CONSIDER THE FOLLOWING sentence, which is one that most literate Americans can understand but most literate English people cannot, even when they have a wide vocabulary and know the conventions of the standard language:
Jones sacrificed and knocked in a run.
Typically, a literate English person would know all the words yet wouldn't comprehend the sentence. (In fairness, most Americans would be equally baffled by a sentence about the sport of cricket.) To understand this sentence about Jones and his sacrifice, you need a wealth of relevant background knowledge that goes beyond vocabulary and syntax—relevant knowledge that is far broader than the words of the sentence present. Let's consider what we as writers would have to convey to an English person to make this sentence comprehensible. (To avoid making this an entire chapter about baseball, the explanation won't go the whole way.)
First, we would have to explain that Jones was at bat. That would entail an explanation of the inning system and the three-outs system. It would entail an explanation of the size and shape of the baseball field (necessary to the concept of a sacrifice fly or bunt), which would require a digression on what a fly or a bunt is. The reader would also have to have some vague sense of the layout of the bases, how many there are, and what a run is. The English person would need to be given a sense of how the pitching team is positioned on the field and how many of them there are. That would require a gloss on the pitcher and the difference between balls and strikes, which would require at least a vague idea of where the strike zone is. At that point we would have gotten started, but by the time an English reader had begun to assimilate all this relevant background knowledge, he or she may have lost track of the whole point of the explanation. What was the original sentence? It will have been submerged in a flurry of additional sentences branching out in different directions.
That it should take so much time to explain a simple sentence containing not a single unfamiliar vocabulary word illustrates the impracticality of the idea that people can simply strategize what a sentence means or look up the knowledge they need for reading comprehension. Time considerations alone require that the background knowledge needed to fill in the blanks must be quickly and readily available to the reader's mind.
The example also illustrates a feature of reading and writing that we touched on in discussing Basil Bernstein's distinction between an elaborated and a restricted code. In ordinary oral speech—for example, in a conversation at an actual baseball game—the whole meaning of our sentence might be conveyed in the restricted code by two words: "He sacrificed." We wouldn't need to say "Jones," because we already know who is batting. We wouldn't need to say that he knocked in a run, since we would know that too. The only aspect worthy of explicit comment might be that Jones intended to do what he did. The original sentence, "Jones sacrificed and knocked in a run," is far more explicit than "He sacrificed." Nonetheless, it still isn't elaborated enough for a typical English person.
Just how elaborated does writing have to be in order for the general reader to understand it? The answer to that question yields an answer to a key question about improving students' reading comprehension: what knowledge do students need to gain to become good readers?
Recently I was reading out loud a wonderful Robert McCloskey book to a four-year-old. The book, called Burt Dow: Deep-Water Man, is about a retired sea captain who goes fishing and encounters a whale. The reading was not a success. The pictures are marvelous, but the text is intended for a more experienced and knowledgeable reader than that four-year-old was—at the least it needed a four-year-old who lives near the sea and boats and knows what rigging and gunnels and sou'wester are, not to mention planking and weather eye. Arequest was made to engage in another activity. It wasn't McCloskey's fault. Many older children and grownups, even if they are not experienced sailors, can understand and delight in his book. Its sense is available to a reader who is neither an expert nor a complete novice in matters of the sea, just as the Jones sentence is directed at someone who is not a complete novice about baseball. Both texts are directed to a "general reader."
The idea of the general reader is a structurally necessary one that enables mass communication to occur.1 The printed text always takes something for granted, always leaves blanks to be filled in by the reader to make it comprehensible.2 Unless writers and their readers internalize the shared knowledge of the wider speech community (such as shared knowledge about baseball), they cannot expect the blanks to be filled in; they cannot be successful writers or proficient readers.
Yet these ideas are rarely, if ever, mentioned in reading programs. That is yet another illustration of the degree to which current thinking about reading is trapped in formalistic conceptions that regard texts as found objects rather than as historically dependent writings which assume the existence of widely shared background knowledge within a particular speech community.
