Style in Children ' s Literature : A Comparison of Passages from Books for Adults and for Children

A computer program and syntactic code based on those used by York University in Toronto provided a statistical analysis of the 20,000 words of selected text. I found that the passages from the children's books had much shorter paragraphs, and slightly shorter sentences, T-units, clauses, and words. T-units were the most consistently and notably reduced elements. The children's books also had more lexical repetition and fewer abstract and Latinate words and tended towards a verbal style. These characteristics support some of the common assumptions about children's literature, but the differences were slighter than anticipated. In the area of syntax, the assumption that coordination would increase and subordination decrease markedly in the children's stories did not prove true. Coordination was only marginally more frequent in the children's passages, and subordination nearly equal in both sets. The reduction of prepositions in the juvenile samples seems of more significance syntactically. In the children's passages there are large increases in the amount of dialogue and in the use of Germanic based words.

In the area of syntax, the assumption that coordination would increase and subordination decrease markedly in the children's stories did not prove true. Coordination was only marginally more frequent in the children's passages, and subordination nearly equal in both sets. The reduction of prepositions in the juvenile samples seems of more significance syntactically.
In the children's passages there are large increases in the amount of dialogue and in the use of Germanic based words.
My general conclusion is that the differences in the children's passages reflect a stronger tendency towards everyday speech, that children's authors borrow more conventions from conversation and from oral traditions when writing for a child audience.

Acknowledgements
First, I wish to thank the members of my committee. Paul Arakelian is an ideal major profesor; he encouraged me to explore wherever my research led me and also demanded that I apply rigorous standards to that research.      Three-class frequencies occurring 30 or more times in more than one author's juvenile or adult samples Table 5.11 Three-class sentence openers occuring four or times in more than one set of samples Percentages of different and of unique threeclass sentence openers Table 5.13. Three-class sentence closers occurring more than four times in more than one set of samples        When a writer addresses other adultR he is at eye level with his readers, and can usually use contemporary language.
But in writing for children, he has to wander through his childhood.
He must be willing to recreate who he was .and then find a syntax that will invite his readers in. The questions that concern me are whether books for children are written in a pared down version of mature literary style, in a radically different style, or in a similar style.
There is some consensus among the critics and scholars but also a variety of attitudes on what is appropriate literary language for the young.
In her widely used book, A Critical Handbook of Children's Literature , Rebecca Lukens gives a balanced assessment of the problem. While she states in her preface that "writing for children should be judged by the same standards as writing for adults," she holds that "writing for children presents some special concerns and problems;" these are that because children's experiences and understanding are more limited, the "complexity of ideas" must not be too great and the "expression of ideas must be simpler--both in language and in form" (Lukens 6). Lukens is not, however, arguing for blandness of language as she makes clear in her chapter on style. Here she considers the use of figurative language, word play, parody, and precise vocabulary, and concludes that the successful children's author does "know what he or she is doing with words" (Lukens 124). In sum, Lukens holds that a certain amount of simplification is necessary when writing for children but argues that this is not so ( ma tter of limitation as of precision and a careful much a lection of stylistic options. se . Nicholas Tucker in The Child and the Book is more emphatic about the need for simplicity, saying that a children's author must be "selective when it comes to communicating with his or her audience. The endless paragraphs of a Proust, the convoluted sentences of a Henry James, or the sophisticated, literary English of a Meredith will not get through to children" (12). He holds that "children usually seem to prefer a style that does not present too many difficulties, using a high percentage of direct speech and a less complex vocabulary" (Tucker 13).
He reiterates these criteria later when explaining the overwhelming popularity that British author Enid Blyton's Noddy books have for children from age seven to eleven. He attributes her success to the fact that Blyton ''leads the young reader without faltering from one stock situation to another, described in an equally stock vocabulary" (Tucker 106 Blyton's level of language is best for children's books, but he does advocate a recognition of the child's need for linguistic security when reading. He also holds that an author must consider the child's cognitive stages (similar to those elucidated by Piaget). Piaget said on language comprehension and the school-age child that "When it is a question of adult speech, transmitting or seeking to transmit knowledge already structured by the [adult] language," that "this predigested intellectual nourishment" may not be "assimilated" (39)(40) Many theorists in the field of children's literature follow Piaget's principles. For example, Cullinan, Karrer, and Pillar in their comprehensive survey Literature and the Child rate the books they discuss into the fairly typical, Piaget-influenced categories of nursery (birth to 4), primary (5 to 9), intermediate (10 to 12) and advanced (13 to 15). They, like many, view the simplicity-complexity spectrum as one to be chronologically transversed.
However , they stress (and again this is typical) that each child progresses at a different pac~ through these levels and the ratings are flexible, not absolute.
They also hold that books should inspire the child's growth of language awareness. "Stories told or read to children," they say, "give them opportunities to hear words in use and, in the process, to support, expand, and stimulate their own experiments with language.
.As children learn, books can help at every stage to fulfill their need to make sense of language and of the world" (Cullinan 13 and 16). They view positively the fact that "Through books we learn to comprehend many more words than we actually use" (Cullinan 16 ). To put it briefly, they, like Lukens, do not view the necessity for varying degrees of simplication as an inevitable limitation so much as a natural progression.
