly shows this point: the cow with Crumpled horn that Farmer Giles likes (Farmer Giles likes the crooked cow of the horn). Although no word in this phrase is ambiguous, the whole phrase is, because we don't know what Farmer Giles likes. Is it the horns or the cow. In addition to "phrasal structure", there is also "semantic structure", which is especially useful for guessing what role different nouns play in a sentence.If you see an intransitive verb such as "sleep," you can be sure that there is only one noun associated with it, the subject.This is the case with the word sleep in any language.Similarly, if a language has a verb meaning "to hit," there must be two nouns associated with it, a subject and a receptor (and perhaps a third noun as instrument—with what to hit).The verb "to give" must have 3 nouns associated with it, because it always needs something to give to the recipient.So any mind group for "give" has 3 empty slots that must be properly filled before you can understand the entire sentence correctly and move on to the next assignment.Sometimes the noun is implied, as in the imperative "give!" - in which case we automatically fill in "money" and "give me". As Bickerton pointed out, a sentence is like... a little play or a little story in which each word plays its particular, limited number of roles.Not all linguists agree on exactly what these roles entail, but most, if not all, will involve subjects ("John" cooks), recipients or donors (John cooks "rice"). ”), target (I gave it to Mary), source (I bought it from Fret), tool (Peel cut it with “knife”), beneficiary (I bought it for “you”), and time and place. No language of any wild animal has such structural features.Wildlife languages ​​have dozens of calls and associated emphasizing patterns (often repetitive, like the waving circle of a bee, or the repeated alarm call of a primate), and combinations of calls are rarely used to convey new information.Some animals have been trained to understand and respond correctly to fixed word orders, such as "Counts, touch a banana with a ball"—the order of words in such sentences distinguishes subject from receptor. But linguists are reluctant to include this understanding of sentences in the category of language.In animal experiments, they wanted to see grammatical sentence formation, not just comprehension; comprehension, they insisted, was too easy.While