Before Diagrams

The practice of diagramming sentences first began in America in the middle of the nineteenth century. Set against the full history of grammatical study, diagramming is a relative newcomer. Language has been studied systematically since at least the second century bce in the western world, and even earlier in India. Yet sentence diagrams—visual depictions of the relationships among words—were developed only after over 2000 years of study. From our vantage, the desire to visualize sentence structure may seem like an intuitively obvious move. But diagrams' sudden appearance and their subsequent popularity owe much to a significant shift in methods of teaching grammar that occurred at the same time. Starting in the second quarter of the nineteenth century, three factors conspired to overturn centuries of pedagogical tradition: increasing dissatisfaction among teachers with traditional methods, new technology introduced into the classroom, and, most importantly, a reformulation of syntax that shifted the focus away from the individual word and to the sentence.

In the first decades of the nineteenth century, the market for grammar textbooks was dominated by Lindley Murray's English Grammar, adapted to the different classes of learners (1795). In addition to numerous editions of the English Grammar itself, there were scores of abridgments, expansions, guides, and imitators. Murray's primary method of analysis was oral parsing, a technique copied almost without change from instruction in Latin grammar. A student would be given a sentence and expected to provide a detailed breakdown for each word, naming the parts of speech and the inflectional relations (case, number, gender, tense, as appropriate), and reciting the rules laid down in the grammar book that explained the particular construction. The procedure required a large amount of rote recitation and very little attention was paid to how words were grouped apart from agreement relationships and case assignment. The following extract from Murray illustrates how such parsing was conducted:

"This bounty has relieved you and us; and has gratified the donor."

This is an adjective pronoun of the demonstrative kind. Bounty is a common substantive. (Repeat the person, number, and case.) Has relieved is a regular verb active, indicative mood, perfect tense, third person singular, agreeing with its nominative "bounty," according to RULE I which says &c. You is a personal pronoun of the second person plural, and in the objective case. (Repeat the government and rule.) And is a copulative conjunction. Us is a personal pronoun, in the objective case. You and us are put in the same case according to RULE XVIII, which says, &c. And is a copulative conjunction. Has gratified is a regular verb active, indicative mood, perfect tense, and third person singular, agreeing with its nominative "bounty," understood. "Has relieved" and "has gratified" are in the same mood and tense, according to rule XVIII, which says, &c. The is the definite article. Donor is a common substantive, of the third person, the singular number, and the object of the active verb "has gratified;" or, the objective case governed by that verb. (Murray 1809, p. 220-221)

Such analysis is fundamentally word-based. To use the terminology of nineteenth-century grammar books, it is "etymological parsing." (In the nineteenth century, etymology included what modern linguists would now call morphosyntax.) Relationships between words are specified only in so far as they explain grammatical concepts such as case. So we are told that has relieved agrees with bounty, but there is no consideration of phrasal units such as this bounty. Murray, of course, knew that this is related to bounty. But pedagogically that fact is never emphasized, and there is no recognition that the two words together form a cohesive unit of grammar.[2] What was lacking, in other words, was a systematic conception of a hierarchical sentence structure.

As if to double down on the inherent tedium of the method, Murray and his close imitators presented the abstract rules first. Only after the rules had been committed to memory did parsing, their practical application, begin. As a pedagogical method, this order clearly leaves much to be desired. It is unsurprising, then, that from the 1820s we find a rising chorus of complaints about the traditional methods of teaching. Goold Brown, conservative both in grammatical description and pedagogy, describes the trend while sniffing disapprovingly at it:

This fact [rote memorization], too frequently illustrated in practice, has been made the basis of the strongest argument ever raised against the study of grammar; and has been particularly urged against the ordinary technical method of teaching it, as if the whole of that laborious process were useless. It has led some men, even of the highest talents, to doubt the expediency of that method, under any circumstances, and either to discountenance the whole matter, or invent other schemes by which they hoped to be more successful. The utter futility of the old accidence has been inferred from it, and urged, even in some well-written books, with all the plausibility of a fair and legitimate deduction. The hardships of children, compelled to learn what they did not understand, have been bewailed in prefaces and reviews; incredible things boasted by literary jugglers, have been believed by men of sense; and the sympathies of nature, with accumulated prejudices, have been excited against that method of teaching grammar, which after all will be found in experience to be at once the easiest, the shortest, and the best. I mean, essentially, the ancient positive method, which aims directly at the inculcation of principles. (G. Brown, 1862, p. 102)

The early reformers did not depart significantly from the basic concepts of parsing, but they took great pains to make the process more understandable to their students. Samuel Kirkham's English Grammar in Familiar Lectures (1823) is representative of this drive towards reform. From a modern perspective, his book may appear virtually identical to earlier grammars such as those of Lowth or Murray, whose rules Kirkham follows almost slavishly. But Kirkham was quite proud of his "systematic order of parsing" (Kirkham, 1829, p. 9). He condemns the older methods of teaching, albeit not the grammatical principles themselves, and emphasizes putting grammatical rules into immediate practical use and explaining things to students in conversational terms:

The systematic order laid down in this work, if pursued by the pupil, compels him to apply every definition and every rule that appertains to each word he parses, without having a question put to him by the teacher; and, in so doing, he explains every word fully as he goes along. This course enables the learner to proceed independently; and proves, at the same time, a great relief to the instructer [sic].... The author is, therefore, anxious to have the absurd practice, wherever it has been established, of causing learners to commit and recite definitions and rules without any simultaneous application of them to practical examples, immediately abolished. This system obviates the necessity of pursuing such a stupid course of drudgery; for the young beginner who pursues it, will have, in a few weeks, all the most important definitions and rules perfectly committed, simply by applying them in parsing. (Kirkham, p. 11)

(This page last updated Oct. 17, 2015)


2. This word-based thinking is reflected, for example, in the traditional definition of a pronoun, which states that it replaces a noun. In fact, it replaces a noun phrase, as can be seen by considering sentences such as, "John approached the woman sitting at the bar and asked if he could buy her a drink." Her in this sentence cannot be taken merely as a replacement for woman. It replaces the whole phrase the woman sitting at the bar.