Textbook Analysis: A Refresher Course

by Gareth Morgan, d. 1996

Texas Classics in Action, Summer 1997.garethfx.jpg (19335 bytes)

[In December of 1968 Gareth Morgan published through the Texas Classical Association a pamphlet entitled "A Consumer's Guide to Latin Textbooks." It was designed to be exactly that: a consumer's guide--complete with charts and ratings comparing the five textbooks up for adoption at that time. As we now face another round of textbook adoption here in Texas, I thought it would be beneficial to reprint edited highlights from this pamphlet. --Ginny]

Consider how many people have learnt their own school Latin from one teacher--a teacher who was very often the only Latin teacher in the school, and perhaps quite isolated from other Latin teachers anywhere. When these students went through a college career it was quite possible for them to have completed several courses without ever having considered the way in which they learnt the basis of the language. They can go into their own classroom and reproduce their own training without having encountered the slightest suggestion that there is more than one way of teaching beginning Latin.

When such a teacher meets this suggestion, the encounter may be traumatic; especially if it comes at a stage when teaching problems seem to have settled themselves, and an easy efficiency has been attained. One advantage of a textbook adoption may be that it makes the teacher consider again whether his efficiency is really an efficiency at doing the wrong thing.


There are two important points of principle that have to be decided by the author of a Latin course. Is his presentation going to be "preceptive" or "inductive"? And is his presentation going to be "vertical" or "horizontal"?


A "preceptive" book will present rules first, and then use reading matter to illustrate and practise the rules that are given. An "inductive" book will present the reading-matter first, and use examples from it to formulate (and, if possible, lead the student to formulate) the rules.

Two examples:

(a) In teaching prepositions, the preceptive book will state that they govern an accusative or an ablative. (The exact grammatical terminology is not relevant.) It may make the point that in can govern both. The reading-matter then has a plentiful supply of prepositional phrases.

The inductive book begins with a narrative that includes prepositional phrases. The prepositions may be treated as vocabulary items, or they may be introduced in contexts which combine with the familiar shapes of, e.g., in, super, contra to make obvious the meaning. Then, after the passage is read, the phrases are used to induce the observation that prepositions are followed by accusatives or ablatives, and the point about the dual meaning of in may be elicited, or stated. (This depends largely on the skill of the author's narrative.)

(b) The preceptive book might introduce relative pronouns with discussion of the way in which sentences can be linked together, and follow this by the rule about the agreement of antecedents. The reading matter would have examples.

The inductive book would treat qui as a vocabulary item meaning who, which, that. The reading would introduce the pronoun in various forms. Then the question would be asked (if it had not naturally arisen already), "Why do the forms vary?" and the antecedent rule would be framed on the basis of what had already been met.

This leads to a very important point in the selection of your textbook. It is obvious that the preceptive teacher can use an inductive book, and the inductive teacher can use a preceptive book, by the simple process of making their pupils read page 102 before reading page 101. The same applies to almost every point mentioned in this study. Any teacher can use any book in any way, if he really wants to. The good teacher may well make his selection by asking, "Which of these books least gets in the way of what I want to do?"

Again, a teacher who prefers inductive methods might be quite happy in teaching prepositions that way, but feel that a quick look at a paradigm of qui would be of advantage before reading examples of it. His ideal might be a chapter that began with reading matter that used qui intelligently and fluently in all the easy cases, then induced a rule and introduced a paradigm, and ended with a reading that used cuius, cui, and quae. The good teacher becomes an eclectic, and will never have the perfect textbook. He ought at least to have one with a lot of good material that fits generally to his patterns of teaching.

In practice the inductive method has been found to be by far the more profitable in language-teaching; and if the publishers' submissions and the committee's selections had adhered to the original specification calling for inductive books, much of this discussion would be hardly necessary.