guessing the meaning is often enough for comprehension, crafting and uttering a unique sentence will show right away whether you use the rules well to avoid ambiguity. But the test for making sentences is more a test of how good a scientist is than a test of how good a language learner is.The first faculty a child possesses is, after all, understanding.The original idea of ​​teaching sign language to deaf chimpanzees also involved teaching them how to perform the required movements; the understanding of what the gestures mean, if at all, occurred much later.Now that the study of ape language has finally answered the controversial question of comprehension, it seems that there are more barriers to sentence-making than anyone imagined, but once the animals crossed these barriers, spontaneous sentence-making increased. Linguists aren't too interested in anything simpler than real grammatical rules, whereas behavioral scientists and comparative and developmental psychologists are.Sometimes, to be fair, we speak of the multiple meanings of language, both in the sense of regular communication, and also in the languages ​​of elites who use advanced syntax.All these are conducive to the growth of multifaceted talents, making people more agile, and therefore beneficial to the growth of intelligence.While morphology and phonetics also contribute to our understanding of cognitive processes, phrase structure, semantic structure, and words indicating relative positions are especially interesting because of their constructive meaning that will give us insight into making "right guesses" The thinking structure of intelligence. Comprehension requires an active intellectual process, listening to a person speak, following the sound of the sound, and trying to infer the meaning and cause of the other person--which is always incompletely conveyed in words.On the contrary, sentence making is simple.We know what we think and what we wish to express.We don't have to guess "what we mean," just how to say it.In contrast to the above, when we listen to someone else, we must determine not only what that person is saying, but also the exact meaning of what he or she said, without knowing what the speaker is saying. Savage-Renburg How much of human language ability is innate?Yes, while the drive to learn new words by imitation may be innate, the drive to learn arithmetic is not.Other