They would agree with Tolkien's often quoted statement that children's "books like their clothes should allow f·or growth." Cullinan, Karrer and Pillar also touch on another issue which is relevant to style in children's books, namely, reader-response theory.
For the purpose of this current study, I define literary style as the use of language to create patterns and deviations from pattern that control or enhance meaning. And precisely how this control of language allows the child as a relatively inexperienced reader to interact with the text is an important consideration. As Cullinan and her collaborators phrase it, "When we consider books for children, the child as meaning maker is ever present as the other half of the equation" (8).
A similar concern led Peggy Whalen-Levitt ·to edit a special section on "Literature and Child Readers" in the Children's Literature Association Quarterly (Winter 1980 ). Whalen-Levitt expressed concern that there had not yet been much research in children's literature that used the recently developed reader-response theories, and yet, "To embrace a theoretical framework that enables us to consider author, text, and reader is to resolve a longstanding impasse in our field" (10). She goes on to say, "Implicit, but rarely explicit, One other aspect of style in children's books that is outside the parameters of this study is that of sound patterns, the rhythm and the phonetic complexity of a literary piece. I do not include an analysis of sound patterns, but they are important in juvenile literature.
Children's books are often read aloud, and knowledge of h . practice undoubtedly influences authors who write for t 1s the young.
Also children are very sound sensitive.
According to some theorists, childhood is the "age of resonance," the time when exact imitation of the sound of words in a given language is most natural. And children delight in tongue-twisters, nonsense words, and the heavy patterning of nursery rhymes. Tucker notes that "the way children respond imaginatively to the sound of words, as opposed to their content, is probably the single most unpredictable topic to try to understand in the whole field of children's literature" (13).
It is a topic that has already produced interesting scholarship (for example, Jacqueline Gueron's study, "Children's Verse and the Halle-Keyser Theory of Prosody") and deserves more work, but is not of direct concern in this research.
Stylistic range in children's books is of interest to editors and publishers as well as to critics and scholars.
"The styles of our children's books ·are as varied and eclectic as the books themselves," said juvenile book words potent and few, as they were in the youth of humanity" (Bernay-Isbert 203). Jane Gardam, speaking of Ow n writing, reports that when she wrote her first book her about childhood, she "tried to describe childhood in very bare words and clear colors, and whether I thought I was writing poetry or painting, I don't know" (Gardam 492).
The poet John Ciardi (whose love for words is certainly well established) selected I Met a Man (Houghton 1961) as his "favorite , because I wrote it on a first-grade vocabulary level" to teach his daughter to read (Hopkins 35 Just as simple as bringing them up.
All you do is take all the sex out, and use little short words, and little dumb ideas, and don't be too scary, and be sure there's a happy ending.
Right? Nothing to it. Write down.
This method, she says, may produce an adult best seller, But you won't have every kid in America reading your book. They will look at it, and they will see straight through it with their cold, beady little eyes, and they will put it down and they will go away.
. The Author has not always thought it necessary to write downward, in order to meet the comprehension of children.
He has generally suffered the theme to soar, whenever such was its tendency, and when he himself was buoyant enough to follow without an effort. Children possess an unestimated sensibility to whatever is deep or high in imagination or feeling, so long as it is simple, likewise. It is only the aritificial and the complex that bewilders them.
(VII: 4) Initially this statement denies any condescension to children, but the passage pulls in two directions.
Hawthorne begins by saying he did not consistently write "downward" to the young audience, but the last two sentences suggest that he avoided complexity and wrote the Wonder Tales simply.
In the stories themselves or the Wind," contends that MacDonald is using language to convey the ever-changing possibilities of existence and not merely "l anguage that deals with a stable reality," and that the . g of that book "is simply the importance of poetry as mean in O f knowing" ( 125).
Along these lines, MacDonald . a way . elf makes a specific connection between poetry as a way hims of knowing and the "patois" of childhood: The man who loves the antique speech, or even the mere patois, of his childhood, and knows how to use it, possesses therein a certain kind of power over the hearts of men, which the most refined and perfect of languages cannot give, inasmuch as it has travelled farther from the original sources of laughter and tears • • • . To a poet especially is it an inestimable advantage to be able to employ such a language for his purposes.
Not only was it the speech of his childhood, when he saw everything with fresh, true eyes, but it is itself a child-speech; and the child way of saying must always be nearer the child way of seeing, which is the poetic way.
(Sir Gibbie 138) Although MacDonald did not systematically bring together his ideas on writing for children, comments on the subject are scattered throughout his works.
For instance, he told his stories for the young to his own children, and a record of such storytelling is given in "Papa's Story," where we see a family gathered for a tale that the smallest child present has requested. This child does not want Scottish dialect (which MacDonald did not use in his children's stories).