Probably most teachers would agree that the conception of case is the greatest difficulty in the early stages of Latin. Fifty years ago there would have been little doubt that the way to start would be to take the First Declension, and go down the column, learning such meanings as "O table," "of the table." "to the table." (Sir Winston Churchill's memoirs have a graphic description of precisely this introduction to the language.) When the pupil was thoroughly versed on the first declension, he would proceed to the second, the third, and so on.

If the declensions are visualised as five parallel columns, this may be called the "vertical" approach.

Then came the realization that the vertical approach had some great drawbacks. It lead to the meaningless chanting which is all that remains in the memories of so many people who studied the language. It presented the cases so close together that this difficult idea did not have the maturing-time necessary for assimilation. It confined pupils for quite a long time in the early stages to a vocabulary of women, water, poets, sailors, farmers, and maps, to which the more desperate (or ambitious) composers added cows, wasps, monkeys, crowns, whales, and erasers.

It was also realised that a horizontal approach was possible, in which one case at a time was introduced in all five declensions.

If you thought the Accusative Singular was sufficiently marked by -m, and the Accusative Plural by -s, you could, even in the first stages, present Latin of considerably more fluency, sense, and interest. Students who met the cases one by one were more able to assimilate the difficulty of the case idea.

The advantages of the horizontal system as far as the reading of Latin is concerned are beyond question. The only argument which can be seriously brought against it by the "vertical" teachers is that it may lead to inaccuracy in writing Latin: even that is doubtful.

A similar, but less important distinction may be made with verbs. A vertical treatment would handle three, or even six tenses of the first conjugation before going on to the others. Horizontal treatment deals with one tense in all four conjugations before moving to another tense.

So far, we have been concerned with the description of the principles of the books. The question of order of presentation, that we have dealt with in the last section, could be extended to all the forms and patterns of the language. Should enclitic -ne appear early or late? Early, if you are using practically any form of oral technique; late, if your exclusive concern is to read prose authors by traditional methods. How should the subjunctive be introduced? With the independent forms if you genuinely feel that your students will be helped by perceiving a general subjunctiveness in the use of the mood in classical Latin; with indirect commands and purpose clauses if you are more concerned with the frequency of use in what they are going to read. These questions, and many more like them, cannot profitably be treated in this sort of guide. They are mostly points of personal preference, provided there is some sort of coherence in the order of presentation, pupils' achievement would probably not be significantly different from one to another.


It is quite feasible to learn to read Latin by the process of reading Latin. This was done by the mediaeval student who started his studies with "Arma virumque cano . . ." and went on through all twelve books, learning each new form and each new word as it appeared. It is done in a rather easier way by the recent [in 1968] Danish book that presents a continuous narrative developing from simple two-word sentences on page 1 to fluent Latin at the end.

These methods require rather sterner resolution than most students possess. We live in a convention where highly-structured "methods" and "exercises" are used to reinforce the material that we learn.

The books that we have before us differ so much in methods that they must be described separately. Perhaps the most profitable way is to move from "traditional" to "modern."


The traditional practice is based on the pair "translate-compose." It goes without saying that the reading-passages, or "narratives" in textbooks are available to be used for translation, and will be so used by the majority of teachers. In a different category comes the provision of sets of sentences for translation into English. This is so much a part of most teachers' training that it is difficult for us to look at it critically. Yet it is possible that in the form in which it is usually found, this is not a profitable strategy. There is obviously little aesthetic merit in presenting artificial sentences with no context, while if the object is to produce mechanical repetition of certain patterns, stricter drills which exclude the possibility of "incidental error" may be more advantageous.

If, for instance, the immediate concern is to practise the recognition of the Dative Singular, a set of sentences containing various Datives may be an inefficient teaching-method just because the student is slowed by the need to understand the rest of the sentence--the subjects, verbs, prepositional phrases, and anything else the author has provided. The most intensive practice may come from such patterns as Caligula pecuniam . . . dedit, where the gap is filled in succession by such words as plebi, equo, poetae.

This is not the whole story, because such "pattern-drills" clearly lack three things which good sets of sentences can provide--variety of material, variety of word-order, and the possibility of cultural interest. It does mean that if sets of sentences for translation are to be justified as an improvement on pattern-drills, at least some of these three elements must be present.