A creature from Mars who had learned the rules of Standard English and the words of a dictionary plus comprehension strategies still could not read the New York Times, or a computer instruction manual, or Gone With the Wind. The Martian would be in the same predicament as the English person reading about baseball. He would not have enough relevant background knowledge to form an accurate situation model. This point was empirically demonstrated some years ago during the cold war, when hundreds of millions of dollars were spent in developing computer programs that could automatically translate Russian into English. The dollars were wasted, because the project was based on an inadequate theory, which assumed that computers needed only formal language rules, dictionary word meanings, and comprehension strategies to translate meaning. The computer translation scheme began to be more workable when researchers started supplying the computers with a large knowledge base relevant to the subjects being translated.3
Every newspaper and book editor and every producer for radio and TV is conscious of the need to distinguish what can be taken for granted from what must be explained. Learning the craft of writing is bound up with learning how to gauge what can be assumed versus what must be explained. The general reader that every journalist or TV newscaster must imagine is somebody whose relevant knowledge is assumed to lie between the total ignorance of a complete novice and the detailed knowledge of an expert. A newspaper baseball story cannot assume an audience as uninformed about the game as our imagined English person or one consisting of baseball experts. Every person who speaks and writes must make an estimate of what can be left unexplained and what must be explicitly stated. Being able to make such an estimate is part of our general communicative competence, and it is therefore something all our children need to be taught. Reading proficiency, listening proficiency, speaking proficiency, and writing proficiency all require possession of the broad knowledge that the general reader is assumed to have and also the understanding that others can be expected to possess that knowledge.
We said that this topic of taken-for-granted knowledge, which is absolutely essential to reading and writing, is neglected in current reading programs. Yet not to teach the knowledge that is taken for granted in formal discourse systematically is equivalent to neglecting to teach a tennis beginner that the ball has to go over the net and stay inside the lines. Most current reading programs talk about activating the reader's background knowledge so she can comprehend a text. But in practice, they are only paying lip service to the well-known scientific finding that background knowledge is essential to reading comprehension. Little attempt is made to enlarge the child's background knowledge. The disjointed topics and stories that one finds in current reading programs seem designed mainly to appeal to the knowledge that young readers may already have, such as "Going to School" and "Jenny at the Supermarket." The programs do not make a systematic effort to convey coherently, grade by grade, the knowledge that newspapers, magazines, and serious radio and TV programs assume American readers and listeners possess.
Referring to the knowledge that American readers are assumed to possess is another way of saying that the range of specific knowledge needed to be a proficient reader is different among American readers than it is among readers in other nations—a point I tried to illustrate in the baseball example. That is why we speak of the tacit knowledge within a particular speech community—of the knowledge assumed to be known by Americans. In the modern world, the background knowledge a person needs for good reading comprehension tends to be largely (but by no means entirely) national in character. For publications in English, American and English (and Canadian, Australian, and New Zealand) newspaper readers take for granted slightly different ranges of knowledge. The range of assumed knowledge is different among Chinese and Russian newspaper readers and speakers, and so on.
Here is the first paragraph of an article by Janet Maslin, taken at random from the books section of the New York Times on February 6, 2003. It is an example of writing addressed to a general reader that a literate American high school graduate would be expected to understand.
When Luca Turin was a boy growing up in Paris, according to Chandler Burr's ebullient new book about him, "he was famous for boring everyone to death with useless, disconnected facts, like the distance between the earth and the moon in Egyptian cubits" Mr. Burr sets out to explain how such obsessive curiosity turned Mr. Turin into a pioneering scientist who, in the author's estimation, deserves a Nobel Prize.
This example shows that the background knowledge required to understand the general sections of the New York Times, such as the book review section, is not deep. It is not that of an expert—of course not, for we cannot all be experts on the diverse subjects that are treated by books. If authors want their books to be sold and read, they must not assume that their readers are experts. They may take for granted only the relevant background knowledge that a literate audience can be expected to possess.
What do readers need to know in order to comprehend this passage? We need to know first that this is a book review, which aims to tell us what the book is about and whether it is worth reading. We need to understand that the reviewer is favorably disposed to the book, calling it "ebullient," and that it is a nonfiction work about a scientist named Luca Turin. We need to have at least a vague semantic grasp of key words like ebullient, boring, obsessive, pioneering, estimation. We need to know some of the things mentioned with exactness, but not others. It's not necessary to know how long a cubit is. Indeed, the text implies that this is an odd bit of information, and we can infer that it is some form of measurement. (Maslin may assume that we remember that the word is used in the Bible story of Noah's Ark.) We need to know in general what Paris is, what the moon is and that it circles the earth, that it is not too far away in celestial terms, and we need to have some idea what a Nobel Prize is and that it is very prestigious. Consider the knowledge domains included in this list. Paris belongs to history and geography; so does Egypt. The moon belongs to astronomy and natural history. The Nobel Prize belongs to general history and science.