In the course of the story she questions the word "garments" and is given "frocks" as a synonym, has "mountains" defined as "higher hills yet," and . n an explanation of wind and storm used is give ho ricallY for human passions (MacDonald,Gifts 311 They are almost more in the nature of poems in prose than stories. .After 1886 Oscar wrote very little poetry . • . . This was probably because he thought that he could give rein to his urge for writing poetry more successfuly and more readily through the medium of prose. So that in The Happy Prince and • . . A House of Pomegranates, he adopted a style which was half way between romantic prose and blank verse; this is particularly apparent when the stories ar e read aloud as it is impossible to read them intelligently without a certain lilt and cadence. In the fairytales the emotional side is ascendant, and the baroque predominates over the t ·c The sentences loosen rather than contract, epigramma 1 • as the statistics of this study will show. Perhaps Wilde considered the cynical humor that pervades most of his adults works beyond the range of a child audience. There is humor in most of the fairytales, but it is quiet and downplayed. As an un~igned notice of The Happy Prince in Athenaeum (September 1, 1888) put it, "There is a piquant touch of contemporary satire which differentiates Mr.
Wilde from the teller of pure fairy tales; but it is so delicately introduced that the illusion is not destroyed and a child would delight in the tales without being worried by this application" (Beckson 60).
Wilde, however, has not given us any clear statement of his purpose in writing for children or indicated any 27 modifications he made in his prose style for this audience.
Among the four authors covered in this study, Wilde shows the smallest change between adult and children's passages.
If he did not articulate a theory on writing for children, perhaps it is because he did not feel it was a distinct literary problem.
John Gardner, on the other hand, has given interviews and written copiously on his theories on style, for any b t also specifically on those concerning a audience, u juvenile audience. A pervasive theory in his book The Art of Fiction is that genre is a major influence on the style of a given literary piece. "Most fictional styles are traditional," he says (163). He gives a more detailed description of his notion of style in On Moral Fiction: The idea that the writer's only material is words is true only in a trivial sense. Words conjure emotionally charged images in the reader's mind, and when the words are put together in the proper way, with the proper rhythms--long and short sounds, smooth or ragged, tranquil or rambunctious--we have the queer experience of falling through the print on the page into something like a dream.

(Moral Fiction 112)
The idea of style as mesmeric, as incantatory, is strong for Gardner, as it is with so many authors drawn to write in one of the juvenile genres.

28
Gardner spoke specifically of his writing for children in an interview with Roni Natov and Geraldine DeLuca for The Lion and the Unicorn.
He identified his audience rather specificaly, saying that, except for the story "Dragon, Dragon" (in which a very youngest brother wins by causing the dragon to laugh at his pretensions), his stories are "really meant for kids who have been through fairy tales and are ready for slight variety" (Natov 119).
In the same interview, Gardner commented that the "main h had learned from children's responses to his thing" e "that if you read these stories to children too books was early, they get bored.
If you read "Dragon, Dragon" to a five year old, the kid will go around remembering that verse Dragon, dragon how do you do? I've come from the !_ing to murder you, but the story has too many dead spots for him" (Natov 130).
Gardner insisted, however, in a tape taken by Stephen Banker, that (as Banker paraphrases it) he "applies the same esthetic principles to children's fiction as he does to adults' fiction.
He simply tries to make sure the fiction is available, or accessible to younger minds" (Howell 94). Gardner's concern that his writing is accessible is not restricted to his juvenile audience.
Asked whether he would simplify his work to make it more salable, Gardner replied that he would not, but added, "I 29 would never be consciously difficult" (Howell 100).

And in
Th.e Art of Fiction he advises novice writers that "A huge vocabulary is not always an advantage. Simple langu~ge, for some kinds of fiction at least, can be more effective than complex language" (144). Evidently, one of these kinds of fiction is the fairytale for juveniles.   He also that the "formulas do not touch on organization, regrets d rd er format or imagery in writing" (24) and wor o · ' lud es that "Formulas measure only one aspect of cone style--difficulty," and even this "imperfectly" and therefore are "not measures of good style" (25).
For a specific example of why something as complex ai literary style is best not reduced to formula, consider R. H.

Oscar Wilde
Picture of Dorian Gray -26 _27, "Laughter is not 11 necessary to me.
.blanch at all." 209-11, "He spurred • • . . knife of bone." 95-96, "So it was . • • melt like snow." The King's Indian "King Gregor and the Fool" 158-63, "The afternoon's . • • . . coming, gleaming." King of the Hummingbirds "The Witch's Wish" 37-41, "The witch told .people happy." "The Gnome and the Dragon" 69-75, "As if wearily .them completely." "King of the Hummingbirds" 6-10, "Olaf worked all .afraid of bears." Dragon, Dragon 48-50, "But the mule .for evidence." 44 The size of the individual passages . varied from 450 to 1,000 words, with approximately 5,000 words per author (Hawthorne,6,200 The York code also notes syntactic functions like whether a noun is a subject or an adjective is a post-modifier. My own major expansion of the York system was the inclusion of the text as well as the code in the material to be computer-processed. This allowed me to drop some classifications, although the special nature of my study to add others. caused me Using a recategorized three-digit d ( see Figure 2.2), I kept many of the distinctions and, co e in general, followed the York principles in assigning words to a certain class or function.
For   "T eachers, librarians, and editors," they say, "have come to accept this matter of length as a given" (Donelson 14). How . 1 educators and publishers take the principle of serious Y . for juveniles can be seen in the history of first brevitY grade readers.
In her study of five different editions of Foresman series, Jeanne Chall found "a continuous the Scott, in the number of words used in these readers. decline 55 l. ·n 1920 the number of running words per average story fhereas was 333, by 1962 it had shrunk to 230" .