Composition is the other half of the traditional duo. It is obviously a hangover from the centuries when writing in Latin was a social necessity, but it has come under very heavy attack in recent years. (It has virtually disappeared from the secondary curriculum in all countries except Britain, Holland, and America; has become an optional subject in many British school and university examinations; and has all but vanished from American national examinations.)

There are two main arguments against composition. The first is that it severely reduces the amount and scope of the Latin one can cover in a given time. In the time taken to compose ten words of Latin accurately, fifty or a hundred could have been read. This argument is very hard to counter. The second is that composition tends to produce and reinforce incidental errors. If a student has just learnt ablative absolutes, and is asked to translate "When the signal was given, the Romans advanced," he may well take the first step towards success by writing "Signo dato . . ."; but the reinforcement will be lost if he has immediately to be diverted to a correction of ". . . Romani procedaverunt." Again, if he has written procedaverunt one evening as part of his assignment, looked it over in the school bus in the morning, put it up on the board during roll-call, and stared at it for five minutes until class begins, by then the word is pretty firmly impressed on his memory (and perhaps on that of his unfortunate classmates). It requires strenuous correction (during which a bad teacher is quite likely to repeat the word, contemptuously, two or three more times) before a truce is achieved.

Yet Roman writers never wrote procedaverunt. In this sort of case, the composition-method erects traps for the purpose of setting students free.

Against these powerful arguments few Latin counterclaims can be made. (I am leaving aside such phrases as "discipline" and "mental training" as being more suitable for the alienist than the teacher.) It can be said that composition brings Latin teaching into line with the "naturalness" of modern languages; and that knowledge of certain features of Latin, e.g., noun-adjective and subject-verb agreement, can be more easily tested by composition than in any other way.

Whatever the rights and wrongs of these arguments, it is at least clear that material for composition, if composition is going to be used, must be of such a sort that only one thing at a time is tested; or, if more than one thing is being tested, the points must be so easy and familiar that the possibility of error-reinforcement is cut to a minimum. In other words, the composition should be confined to a word or phrase (or even an ending) in the context of a given Latin or English sentence; or, should be of material already acquired with considerable degree of certainty.

It should also make some natural sense in English.


Weariness with the more stultifying forms of the translation-composition method has led many teachers to try to test comprehension without translation.

Questioning in Latin is principally a test of structure. For example, the passage Mercator Romam ibat ut servum venderet generates a series of questions:

Quis Romam ibat?
Quo ibat mercator?
Quo consilio ibat Romam? etc.,

which may be answered in whole sentences (insisting that the answer is stressed by coming first), or in single words or phrases. The greatest advantage of this method is that it doubles or trebles the quantity of Latin that the student hears, and tests knowledge of structure thoroughly. It would obviously get boring if pushed too far (after a class has grown familiar with the method one question per sentence should be quite enough), and it does not test knowledge of sense. This can be done by series of questions in English, and pure translation need only be used where misunderstanding is revealed.

(To teachers, whose skill and experience in both languages is presumed, translation appears the most direct and simple way of showing understanding. This is not necessarily true at an earlier age: a child can often understand without being able to translate.)

Traditional teaching assumes that Latin must be approached through the mother-tongue. Modern teaching assumes that the mother-tongue is a hindrance in learning Latin, and should vanish as far as possible from the Latin classroom. To use a rather crude symbolism, the different approaches to the Imperfect marker might be:

imptense.gif (10854 bytes)

The possibility or probability of acquiring a language without the intervention of your own is a question for psychologists and neurologists. The extensive use of this method in modern-language teaching in the last twenty years suggests that it is not impossible. We may still doubt if success in training radio-monitors and tourist-agents is necessarily an indicator of success in training Latinists.

In the achievement of this aim, of learning Latin without English, two principal methods may be used.