animals learn gestures by imitation, but preschoolers learn 10 new words a day.As far as imitation is concerned, obviously non-animals are in the same category.The British neuropsychologist Richard Gregory emphasized that the right tool for the job imparts intelligence to its user—words are social tools.So this drive alone may explain why humans have significantly higher intelligence than apes. Preschoolers also have an intrinsic drive to master the rules of composition, which we call the grammar of thought.This is not an intellectual exercise in the usual sense, since even low-intelligence children seem to learn syntax effortlessly by listening.This is not learned in an obvious incremental way of trial and error, as the child seems to transition quickly to syntactic structures.学习显然是起作用的，但是语法的某些僵硬性提示存在先天的布线。如比克顿曾指出的，我们表达关系的方式是 不能增加的（如所有那些上/下的词），虽说你总是能加上更多的名词。由于儿童在学习说话时发生的“错误”在不同的语言之间具有的规律性；由于语法的各个侧面随语言而变的方式；由于那些成年亚裔移民；也由于在任何已知语言中存在着某些似乎是禁用的构成方式；由于所有这些，有些语言学家，如诺姆·乔姆斯基（Noam Chomsky）猜测，在这种普遍性的背后可能存在某种生物性的东西参与其中，也即人脑天生就作好布线，为句法需要的树形结构所用，就像作好布线为直立行走所用一样；正常的说话大部分充斥着片断、开始时说错、缩合方式，以及作为其基础的理想化形式发生的走样。尽管如此…。儿童所习得的是作为其基础的理想化形式。这是一个值得注意的事实。我们也必须记住，儿童构建这种理想化形式毋需接受明晰的传授，他获得这种知识是当他还不能在许多其他领域从事复杂的智力活动之时，这方面的成功相对地独立于智力……确实，脑中存在着一个“语言模块”，它位于大多数人左耳上方的油区中，“通用语法”可能在出生时便布线其中。猴缺少这个左外侧语言区，它们的叫声（以及人类情绪激动时的叫喊）利用的是讲服体上方的更原始的皮层语言区。至今，无人知道猿是否有外侧语言区或相似的组构。如果一只幼倭猩猩或黑猩猩有人类幼儿所具有的两种内驱力；学习新词和发现规则，在脑发育的合适时候又达到足够的量，那么在它们的皮层中是否会自行组织成像我们一样的一个语言区，并用它来从词的混合中结晶出一套规则？或者，虽然脑中的神经布线是先天的，但失去了内驱力或机会，就没有相关的经验，也根本就不再用这些布线了？在我看来，这两种可能性都是和乔姆斯基的观点相一致的，他认为，“通用语法”可能是自组织规则“结晶”的结果，就像“电路自动断续装置”和“滑行装置”由蜂窝状自动装置产生一样。 你实验性地把独特的人的先天布线与信号输入所驱动的结晶化加以区分，你这样做是试图以聪明的动机执意地把词汇和句子强加在出色的猿类学生头上，并以这种模式来取代儿童的非传授的可获得性。我以为；幸好猿是处于语言学家谓之的真正的语言的边缘，因此借助于研究它们的奋斗经历，我们才最终有可能得以了解思维语法的功能基础。在人类的进化过程中，这垫脚石可能已经铺设好，只是已为超结构所窒息，被一体化得无法辨认了。 个体发育有时再现种系发生（婴儿试图站起来再现了四足行走向二足行走的种系发生过程；婴儿第一年中喉的下倾部分再现了从猿到人的变化）。然而，发育发生得如此迅速，以致你不能看到进化过程的再现。不过，如果我们能从倭猩猩看到向更精致的构建的演变，我们便可能发现何种学习能增强句法能力；何种别的作业会有竞争作用从而妨碍语言的发展；与人类相比，猩猩的哪些脑区在用脑血流成象技术观察时看起来会“发亮”“。除了对我们理解人类是多么独特有重要意义外，对猴的语言基础的认识可能也有助于对语言受损者的教育，甚至揭示可望增强语言学习或获得更佳猜测的方法。只有通过倭猩猩这位合格的教师的努力，我们才会有可能回答关于垫脚石的问题。 你正是用句法来构筑更巧妙的思维模式，这些模式包括谁对谁做了什么，为什么，何时，用什么方式等。如果你想传递这样一种精雕细琢的认识，至少你必须把对于这些关系你所具有的思维模式转译为语言的思维语法规则，这种规则将告诉你如何对词加以排列或作屈折变化，从而使听者能重建你的思维模式。当然，更容易的是首先“用句法来思维。”在这种意义上，我们可期望句法能力的增强会导致进行正确猜测的智力的大幅度提高。 最重要的是在听者的头脑中重建你的思维模式。为了把一串词解码为与你本意近似相同的精神上的理解，对你的信息的接受 者将需要知道相同的思维语法。因此，句法就是在你所使用的思维模式中建立各语法项（通常是词）之间的结构上的关系，而并非事物的表现情况，如SVO或屈折变化，这些仅仅是线索而已。