Primers are the extreme example, but they are symptomatic of the tendency to write briefly for children.
In the next unit considered, the paragraph, some important differences do emerge.
Paragraphs are a purely literary form. Their parallel form in an oral culture is the poetic stanza, which is clearly marked by sound patterns and may, but need not, be a division by content or ·thought. The paragraph, on the other hand, is an intellectual, visually perceived unit, a kind of spatial punctuation that allows for a high degree of stylistic manipulation. Paragraphing in books for very young readers is often emphatic and precise.
In picture books each such unit is frequently marked off by an illustration, or several subject-connected paragraphs are grouped with the relevant picture. Paragraphing, in the hands of a skillful author, especially an author-illustrator, can become an art form in itself. Graham Greene noted that Beatrix Potter's paragraphs "are fashioned with a delicate to complete a movement, but mutely to criticize irony, not . n by arresting it. The imperceptive pause allows the actio t k · n the picture" (Egoff 293). the mind to a e i Other factors also make paragraphing important in children's literature.
For instance, the conventional journalistic wisdom is that short pargraphs aid readability.
And the high incidence of dialogue in juvenile fiction, with its concomitant convention of paragraph per speaker, further increases the relative number and relative brevity of paragraphs in many juvenile works.
In the use of the 56 paragraph, then, whether by authors or the editors of books, differences between the longer, supposedly more complex units for adults and the shorter, supposedly simpler, more digestible blocks of prose for children should be highly visible.
The evidence for the four authors does indicate that brief paragraphs prevail in children's literature. As Table   3.1 clearly shows, there is, for three · of the four, a marked difference between the length of paragraphs for adults and the length of those for children. Only Wilde has a minimal difference, an average 69 words per paragraph for adults and 61 for juveniles. Those under•20 words are designated as short because, according to Cluett ' s study, the average sentence l ·n literature over the last two centuries has hovered 1ength around 20 words.
Any paragraph shorter than an average seems quite brief. The figure of 100 words or more sentence for long paragraphs was arbitrarily chosen. And the youth answered, and Tangle heard him, though not with her ears:--"I am.
What can I do for you?" (Golden Key 56) The dialogue is deliberately separated from its tag, the space adding further pause (and emphasis) after the already double punctuation of colon and dash. MacDonald also uses this paragraphing technique in one of the other passages analyzed to put emphasis on the importance of the golden key: The first words the lady said were,--"What is that in your hand, Mossy?"  Gray and the entire stories "The Happy Prince" and "The Young King" (almost 9,000 words total) yielded an average sentence length of 12 for the adult novel and 19 for the fairy tales.
A quantification of Wilde's entire opus might reveal a consistent tendency towards longer sentences in his children's works.   The mass was chiefly made up of the shadows of leaves innumerable, of all lovely and imaginative forms, waving to and fro, floating and quivering in the breath of a breeze whose motion was unfelt, whose sound was unheard.
(Golden Key 38) · "floating" quality would be lost in shorter The eerie, sentences. Or consider the longest sentence (88 words) in his children's samples: She was standing at the foot of a tree in the twilight, listening to a quarrel between a mole and a squirrel, in which the mole told the squirrel that the tail was the best of him and the squirrel called the mole Spade-fists, when, the darkness having deepened around her, she became aware of something shining in her face, and looking around, saw that the door of the cottage was open, and the red light of the fire flowing from it like a river through the darkness.
(Golden Key 29-30) 63 A modern editor for a juvenile department would probably itch to dissect that sentence into several, and yet it is both easy to follow, due to its temporal sequencing, and a successful blending of content and form. But that speculation to the side, Gardner follows the pattern 3 3 shows for Hawthorne and MacDonald--shorter t Table · tbS.
te nces when writing for children--though with less sen than those two have. Gardner varies from the difference his increased use of sentences over 30 words in pattern in ·uvenile passages (18.8% as compared to the adult's the J 14 • 9 %) and as the only author to have his longest sentence occur in the children's sample.
This 109-word sentence is, g rammatically five sentences separated by however, semicolons: The ants on whom he had refrained from stepping came and paraded by while he worked; the mice he'd fed cheese came and polished the copper pots by rubbing their backs against them; the owls he'd allowed to roost on the rafters flew down to him and fanned Olaf's fire with their wings; the wolves he'd allowed to hide under his bench when there were hunters about came and helped him line up the pots when he'd finished with the mending; and the huge burly thieves he'd allowed into the cellar when they escaped from the sheriff (who'd gotten trapped in conversation with the mayor) sang him barbershop quartets. The adult sentence that tops 100 words is syntactically complex: He'd been wanting, as he walked slowly through the palace, thoughtfully stroking his beard with his right hand and swimming with his left, bowing to his brave and gallant knights as they swam by arm in arm with their elegant ladies, or throwing a word of encouragement to some elderly minister who was puffing hard and looking very doubtful that he'd make it as far as the safety of the stairs--he'd been conscious of wanting to embrace them all, both the beautiful and the ugly, and cling to them as a sweet uninhibited child clings for dear life to his parents.
("King Gregor and the Fool," The King's Indian 160) The sentence breaks at the dash and repeats the subject with a variation on the main verb, so one could argue that there 66 are 2 T-units (72 and 31 words long), but as the predicate is not completed until the end, the sentence can also be regarded as one very long T-unit.