The first is the Latin-to-Latin questioning that we have described, but intensified, and taken beyond mere questions of structure to questions of sense. This is not revolutionary, of course. It was the regular method of learning Latin until the late Renaissance, was revived by Rouse just before World War I, and given form by Peckett and Munday in Principia, a book that remains the best available for teaching children up to the age of about eleven.

The second is the use of "pattern-drills." These consist of a short pattern-sentence which must be manipulated by the substitution, or addition, of new words, or the transformation of the structure.

Here are two examples:

Puella equum amat. Puella equum amavit.
Puella equum tangit. Puella equum tetigit.
Puella equum caedit. Puella equum cecidit.
etc., etc.
Puella equum amat. Equus a puella amatur.
Puella equum tangit. Equus a puella tangitur.
etc., etc.

Such manipulations have been the stock-in-trade of language-teachers for years. The difference in the modern method is in the vastly increased quantity, and in the fact that they are done orally, and often with the aid of a language laboratory. It should be stressed that a language laboratory is not necessary, or even desirable, for the main work in this method. The process of taking children to a different place, putting them with strange equipment in places where they cannot fully be observed by a teacher, and then turning on a steady murmuring, has proved disastrous more often than profitable. For day-to-day work in this method a tape-recorder in a classroom is more effective, and the laboratory can be reserved to give flexibility in remedial and accelerated activity. Not even the tape recorder is a necessity. The teacher with good pronunciation and durable larynx can do better work viva voce.


The attitude of a textbook is a difficult matter to assess. In essentials, the problem of a Latin teacher is that he himself knows from personal contact that there is greatness in Latin Literature; but that his students do not. They must be brought to know this, and even before they know it they must at least have the impression of greatness. This cannot be given by telling them that it is great, because such a statement is liable to arouse more antipathy and resentment than agreement--whether the statement is made by the teacher or the textbook. In the early stages of a student's learning, he is most likely to gain this impression of greatness if the attitude of his teacher and textbook is consistent with greatness: if, without any direct propaganda, there is a constant feeling that reading Latin is something important.

As far as the textbook is concerned, the two main factors that can produce this feeling are the quality of the reading-matter that is put before the students, and the tone and proportion of the ancillary material.

Reading Matter

There is a convention, for good or bad, that Caesar's Gallic Wars, more or less straight, be introduced at some moment in the second year. Accepting this as a common factor, we can concentrate on our survey on the following points:

(a) subjects and quality of the pre-Caesar readings;
(b) timing, introduction, degree of adaptation in Caesar reading;
(c) readings from other Roman authors;
(d) gradient--the evenness with which new difficulties are met.

The construction of reading-matter during the first year has always been a difficulty for textbook authors. The ideal--the unattainable ideal--would be to say: "By a certain time, I want students to know the x most frequent words in Caesar. Before then they will read y paragraphs of Latin. Therefore, each paragraph should introduce x/y new words, and at the same time include only forms and patterns already met or being taught. The words, forms and patterns introduced should be repeated at specific intervals (at first short, and then longer) to reinforce adequately."

This mechanical principle can be followed fairly closely for the first few months of a student's career, but only at the risk of producing grotesque reading-matter. It is a good measure of an author to ask how well he reconciles the mechanical necessity with the aesthetic necessity: that the readings be relevant to Roman culture, pitched to the sensibility of the adolescent child who will be reading them, and sufficiently coordinated to give some sense of continuity and coherence. The last point is particularly important. If there is no coherence, each reading may be regarded as an artificial exercise to cap a linguistic point. With coherence you at least have a chance of considering linguistic points as a step on the road to reading--which is the attitude we should prefer.

It is always assumed that mythology is essential in reading classical literature. In fact, when we read Vergil or Horace or Ovid, the mythological allusions that we need are not usually the sort of simple things that we meet in the stories of Pandora, or Phaethon, or Icarus. They need a more sophisticated knowledge. In any case, the important things about mythology are more likely to be concerned with Jocasta and Kronos. We must seriously ask ourselves if the marginal advantage to Latin background of a first-semester diet of myth can really justify the risk of adolescent contempt for our subject right from the beginning.