We may infer from this example that only a person with broad general knowledge is capable of reading with understanding the New York Times and other newspapers. This fact has momentous implications for education, and for democracy as well. A universal ability of citizens to read newspapers or their equivalent with understanding is the essence of democracy. Jefferson put the issue unforgettably: "The basis of our government being the opinion of the people, the very first object should be to keep that right; and were it left to me to decide whether we should have a government without newspapers or newspapers without a government, I should not hesitate a moment to prefer the latter. But I should mean that every man should receive those papers and be capable of reading them."4 The last phrase, "be capable of reading them," is often omitted from the quotation, but it is the crucial one. Reading achievement will not advance significantly until schools recognize and act on the fact that it depends on the possession of a broad but definable range of diverse knowledge. The effective teaching of reading will require schools to teach the diverse, enabling knowledge that reading requires.
But what exactly does that enabling knowledge consist of? That is the nuts-and-bolts question. The practical problem of helping all students achieve adequate reading comprehension skills will depend on our schools being able to narrow down what seems at first glance to be vast amounts of heterogeneous information into a teachable repertory that will enable students to understand the diverse texts that are addressed to a general reader. Our sketch of the background knowledge needed to understand Maslin's short passage offers clues to the kind of instruction that is needed to advance general reading comprehension ability. It will be broad instruction in the worlds of nature and culture as a necessary platform for gaining deeper knowledge through listening and reading. But what should that broad general knowledge be?
It is assumed by the American educational community that any "representative" knowledge will do.5 My colleagues Joseph Kett and James Trefil and I set out to develop more useful guidance for schools than this imprecise and inaccurate notion back in the 1980s. We asked ourselves, "In the American context, what knowledge is taken for granted in the classroom, in public orations, in serious radio and TV, in books and magazines and newspapers addressed to a general audience?" We considered various scholarly approaches to this problem. One was to look at word frequencies. If a word appeared in print quite often, then it was probably a word whose meaning was not going to be explained by the speaker or writer. We looked at a frequency analysis of the Brown Corpus, a collection of passages from very diverse kinds of publications that was lodged at Brown University, but we found that this purely mechanical approach, while partially valid, did not yield altogether accurate or intelligent results. For example, because the Brown Corpus was compiled in the 1950s, "Nikita Khrushchev" was a more frequent vocabulary item than "George Washington."6
Another shortcoming of this "objective" approach was its failure to discriminate between language and knowledge, that is, between words and what they refer to. That shortcoming is apparent in the sentence about Jones and his sacrifice. Knowing the basic meaning of the word sacrifice does not suffice if one is reading the sentence. A mechanical, purely language-based approach to the needed knowledge is inherently misleading. A much better way of finding out what knowledge speakers and writers take for granted is to ask these people themselves whether they assume specific items of knowledge in what they read and write. This direct approach proved to be a sounder way of determining the tacit knowledge, because what we must teach students is the knowledge that proficient readers and writers actually use. From people in every region of the country we found a reassuring amount of agreement on the substance of this taken-for-granted knowledge.
We had predicted this agreement. The very nature of communicative competence, a skill that teachers, reporters, doctors, lawyers, book club members, and writers have already shown themselves to have, requires that it be widely shared within the speech community. These are successful communicators, and shared, taken-for-granted background knowledge is what makes successful communication possible. Several years after our compilation of such knowledge was published, independent researchers investigated whether reading comprehension ability did in fact depend on knowledge of the topics we had set forth. The studies showed an unambiguous correlation between knowledge of these topics and reading comprehension scores, school grades, and other indexes of reading skill. One researcher investigated whether the topics we set forth as taken-for-granted items are in fact taken for granted in newspaper texts addressed to a general reader. He examined the Times by computer over a period of 101 months and found that "any given day's issue of the Times contained approximately 2,700 occurrences" of these unexplained terms, which "play a part in the daily commerce of the published language."7
An inventory of the tacit knowledge shared by good readers and writers cannot, of course, be fixed at a single point in time. The knowledge that writers and radio and TV personalities take for granted is constantly changing at the edges, especially on topical issues. But inside the edges, at the core, the body of assumed knowledge in American public discourse has remained stable for many decades.8 Only students who are conversant with this assumed knowledge can understand more specialized matter. This core of knowledge changes very slowly, as sociolinguists have pointed out. If we want to bring all students to reading proficiency, this stable core is the enabling knowledge that we must teach.