As Kellog Hunt's study (1965)  '    In his adult passages Hawthorne has "Adam," "Bunyan," "Camus," "Eve," "Gothic," and "Tophet." MacDonald refers to the dawn as "Aurora." Osca:r Wilde, whose children's samples so often prove the exception in this study, has 8 allusions of this sort in his adult samples ("Antinomianism," "Bacchante," "Christ," "Darwinismus," "Gretna Green," "Omar," "Sarani," and "Silenus") and 5 in the children's ("Adonais," "Bithynian," "Endymion," "Hadrian," and "Narcissus"). Of course, Hawthorne's children's tales are filled with Greek names, usually unfamilar to children, but they are in the stories and explained in the text. characters The next consideration, the relative levels of abstraction in the two genres, presents a difficulty. As p.J. Gillie  any of the readability formulas, numerical scores may imply an unwarranted degree of precision" (215), we can Table 4.4 that the four authors are, in all samples, see in C oncrete end of the scale, ( 1 sample is "standard," 00 the "fairly concrete," and 3 are "concrete"). This is to 4 are be expected in works of fiction. The formula-derived differences between the adult and juvenile passages are not very great for any one author.
Hawthorne's child and adult passages both are "fairly concrete," MacDonald and Gardner vary only one category each, and Wilde misses the same "fairly concrete," "concrete" ratings by only one point. Notice, however, that there is an increase in the "concrete" character of each of the samples from children's books, and that in one factor, abstract nouns noted by suffixes, there is a very marked decrease in the juvenile samples, the ratios being 3 :1 for Hawthorne, MacDonald, and Wilde and 2 : 1 for

Gardner·
These findings indicate a slightly lower level of t a ction in the juvenile selections and suggest that one abs r more assumption about the style for young audiences may be true.
A still somewhat subjective but more easily measured quality in literature is the presence of absence or words with negative meanings.
It is usually assumed that children's books should present an encouraging view of life, more optimistic .than that of mature literature with its tendency toward irony and its frequent acknowledgement of the tragic. So one would expect to find a lack of negative language in children's literature. When the standard negatives in the language (no, not, never, neither, nor) were considered, this did not prove true.      negative terms seem to be powerful ingredients.
There is another class of words which, unlike negatives, we expect to find in children's books--easy words. One of Klare's criteria for easy words was "words learned early in life" (Klare 19 and pets are of great concern for them. Another reason for including pets or animals is that Spache lists "animalness" as one of the positive elements of style for young readers (Spache 15       In every case in Table 4.11 there is a decrease in 95 nouns and an increase in verbs in the . juvenile samples, and the modifiers connected with these two classes follow suit. But what does less nominal or more verbal mean in terms of style? Rulon Wells, in his article "Nominal and Verbal Style, concluded that "A nominal sentence is likely to be longer, in letters and in syllables, than its verbal counterpart" (Freeman 301).
A good example of this can be I ;i · Joseph Williams' rhetoric, Style: Ten Lessons in found in ciaritY and Grace. Williams begins with the problems created by over nominalization and compares these two sentences: There will be a suspension of these programs by the Dean until his reevaluation of their progress has occurred.
The dean will suspend these programs until he reevaluates their progress. For instance, Don P. Brown and ver a biS co-authors of a composition text claim that when "a .
of writing seems unclear, you will often discover piece 97 that it has a low verb density" (Brown 51).
The range they defined ran from 11% verbs for low density to 20% verbs for high density.
(Their count included finite and auxiliary verbs.) Cluett's verb count included these and participles, gerunds, and infinitive~, and he concluded that "as a style reaches 18.5% in its total verb items, it tends to become a verbal style" (Cluett 74 Furthermore, when the "M" statistic is split into word Se s changes in the percentages of adjectives and cias · ' adverbs do not to prove much either. The decrease in adjectives and increase in adverbs in the children's samples seem to reflect little more than a correlation with the decrease in nominals and the increase in verbals. The ratios of noun to adjective and verb to adverb are almost identical in the juvenile and adult samples. The nominal-verbal proportions seem to be the determining factor here rather than any choice to be more or less descriptive.
With an average of 15.3% modifiers in the adult samples and 15% in the juvenile, the four authors are, however, more prone to use modifiers than many of the 50 authors in Cl~ett's study.
The  Note that, even excl uding th e neuter plural (because "they" can have either p ersons or objects as anteceden t), we see a heavy use of human referent pronouns in all the samples. • yet can never be certain," and in "King Midas" we have, "a very pretty piece of work as you may suppose." But these account for only 4 of 38 you's in all the samples. The rest are in dialogue, one speaker addressing the otper as "you." In his adult passages, Hawthorne uses "reader," "whoever," and "we" to draw the reader into the text. Such direct address is reputedly common in children's books, but this impression is probably due to the popularity of Victorian juvenile literature, which shared with its adult counterpart a proclivity for addressing the "dear reader." Hawthorne, as we have seen, and MacDonald employ it in both adult and juvenile writing. The use of you is one to rical trick for involving the reader. rhe It may carry weight this way even when the you is a character some ~spo ke n to in dialogue.
In this sense the children's samples may be slightly more personal than the adult ones, but pronouns do not change in any major way between the two sets of samples.