你作为一位听者的任务，就是确定何种树形图可以贴切地与你所听到的那单词相吻合。打一个比喻，就好像给你一张空白表格和一组数值，然后让你猜测把这些数值相关起来的公式。一种可能的工作方式是，你先试一种简单的配列（主体动作、受体、修饰词），可以先把一个词留下。你再试另一种树形图，发现有一些未填的但必须填上的空位。然后你利用复数和动词所提供的关于树形图的线索（如你知道动词“给”要求有接受者和所给予的东西）。如果没有词（明说的或暗指的）可填充需填的空位，你就把那树形图勾去，继续作另一种尝试。你可以同时试几种不同的树形图，而不是一个一个来，因为理解（即为那串词找到一种足够好的解释）能以使人目眩的高速度来进行。最后，几种树形图都可能符合要求，而不留下任何词，因而你必须作出判断，在你所处的情况，哪一种解释最合理，那时你就大功告成了。这就是理解——至少以我的版本的语言学家模型来看，当然这个版本肯定是过于简化了。让我们来看看单人纸牌游戏，在这种游戏中一直要到你成功地把所有面朝下的牌都翻过来，你才算是大功告成，虽然遵循递减的顺序和交替的颜色的规则在洗了几回牌后你实际上是不可能成功的。你输了那盘，洗牌后再来。对于某些词的组合来说，不管你对给定的词作何种配列，都不能发现一组有意义的关系，那么你就无法构成一个故事，描述谁对谁做了什么。如果某些人对你说了这么一串模棱两可的词，那么他们就未能通过一种重要的语言能力的测试。对于一个能胜任语言的人所造的大多数句子，你碰到相反的问题，你能构建多种场景，以不同的方式来理解句中的词串。一般来说，其中的一种会比另一些更好地满足于语言的惯例或当时的情景，那么那串词就变成了那次交流所传达的“涵义”。句子中的某些词和词组按上下文可以产生确切的涵义，省得言者说更多的话（代词就提供了这样一种捷径）。 你在中学时学到的正确写作的那些规范性法则，在日常谈话的不完整的表达中实际上无时无刻不在违反。但对日常谈话是足敷应用的，因为真正重要的在于你是否已把“谁对谁做了什么”的你的思维模式传递给了你的听者，按上下文，他们可以把漏掉的词填满。由于书面信息必须在没有多少上下文也没有反馈（如听者脸上豁然开朗或疑惑不解的表情）的情况下让人明白，我们在书写时必须比在讲话时做得更完全，当然也更“累赘”；更充分地使用句法和语法规则。 语言学家想要了解句子是如何从一种类似机器的方式产生和被理解的，这种方式使我们对句子的理解可以达到令人目眩的高速度。我喜欢将其称之为“语言机” （language machine），当然，这堪与古典戏剧中的解围之神（densex machine”）媲美，这是舞台上一个有轮的平台（神机器），神站立其上训示其它演员，现在常泛指剧情困难时杜撰的任何一种安排。在我们的“剧本写作”技术还没有臻于完善之前，我们对于理解句子的算法似乎也是杜撰的。 我将描述这样一架语言机怎样能把短语结构和语义结构以一种算法的形式结合起来进行运转的。语言学家可能会认为这个系统和别的图解式系统一样都太牵强附会。以下是我设计的“真空升降机装箱、传送系统”的若干要点，其中包含的过程就像装运部或生产线上的过程一样简单。 假设我们刚听到或谈到一句完整的句子：“穿一只黑鞋的高个子金发男子，把另一只给了她。”我们怎样来构筑这个动作的思维模式呢？最初，我们需要把其中一些装箱，从介词短语起步值得一试：我们的机器认识所有的介词，并把与之相邻的名词（若句子是英语为后署名词，若是日语则为前置名词）放入相同的箱子。我将用圆角的箱子来装“穿一只黑鞋”和“给她”这样的短语。有时，为了正确地装箱，非语言性记忆签要起作用。例如，在那个模棱两可的组语“the cow with the crumpled horn that Farmer Giles like。”（农夫贾尔斯所喜欢的弯牛角的母牛）中，要是知道贾尔斯在壁炉上有一堆收藏的牛角，这会有助于你作出判断：“that Farmer Giles likes” （农夫贾尔斯喜欢的）应和“cow”（母牛）装箱，还是和crumpled horn”（弯牛角）装箱。 动词由于它们所起的特殊作用将有特别的箱子来装。如果有以“－ly”结尾的副词，或有一个如“必须”这样的助词，我会把它与动词一起装箱，即使它们并不相邻。然后我们把名词短语装箱，把所有修饰它们的介词短语的圆角箱都装进方箱内。如果有一个营句，对于下一级处理来说能把它们看作是一个名词。现在我们把所有的都装起箱来了（至少有两箱，但通常更多）。 下一步，我们需要把它们作为一个整体“提升”起来，在想象中把这个最终理解的混合体从工作地点运走。它会掉落在地上吗？在我的真空升降机上有几种不同类型的手柄，我们必须用的手柄取决于我们所确定的动词［在这种情况是“给”（give）的过去式〕。对于包含主语的名词短语箱有一台真空吸引器（我已把它画成一个小锥形）。不可能有一句既无主语又无动词的句子。如果主语缺失，在开口处空气将被吸入。不可能形成真空，包装箱升降机就不会将它提起。（那就是为什么我在这儿用的是吸收器而不是吊钩的理由，这使目标物成为强制性的。） 但是，正如我在先前指出的，“give”（给）这个动词有点特别，它需要的两个宾语（你不能说“I gave lober”或l gave it”，因此这升降机手柄有两条附加的吸引线。可以有某些非真空型吸引线，即简单的带钩绳，用来传送许多可供选择的名词短语和介词短语，只要该动词允许。 有时需要对吸引口和可供选择的吊钩作某种引导，使之找到合适的目标，例如， SOV可以帮助主语吸引口找到合适的名词短语（如可以是一种格的记号），如“he” （他）。其他的屈折变化也会起作用，例如动词和主语的性、数的一致。