That's more easily said than done. One essential, preliminary question that we faced was, how can this necessary knowledge be sequenced in a practical way for use in schools? We asked teachers how to present these topics grade by grade and created working groups of experienced teachers in every region of the country to produce a sequence independently of the others. There proved to be less agreement on how to present the material grade by grade than there had been in identifying what the critical topics are. That difficulty too was predicted, since the sequencing of many topics is inherently arbitrary. While it's plausible in math that addition needs to come before multiplication and that in history Greece probably ought to come before Rome, maybe it's not plausible that Greece should come before George Washington.
We collected the accumulated wisdom of these independent groups of teachers, made a provisional draft sequence, and in 1990 held a conference where 145 people from every region, scholarly discipline, and racial and ethnic group got together to work extremely hard for two and a half days to agree on an intelligent way to teach this knowledge sequentially. Over time, this Core Knowledge Sequence has been refined and adjusted, based on actual classroom experience. It is now used in several hundred schools (with positive effects on reading scores), and it is distinguished among content standards not only for its interest and richness, but also because of the carefully-thought-out scientific foundations that underlie the selection of topics.9
But while the general course of study outlined in the Core Knowledge Sequence (which focuses on history, science, and the arts as disciplines but lacks a reading program per se) is bound to improve reading comprehension in the long term, a great opportunity is being lost when this efficient and coherent approach to the knowledge required for reading is being neglected in the very place where it is most needed—namely, in the long hours devoted to the subject of reading. Since at least ninety minutes per day are currently allotted to reading in early grades, about an hour could be devoted to the language and world knowledge that is most important for competence in listening, talking, reading, and writing. The substantive topics in literature, history, the arts, and the sciences that literate Americans take for granted are deeply interesting and highly engaging to children.
For many years, the great reading researcher Jeanne Chall complained that the selections offered in language arts classes did not provide students with the knowledge and language experiences they need for general competence in reading. She observed that far too much time was being spent on trivial, ephemeral fictions and far too little on diverse nonfictional genres. In the two decades since Chall entered this complaint, little has changed. Most current programs still assume that language arts is predominantly about "literature," which is conceived as poems and fictional stories, often trivial ones meant to be inoffensive vehicles for teaching formal skills. Stories are indeed the best vehicles for teaching young children—an idea that was ancient when Plato reasserted it in Republic. But stories are not necessarily the same things as ephemeral fictions. Many an excellent story is told about real people and events, and even stories that are fictional take much of their worth from the nonfictional truths about the world that they convey.
The association of language arts mainly with fiction and poetry is an accident of recent intellectual history that is not inherent in the nature of things. Older American texts that were designed to teach reading, such as the McGuffey Readers, contained moral tales and historical narratives as well as fictional stories (not that we should go back to the McGuffey Readers, which have many shortcomings). Ideally, a good language arts program in the early grades will contain not only fiction and poetry but also narratives about the real worlds of nature and history. Ideally, such a program will fit in with and reinforce a well-planned overall curriculum in history, science, and the arts. The recent finding that word learning occurs much faster in a familiar context implies that the overall program should stay on a subject-matter domain long enough to make it familiar. As we've seen, such integration of content in reading and subject-matter classes will serve simultaneously to enrich background knowledge and enlarge vocabulary in an optimal way.
That fictional stories can convey factual and moral truths is the traditional ground for defending their value and importance in education. The truth-telling and knowledge-enhancing aspect of fiction is emphatically just as important as the aspect of fiction and poetry that stimulates children's imaginations. The romantic idea that literature should mainly nurture the imagination fits in well with the generally romantic flavor of early childhood education in the United States today. I do not wish to appear in any way hostile to developing children's imaginations. But the second- and third-rate fictions that are too often presented to children in the early grades are far less stimulating to their imaginations than classical stories and well-presented narratives about the real world.
We need to reconceive language arts as a school subject. In trying to make all students proficient readers and writers, there is no avoiding the responsibility of imparting the specific knowledge they will need to understand newspapers, magazines, and serious books directed at the national language community. There is no successful shortcut to teaching and learning this specific knowledge. Those who develop language arts programs at the school level or in publishing houses must understand that the skills they wish to impart are in fact knowledge-drenched and knowledge-constituted. The happy consequence will be reading programs that are much more absorbing, enjoyable, and interesting than the disjointed, pedestrian programs offered to students today.