To sum up the meaning of the many vocabulary related factors that have been considered, let us take them in ascending order of their importance. The least significant factors found in this study seem to be the incidence of foreign words (too few to allow for any conclu~ion) and the use of negative function words (which revealed no pattern of change from adult to juvenile). Neither did the "M" statistic show any consistent tendency towards more or less modification in children's books.
Some factors that suggested patterns, but with far from conclusive statistics, are total negative words (where the ratio was 2:1, adult to child. However, the average difference was only 1.2%. The word little occurred 37 times in the children's samples to 13 times in the adult and may well be, as it is often designated, a "children's word." But other size words were not so clearly Therefore, the study shows that vocabulary in children's books is likely to favor concrete words in Anglo-Saxon based English, repeat words more often than adult prose usually does, and tend to a verbal style and a corresponding high usage of adverbs. Other factors, such as words relevant to childhood, intensifiers, diminutives and exclamations, may increase slightly, but the evidence for this is inconclusive in this study. A literary dialect of childhood is only partly defined by this look at the vocabulary of sample authors.

Syntax for Juveniles
How Complex?
The changes in the norms of syntax are as significant as the ones in vocabulary, if not more so.
(Robert Cluett, Prose Style and Critical Reading, p. 258) When words are considered not for their individual content but in their grammatical relations to each other, 107 we are dealing with syntax, a word from the Greek, meaning "to arrange together." There are many ways to analyze an author's use of the grammatical arrangements a language offers, but in this study I give most attention to those syntactical choices that can be readily measured statistically, things like the incidence of function words (prepositions, determiners, connectives) and distinctions within these and other word classes.
The computer program also counted the frequency rate of three-class patterns  Pronominal tags may influence the pronoun count in the children's passages, but Gardner has the greatest increase in dialogue and his pronouns decrease. Only in Hawthorne's passages is there much correlation between relative amounts of dialogue and relative pronoun usage. Perhaps the high incidence of you is also a factor. Whatever the explanation, we are left with an increase in pronouns in half of the children's passages and a decrease in the other half.
Investigation of the subject has convinced me that oun proportions have very complex causes. Cluett, for pron instance, finds Hemingway's style pronominal and feels that this contributes to its reputation as simple plainstyle (Cluett 152)    A slight increase in the proportion of determiner to noun may also exist, but here the evidence is less conclusive.
Prepositions, those small but powerful function words have been ignored or down-played in many stylistic studies.
Josephine Miles did not include them in her count of word classes in Poetry and Change or in her earlier studies.

Charles Fries gives them less than a page in his book The
Structure of English, and yet Cluett notes that the sequence preposition-determiner-noun is the most common in the language (Cluett 68). As a word class, the preposition ranks second, third, or fourth in usage in the sampl es of my study (behind nouns and/or pronouns or determiners), and Table 5.3 shows that these four authors are quit e typical in the number of prepositions they use: 64% of the 50 I I authors that Cluett tabulated fall between 9% and 13% usage for prepositions .
Here, only Gardner's children's sample falls below this range. proportions (see Table 4.11).
Of more interest for this study is the consistent pattern of decreased prepositons in the juvenile samples. The reasons for the reduction seem multiple.
It is, after all, impossible to judge whether fewer nouns caused the reduction or vice versa. An 11 in modifiers could also explain the reduction increase because the information in most prepositional phrases can be expressed by an adjective or adverb, but Table 4.11 recorded a slight tendency to reduce modifiers as well as 115 nouns. Therefore, some other influences must be behind the decrease in prepositions.
For one thing, regarding the characteristics of a dialect of childhood, the findings of several studies show that immature speakers and writers do not use the prepositional option as often or as the mature do. We have tended to think of modification by subordinate and relative clauses as the area of syntax difficult for children, but research has called this into question.
Walter Loban includes prepositional phrases among "those syntactical strategies for classifying thought relationships" (12). He and others suggest that the use of genitives, adjectives, verbals, and prepositional phrases may more accurately signal a mature style than the use of dependent clauses.
Menyuk, studyitig language in very early childhood, reports in her monograph Sentences Children Use that prepositions begin to appear at about age 2, usually lumped under the sound "uh" (grammatically though not phonetically distinguished from the indefinite article at this stage).
Prepositions of place appear from ages 3 to 9, but the greatest overall increases and most frequently significant increments from grade level to adjacent grade level were found in the use of adverbial infinitives, sentence adverbials, coordinations within T-units, and modifications of nouns by adjectives, participles, and prepositional phrases.
In the theory of transformational grammar, all these constructions are explained as being produced by application of deletion rules.
(O'Donnell 90) They go on to say that, although the amount of subordination has long been used to calculate syntactic maturity, their findings call into question the sensitiveness of this measure. They found, besides the high early incidence of relative clauses, that "Nominal, 118 adjectival, and adverbial clauses were all used quite often by kindergarten children, and none of the types was employed in speech in any grade at a rate significantly higher than in the grade below'' (O'Donnell 98). Therefore , the fact that Hawthorne, MacDonald, Wilde, and Gardner did not make major changes in their choice of connectives when they wrote the children's passages may not be as surprising as it initially seemed to me.
