Prerequisites to Diagrams

New Technology

Even with Kirkham's innovations, etymological parsing still appears to be monumentally dull. And yet it did have one pedagogical virtue: it was easily employed in a largely oral environment, where students spoke their answers without writing anything down. At the start of the nineteenth century, equipment we now consider basic to classrooms was largely lacking in American schools. Commonplace books for students to write in were available but used sparingly, given the relative expense of paper. There were no blackboards. Even the slate, a portable, reusable writing surface that had been used in Europe for centuries, was apparently seldom used in American classrooms before the 1830s, or at least was not normally applied to the teaching of grammar. Lyman quotes one William A. Alcott, writing in 1830, who notes that "the idea of studying grammar with slates and pencils was so novel that I found no difficulty in gaining general attention." (Lyman, 1922, p. 148).

This traditional reliance on oral methods for teaching grammar is one reason that diagrams did not appear earlier. Diagrams are a means of visualizing structure. One necessary prerequisite for their use, therefore is a suitable medium for their display. They are unusable in the classroom if teachers and students cannot show them to one another. The major technological innovation that made diagrams possible in the classroom was the blackboard, an oversized version of the slate, which begins to appear in schoolrooms as early as 1818 but was not standard equipment for schools until the 1830s or 1840s (Lyman, 1922, p. 149).

The blackboard let teachers show connections among words in new ways. Lyman quotes James Ray giving advice in 1830 about the use of the blackboard, including an early, ad hoc form of diagramming:

In the study of Grammar the blackboard may be used to exhibit the inflections of the various parts of speech; it may also be used in syntax, to point out the connection of the principal words to each other. The method of doing this is by writing on the board the sentence to be parsed, and then connecting by curved lines those words that have any grammatical connection with each other. The instructor at the same time pointing out what that relation is. It may be observed that in teaching grammar the use of the blackboard is confined to the teaching of the elementary principles of the science, [and] is used by the teacher for the purpose of illustrating these principles. (Lyman, p. 148).

The drawing described here is an attempt to illustrate the traditional relationships of government rather than full system of diagrams, but it's clear that in the decades before the appearance of the first true sentence diagrams, teachers were beginning to hit on the basic technique of depicting syntactic relationships visually.

From Word to Sentence

As has been noted, the word-centered methods of traditional grammar discouraged looking at hierarchical relationships within the sentence. Diagrams, however, are an attempt to illustrate constituency within the sentence. A second prerequisite to diagrams, therefore, was a shift in thinking about syntax to a hierarchical, sentence-based orientation.

Traditional grammar was not wholly lacking in a consideration of sentence structure. The terms subject, predicate, and clause are notably lacking from Lindley Murray's work, but they predate Murray by centuries.[3] Yet the word-by-word approach to analysis dominates these earlier discussions. James Harris, for example, begins his work Hermes with a consideration of the sentence (Harris, 1751), but almost immediately, he turns his focus to individual words and their parts of speech. Similarly, when Goold Brown presents his syntactic parsing, he works through his example word by word, in sentence order. At each stage he describes the syntax by naming a word that it governs or is governed by, or related to. The first part of his sample syntactic parsing in The Institutes of English Grammar is the following:

"This enterprise, alas! will never compensate us for the trouble and expense with which it has been attended.''

This is a pronominal adjective, of the singular number, not compared: and relates to enterprise; according to Rule IV, which says, "Adjectives relate to nouns or pronouns." Because the meaning is—this enterprise.

Enterprise is a common noun, of the third person, singular number, neuter gender, and nominative case: and is the subject of will compensate; according to Rule II, which says, "A noun or a pronoun which is the subject of
a verb must be in the nominative case." Because the meaning is—enterprise will compensate. (G. Brown, 1845, p. 108}

Brown is so captive to the word-and-accidence model of syntax that he appears not to notice that his rule for the subject is circular. It's in the nominative case, he says, and we know it's the subject because it's in the nominative case. But how do we know it's really nominative? English nouns have no distinct forms for nominative and objective cases, after all. Because it's the subject, which is presumably self-evident from the "meaning," a formulation that explains nothing.

From the 1830s, however, writers of grammar books began to incorporate a more systematic analysis of sentences beyond word-level parsing. The ultimate source of this shift was German philology, especially as filtered through the intermediary of the influential Latin grammar of Andrews and Stoddard (1836), which introduced a method of analysis deriving "the rules of Syntax from the logical analysis of the sentence, and its distinct specification of the particular use of each of the several words of which a sentence is composed." (Andrews and Stoddard, p. 4).

English grammars quickly took up the sentence-based method of analysis. Noble Butler, author of the popular A Practical Grammar of the English Language (1846) and one of several authors who explicitly credit Andrews and Stoddard, gives a concise account of the method of analysis he championed:


Tell whether the sentence is simple or compound.
If it is a simple sentence, name the logical subject and the logical predicate.
Name the grammatical subject.
Tell by what the grammatical subject is modified.
If the words which modify the grammatical subject are themselves modified, tell by what they are modified.
Name the grammatical predicate.
Tell by what the grammatical predicate is modified.
If the words which modify the grammatical predicate are themselves modified, tell by what they are modified.
If it is a compound sentence, name the several clauses.
Name the independent and the dependent clauses.
Tell by what each dependent clause is connected to the independent.
Analyze each clause in the same manner in which the simple sentence is analyzed.
(Butler, 1847, pp. 201-202)

Unlike the earlier methods of parsing, which proceed word by word in order, Butler's process encourages a top-down view of the sentence. The student's first job is to divide the sentence into clauses, it there is more than one, then to divide each clause into subject and predicate. By "logical" subject and predicate, Butler means what would, in later works, be called the "complete" subject and predicate. Then, the student analyzes how the subject and predicate, respectively, are constructed.

(This page was last updated October 17, 2015.)


3. The concepts of subject and predicate were introduced by Boethius in the early sixth century (Vineis and Maieru, p. 151), and clause (Latin clausula) appears in the grammatical sense around 1300.