吸引口和吊钩可以带有小标签，如“受惠者”、“工具”、“否定”、“强制”、“目的”、“拥有” 等等，仅和与这些范畴相应的词配对。在这架特定的语法机中正是以下的这些组分组成了对句子的识别：能拉起“动词”的手柄，把所有包装箱运走，一只也不留下，而且所有的吸引口均不闲置。如果一个吸引日没有找到合适的目标，那么当你拉起手柄时就不形成真空，你的构建物就不能被运走，也就不能算是大功告成。 如前指出，每个动词一旦为语言机所认定，就有一种特殊类型的手柄。例如，为不及物动词（如“睡觉”）的手柄仅有一个吸引主体的吸引口。但是它们有备用的吊钩为存在的任何有待运送的额外短语所使用：“睡觉”将成功地扮演可选作角色，如时间（“在晚饭后”）和地点（“在沙法上”），但不是受动者。通常有一个真空吸引口是为“主动者”所用（虽然有时并没有主动者，如“冰融化了”），也许会有另一些扮演各种角色的真空吸引口，也有一些备用的吊钩供动词陈述的故事中其他可能的角色所用。 当然，相同的篇中之箱原理（允许一个介词短语起名词作用）允许我们有句中句，就像在从句中，或在“我想我看到了……”这样的句子中。 这就是我们装箱传送系统的简述。如果它似乎是值得小题大做，那么请记住他是进化的守护神”。设想在一间挤满了玩宾戈（Bingo）游戏”的人的屋子里那样，以平行的方式作多种尝试来解读句子，在不同原型句子的支架上重叠地放置着后选句子的许多拷贝，其中大多数因为有残留词和闲置的吸引口而不起作用。其中有一个拷贝，它的动词手柄把所有这些都提升起，这个拷贝便叫道“Bingo/游戏就此结束（当然除非出现平局）。 能够一件不漏地提升一句句子的所有部分是句子配置是否信当的一种试验。请注意，一旦成功，词序和屈折变化都不再是问题，因为角色已经被指定。这架语言机会把某些无意义的句子也提起来，如乔姆斯基著名的例子：“colorless green idears sleep furiously”（无色的绿色想法气愤地睡觉），但是不会有这样的句子：“Colorless green，deas sleep them。”（无色的绿色想法睡觉它们）（动词“睡觉”手柄并无为剩留的宾语的吊钩或吸引口所用。） 虽然对于各种关系建立一个有意义的思维模式可以是交流要达到的目标，不合语法的句子无法解读（除非通过简单的词间的关联），然而还是能使词的组均符合语法，这种结构与句号的要求相吻合，但并不具有与之相关联的任何合理的思维模式。语又试验不同于语法试验。语义也是一种出现平分时的决赛，以决定同时有几个胜者时何人为决胜者，正如在拳击赛时未出现击倒而以点数决胜的情况一样。在猜测农夫贾尔斯可能喜欢的是什么（母牛还是牛角）时，我们也就是那么干的。 虽然每个句子都是一则小故事，但我们也组建一些远比句子更大的以语符列为基础的概念性结构。这些结构也要起到许多强制性和可选性作用。它们紧随语法之后出现，一如作家凯瑟琳·莫顿（Kathryn Morton）所观察到的： 一个幼儿作为人而不是一头吵闹的宠物的第一个标志是，他开始叫得出这世界，并询问其各部分之间是什么关系。一旦他略知其然，他便会教他的玩具熊，把自己的世界观强加于沙地上的任其摆布者。他会一边玩耍，一边自言自语地叙述自己在做什么。他屯会讲述长大以后将于什么。他会注意观察别人的动作，一旦发现不懂流会询问照管他的人。他在临睡前想要听故事。 我们超前计划的能力就是从儿时的讲述发展而来的，这是作出道德上选择的重要基础：我们想象一个动作过程，想象其对别人的影响而决定做还是不做。 借用句法的思维结构来判断各种其他组合的可能行动，我们就能扩展超前计划的能力和智力。在某种程度上它是这样实现的：默默地自言自语，叙述下一步可能会发生的事情，然后应用与句法相似的组合规则来评估（又是用点数来作决定）某一情景是危险而又荒谬的？仅仅是荒谬？可能？很像？还是合乎逻辑的？但是我们动用智力的猜测并不限于语言这样的结构。当我们豁然顿悟某种组合的思维关系，但又难于在此后几周用言词来表达时，我们会叫出声来“我发现了！”“那么，是人脑中的什么使我们能如此擅长于猜测复杂的关系呢？ 我们并不意识到我们的初始仅设对我们的前瞻以及解释收集到的信息的方式的影响有多深。我们应该认识到，非人类的动物并不需要为了使自身具有那些值得严肃研究的特性符合关于人类的语言、工具使用、思维或意识的每一种新的定义。我们强调将人类与地球上所有生命区分开来的那些定义，已经做得太过头了。我们必须重新溶入我们由之产生的生命的洪流中去，努力地去看到在其中间我们大家现在的家系和可能变成的家系。

We can understand neither ourselves nor the world without fully understanding what language is and what it has done for human beings.While it is language that creates us and the world we live in, it unleashes the power to understand and control our environment rather than to expand our own agency.We've been on this path of world control and domination until even the most daring of us have grown terrified of where it might lead us.Now, the engine of our quest for power and knowledge should itself be what we seek to understand.