By looking carefully at Table 5.4, "Types of Connectives and their Percent of the Total Words''; Table   5 .5, "Frequency of Types of Connectives," and Table 5.6, "Percentage of and Placement of Subordinating Elements," which are grouped together, we can gain an overview of this complicated matter.    (1962), Loban (1963), and Hunt (1965) which "found frequencies of main-clause coordination to vary inversely with advance in grade level" (O'Donnell 21).
In regard to the language in literature for children,  (265), and, Cluett notes in his study that "prior .was to 1825 a · writer with fewer that 5% coordinators •• the exception'' (227). He speculates that the modern decline is connected with the shorter sentence and the decline in "formal parallelism and in the copia that is often associated with it" (Cluett 227 (68) and  found the use of the infinitive developed early and seemed to pose no problems of easy acquirement.
As   After repeating the percentages of finite verbs in the total samples (already given in Table 4.11), Table 5.8 then gives the percent that each subdivision is of the finite verbs. Those classes that showed a fairly consistent pattern of change are the passive form, the auxiliary will, modal auxiliaries, and the ~rogressive tense. Total auxiliaries decreased in 3 of the juvenile samples.
The first of these findings, that the passiye form is Note also that when each author is compared with himself, adult against juvenile passages, the differences in sample size must be taken into account.
These caveats stated, the results appear in Table 5.9 So the evidence on syntactic variety is inconclusive.
Neither does a perusal of the three-class frequency lists given in Table 5.10 offer much information relevant to this study. The table gives only those sequences which appeared 30 times or more in more than one author in the juvenile set or the adult set.   Table 5.11.        Table 5.14 shows.

Conclusions
There is surely no doubt that the child's achievements in systematizing linguistic data, at every stage, go well beyond what he acutally produces in normal speech.  The items are marked with one of three symbols: "+" or "-" or "?". The plus mark designates findings for which the diff e r e nce b e twee n the adult and juve nile samples i s greater than 10% of the figure in the adult column; the minus symbol designates cases in which the difference is less than 10% of the figure in the adult column; the question mark designates borderline differences where the significance is problematic, for example, too close to 10% to allow for a margin of error or concerning small classes where further statistics are needed to confirm the data.  The hypotheses on length can be treated briefly.
150 Table 6.1 shows that the most marked differences are in average paragraph lengths and average T-unit lengths. Two very opposite forces seem to produce these differences.
Paragraph size is a highly conscious and easily e dited aspect of composition. T-units, on the other hand, can be masked by punctuation. They represent syntactic rhythms that are less consciously but more consistently us e d by an author than those of sent e nc e l e ngth. As T-uni ts a re th e basic unit that must be comprehended as interrelated syntax, their reduction for a child audience s eems natural, and T-unit l e ngth was r e duc e d in all f our childre n's    Word classes were discussed in · some detail when they were covered in Chapters 4 and 5. Notice that the averages in Table 6.3 confirm only two of the hypotheses: the increase of definite determiners and the decrease of prespositions. Pronouns and modifiers, which had been projected to increase, also remained virtually the same. Across the board, word-class distribution did not prove a good distinguisher between children's and adult literature, nor did the syntactic patterns which the program calculated.
As Table 6.4 suggests, no major syntactic differences in word order or its amount of variation emerged as typical of the children's genre.
I I There is very little change in the amount o f syntact i c vari e ty. The "D" statistic ave rage shows none a nd th e authors split on the matter, Hawthorne and Gardner ' increasing theirs very slightly in the juvenile passages, MacDonald and Wilde decreasing theirs (see Table 5.8). The frequency statistics on sentence openers indicate tendencies to begin more sentences with connectives and to avoid the Preposition-Determiner-Noun sequence as an opener when writing for children, but these need further confirmation. Sentence closers (given in detail in Table   5. It can be seen, therefore, that the arrangement of word classes into specific syntactical patterns is no more susceptible to major changes between the two genres than the statistical distribution of those classes is. There are some differences, but they are not nearly so l~rge or consistent as those found in vocabulary and length. The work of the last decade or so on children's use and comprehension of language suggests that syntactical simplification in children's books may not be necessary in any great degree, even for quite young children. Menyuk found, for instance, that "All the basic structures used by adults to generate their sentences were found in the grammar of the nursery school group" (Syntactic Structures 298). These developing theories ~bout juvenile language assimilation should be taken into account by authors and I editors for the young. Routine oversimplication of children's books could actually inhibit language acquisition. In his seminal work, Language: Its Nature, Development and Origin, Otto Jespersen, speaking of the way small children acquire language, describes it as "The 'little language' which the child makes for itself by imperfect imitation of the sounds of its elders" (106). He is considering sound here, not syntax, and finds it very imperfect, "meaningless babbling" of "long strings of sounds" (Jesperson 108), yet there is an analogy. It is 1 I because adults speak to children in an established language I that the child ceases babbling and forms words. If adults babbled back, children would never learn. The same holds true for syntax. Children will not perfect it without models. Children's literature should not lag behind its readers' abilities. Apparently serious authors for children (among whom are the four covered in this study) do not severely trim their syntax.
What children cannot handle is too much new information at once in the form of too many strange words.
Words do not represent an interconnected set of rules that can be internalized. Semantically, words are arbitrary and individual and must be learned individually over time.