The physical presentation of the reading matter is a point which has received surprisingly little attention. Our ideal (however rarely it may be attained) is that a person should sit down and read a Latin book in exactly the same way that he sits down and reads an English book. But English books--or at least the sort of English books that are read casually and pleasantly--do not have glosses, footnotes, explanations and marginalia. Religious publishers found out long ago that people who would not consider reading an old-fashioned Bible would read one that was printed like other books. They removed verse numbers, italics, parallels, chapter headings, called the result names like The Reader's Bible, or The Bible designed to be read as Literature, and made a profit.

If we want students to feel any possibility of reading a Latin text like an English one, it must look like an English one. If notes are necessary they should be overleaf, so that in class, at least, a teacher can more easily encourage the habit of intelligent guessing. Since publishers cannot be persuaded to this, devices such as footnotes, marginal notes, indices (asterisks, raised numbers, etc.) and line-numbers are used.

Ancillary Material

We shall deal with this under three heads: the visual impression of the books; the background essays and notes; and the treatment of vocabulary.

The outer appearance can probably be ignored. In nine cases out of ten it is going to be hidden under a brown-paper jacket covered with advertisements for artificially-sweetened chemical drinks. Print and paper, too, are matters of considerably more interest to the adult than importance in the classroom. We hope that no teacher will choose a book on any of these grounds.

Maps and photographs are a very different matter. They can provide just the extra help and inspiration that make a text intelligible. Or, if the plates are badly chosen, and have no clear relationship to the text, they can produce a feeling of intense dullness and irrelevance. Much of this is going to be a question of subjective preference, but we can at least recommend that a critical eye be cast on illustrations of potsherds, nineteenth-century oil paintings of classical scenes, and archaeological excavations under heavy cloud-cover.

Our judgment of the value of the illustrations happens to coincide with their quantity, though there are reservations at both ends of the list.

Whenever background material is good, you will find the bad teachers who bring it to the foreground. We have all met the teacher who sets his students to read the introduction to an author and then gives them a test on the subject before they have read one word of what the author wrote. Good background material should increase interest, and accordingly increase the amount of Latin read. There is always a risk, with this sort of teacher, that more background means less reading.

If there are teachers of this sort, one weapon in their armory will be the misuse of vocabulary aids. This is so widespread a malpractice that special consideration is needed.

The best way to learn Latin words is to meet them often enough in context. This aspect of the psychology of learning has received so much attention that formulas for the number of repetitions in various times are well known to the authors of textbooks, as far as can be judged without very tedious experiment, all these books should be adequate in producing fairly low level of vocabulary recognition. For improving this, and enriching the connotations of the words that are recognised, associating them with other words, especially English words, has been widely and rightly practised--though not always with a sense of expense and profit. It is obviously profitable to link erumpo with "eruption" and manus with "manual." When the link becomes twisted, or the derivative unfamiliar, the waste of time in teaching it may outweigh the profit to the student's knowledge of Latin. It would be interesing to know how profitable, for instance, are umbra--"umbrella," humilis--"humble," plicare--"applicable," celer--"celerity."

Latin teachers are almost certainly interested in words as words. Their students may not be. When working with derivatives, many teachers may be tempted to go far beyond the profitable use of the technique. They may tend to forget that derivatives are only justifiable in a Latin classroom as reminders to the meaning of words that have been forgotten, or clues to the meaning of words that must be guessed. Any testing of lists of derivatives is probably a misuse of a student's time.

The implications of this, as far as textbooks are concerned, may be that derivatives should be regularly supplied as appendix to vocabulary definitions, that where casual vocabulary is needed for a particular passage it may be noted in the form "X comes from this--Guess;" and that useful equivalent-groups, such as -itas/-ity should be treated when the student has met sufficient examples to make them significant. For interest the ancillary material might contain the occasional eccentric or linguistically stimulating derivation--tandem, "lieutenant," and so on. But exercises on derivatives are an indication of a failure in attitude.