Therefore, it is not surprising that some of the greatest differences this study found between the two genres were in the range and type of words used.

Alan Garner, in his address to the Tenth Annual
Conference of the Children's Literature Association, spoke on this matter of word choice when writing for childre n.
He estimates that children "by the age of five, use about two thousand words, by th~ age of nine, six thousand (or eight thousand, if encourag e d to r e ad). By the ag e o f twe lve, the child will have a voc a bulary of twe lve thousand words" (Garner 7). This, he notes, is one-third of his own (and a typical writer's) vocabulary. Does this d i screpancy bo ther him whe n he write s for childre n? He says, My experience, over twenty-seven years, is that richness of content varies inversely with complexity of language. The more simply I write the more I can say. The more open the prose as the result of clarity, _the more room there is for you, the reader, to bring something of yourself to the act of translating the story from 161 my subjectivity to your own. • • • The reason why I have no dilemma over choosing the one shared word in three is that the vocabulary I use in writing is almost identical to the twelve thousand words of childhood and of most adults. They are the words of conversation rather than of intellectual debate; concrete rather than abstract; natural rather then imposed; Germanic rather than Romance.

(Garner 7)
One phrase especially of Garner's may hold the key to style in children's books: "words of conversation." The Germanic, natural, concrete qualities of juvenile prose may ve ry well flow from the fact that children's authors try more consciously than authors for other audiences to sound as if they were talkitig. Two reasons make this an effective strategy for children's books. For one thing, a conversational style is easier to follow. One study by F.E. Engleman found that fourth to seventh graders "preferred factual content written in conversational style and read it faster than narrative expository style" . Another reason is that children's literature r e mains more closely linked with an oral tradition than adult literature now is.
It keeps many of the conventions that create a bond between teller and listener in the oral tale. Perhaps the increase of the pronoun you found in this study reflects this. Also children's books are often intended for reading out loud which means that, if successfully composed, they will sound natural and allow for exchange between reader and listener (as between teller and listener in an oral culture).
Walter Ong, tjistinguishing between orality and literacy in his book The Interfaces of the Word, holds that, "It is at least likely that in some way a child in technological society today passes through a stage something like that of the old oral culture," but he adds, "only somewhat like the old, for it remains a child's stage and cannot be protracted into adulthood'' (299). He is connecting the oral tradition with formulaic repetition rather than casual conversation, but the distinctions he makes between natural "mother" or "native" languages, picked up by mouth in infancy, and "male languages," the intellectual ones (like classical Greek and Latin), learned at least partially by eye, is important here. Almost all of the charact e ristics that proved typical of th e children's samples in this study are also characteristic of oral language: brevity of units, lexical repetition, and concre te basic words.
As far a length is concerned, we do not normally speak in long involved sentences, and the interactive nature of conversation, speakers responding to, interrupting each other, inhibits long stretches of monologue (the equivalent of lengthy paragraphs). We repeat words more often in speech than in writing, and, unless the subject is technical or academic (tainted by our literacy Ong would have it), we tend to converse in concrete words which are basic to the language. We do not, however, when conversing, tend to concretize by means of elaborate modification--strings of adjectives, adverbs, or prepositional phrases. And these did not prove characteristic of the children's passages either. Because speech must exist in time, simple sequential linking by coordinators may be more likely, but all types of conjunctions (most are from the Germanic base of the language) are natural to English. Also note that in the children's samples there is a marked ·increase in dialogue, which is a deliberate, direct imitation of oral language, of conversation. Early in this study I mentioned that sound is an important element in children's literature.
This aspect of style was not analyzed directly, but the conversational "sound" of children's books has, in a sense, been indirectly assessed by adding together the more significant statistics of this study.
Writers who use a conversational tone as an invitation to the reader may, when aware of a specific audience, appproximate the level of diction they think that audience will respond to. When the audience is children, the dialect of childhood is an appropriate choice. I have been trying to define this dialect throughout this study. It is, it seems, "the words of conversaton,'' as Alan Garner notes, and shares much with everyday adult speech. If the differences found between the two sets of samples are not that great, perhaps it is because they need not be. The qualities of the adult samples that differ most from the qualities of the children's (greater lengths, less repetition, Latinate vocabulary) mark adult genres as more bound to literary than to oral conventions.
One last issue should be raised. The study shows a large syntactic range in the juvenile genre, but a more limited vocabulary. The implications ·of this limitation are diverse and not necessarily negative. There is, as we have been discussing, the naturalness and clarity that e veryday words can foster. Beyond this, authors and critics have noted that some of the stylistic features of juvenile prose can be related to those of ten associated with poetry. As George MacDonald put i t, child language is ' I I close ~o the child way of seeing things--the poetic way. That children can appreciate this poetic approach to language is confirmed by experiences like those of Kenneth Koch who spent some time teaching children to write poetry.
Describing the success of the experiment, he says, "Treating them [the children] like poets was not a case of humorous but effective diplomacy as I had first thought; it was the right way to treat them because it corresponded to the truth" (Koch 29). This was not merely because of their sensitivity to words. On the use of repetition, Koch says, "Repetition is natural to children's speech. It l ef t their poetry free for the kind of easy and spontaneous music so much appreciated by contemporary poets" (21