Until we know more about speech-psychology and the ultimate processes of language-study, it is doubtful whether we can embody in the form of a concrete principle the subject treated in this chapter. The writer would prefer, at this stage of our knowledge, simply to submit the following considerations in the hope that future research will throw further light on the subject and render it possible to co-ordinate it with those branches of linguistic pedagogy which are more familiar to us. Indeed, when we have ascertained experimentally the exact nature of what we shall call ‘memorized’ and ‘constructed’ speech-material, it is conceivable that the whole subject will become so clarified that it will be possible to reduce to one main principle all or most of what has been said in the foregoing chapters.

Now, whenever we open our lips to speak, or whenever we set pen to paper, it is with the object of producing one or more units of speech. These units may be short and simple, such as: Yes, No, Here, or they may be word-groups, such as: Very well, I don’t know, Yes, if I can, or they may be complete and even complicated sentences containing one or more subordinate clauses. But whatever the unit may be, long or short, simple or complex, one thing is clear: each unit has either been memorized by the user integrally as it stands or else is composed by the user from smaller and previously memorized units. This is a fundamental fact about speech which stands out clearly and unmistakably; it is not a fanciful supposition or an idle conjecture, it is an axiomatic truth.

Now let us term ‘memorized matter’ everything that we have memorized integrally, and ‘constructed matter’ everything that we have not so memorized, but which we compose or build up as we go on. Can we distinguish the two things? In most cases we can. Monosyllabic words have generally (although not necessarily) been memorized as they stand; we say and understand the word cat, because once upon a time we had the occasion to hear the word in question and the opportunity to connect it with its meaning and to retain it. The word cat is included in our memorized matter. Probably most words of two or even more syllables have been acquired as memorized matter. Great numbers of compound words have also been acquired in the same way. A considerable number of word-groups and sentences are included in our memorized matter. Such sentences as I don’t know, Just come here, Pick it up, I don’t want it, are most probably memorized with most speakers.

Now consider a unit of speech such as: I saw Henry Siddings between six and half-past at the corner of Rithington Lane. Is it the sort of unit which we should use as a result of having memorized it integrally? An actor or reciter may indeed have occasion to do so, but apart from those whose duty or hobby it is to memorize ‘lines’ it is an extremely unlikely specimen of memorized matter. The writer has just composed it, and does not even know whether there exists such a surname as Siddings or a place called Rithington Lane; there are millions of chances to one that it is an entirely original sentence. Most of the things we utter or write come into the category of constructed matter; their component parts have been memorized integrally and so constitute memorized matter; but the complete units are constructed, they are the result of rapid and probably unconscious acts of synthesis.

This is no place for statistics, even if data were available; it must be left to investigators to ascertain the relative amount of memorized and constructed matter used by the young child in his first months or years of speech. Inquiries of this sort should afford some valuable and surprising evidence; the writer has had occasion to note that a French-speaking child of about ten was even unconscious of the composition of units such as pomme de terre or quatre-vingts, just as the average adult English person is unconscious of the composition of fortnight or nevertheless.[6] What will certainly complicate such research work is the paradoxical fact that constructed matter may become memorized by dint of frequent repetition. A further complication is added by the fact that the two types of matter may also be considered from the point of view, not of the speaker, but of the auditor.

One of the questions that concerns us at present is to ascertain what should be the right proportions of memorized and constructed matter in the initial stages of learning a foreign language.

Too large a proportion of memorized matter will render study unnecessarily tedious, for memorizing work, even under the best of conditions, is less interesting than the piecing together of known units. Too large a proportion of constructed matter, on the other hand, will certainly result in an artificial sort of speech or a pidgin form, with all its evil consequences. At the present day, as in the past, the tendency in language-study is to pay far too much attention to constructing and not nearly enough to memorizing.

What concerns us still more is to ascertain definitely by experiment what is the exact nature of those processes by which constructed matter is derived from memorized matter. We must find out what really does happen in the case of young children in the first stages of their speech-experience, and by what mental processes those persons called born linguists attain their results.

There would appear to be three distinct manners of producing constructed matter; these may be termed respectively:

(a) Grammatical Construction

This process consists in memorizing ‘dictionary words’ (the infinitives of verbs, the nominative singular of nouns, the masculine nominative singular of adjectives, etc.) and of forming sentences from them (with or without the intervention of translation) by means of applying the various rules of accidence, syntax, derivation, and composition.

The following is a typical example of the process. An English student wishes to form as constructed matter the German sentence: Ich habe mit grÖsstem VergnÜgen seinen freundlichen Vorschlag angenommen, from the previously memorized units ich, haben, mit, gross, VergnÜgen, sein, freundlich, Vorschlag, annehmen. Besides having to determine (in accordance with rules of word-order) the relative position of the nine primary units, he has to perform the twelve following operations:

It will be noticed that most of these operations require, in addition to a perfect memory of the grammatical rules (including numbers of word-lists), a fine power of logical discrimination. Needless to say, no speaker of German actually does perform any of these operations (except perhaps on very special and rare occasions), and we dismiss as a patent absurdity the supposition that the young native child constructs his matter in any such way.

(b) Ergonic Construction

In this process we work from an entirely different sort of memorized matter; instead of being merely ‘dictionary words’ it consists of (a) more or less complete sentences, and (b) units of speech which we may term ‘ergons,’ i.e. ‘working units’ derived and inflected in advance by the teacher (or the author of the course), each ergon being thus quite ready for use.

The following is a typical example of the process:

A fairly simple sentence is memorized; let us say, Ich kann meinen Stock heute nicht nehmen, “I can’t take my stick to-day.” Appropriate groups of ergons are also memorized, such as:

The student will then form (as constructed matter) as many of the 16,128 resultant sentences[7] as is considered necessary for this particular vocabulary. This will be done by means of drills and habit-forming exercises based on the following substitution table:

The essential difference between grammatical and ergonic construction lies in the sort of memorized matter used in either case. In grammatical construction the memorized matter consists exclusively of what we have called ‘dictionary words’ (a large proportion of which require modifying in some form or other before being available for use in a sentence), whereas in ergonic construction two sorts of memorized matter are required: a more or less complete sentence and a number of ergons (units of language inflected or composed in advance for the student, instead of by the student).

(c) Conversion

This process consists in memorizing a number of sentences all composed in a more or less uniform way.

When these sentences have been memorized, the student is taught by a series of appropriate drills and habit-forming exercises to convert each sentence into another form.

The following is a typical example of the process. The student memorizes the ten following sentences:

He then listens to the teacher, who says:

and after one or more repetitions performs the conversion himself in the same way, with or without prompting by the teacher or the book.

The teacher will then change the sentences in some other manner, for instance:

The student listens, and subsequently performs the same series. On other occasions each of the ten sentences may be converted into forms such as:

In the case of conversion the difference between memorized and constructed matter is not so marked as in the two synthetic operations, nor is the yield of constructed matter so great. Indeed, in extreme cases, the form into which the original sentence is to be converted will have to be learnt integrally, and so becomes in itself memorized matter. On the other hand, some forms of this type of work are practically identical with exercises based on ergonic construction, and for these two reasons it has been held that conversion is not a distinct process for forming constructed matter, but merely a modified form of ergonic work. Whether this view is justified or not is a matter more of academic than of practical interest to the language-teacher.

These then appear to be the only three processes known by which memorized matter can be developed and expanded into original composition. What we have called grammatical construction is the classical and almost universal method. What we have called ergonic construction is embodied more or less unsystematically in a number of language-courses and the more enlightened books of instruction. Conversion is also practised, but still in a sporadic and desultory fashion.

Now, some thirty years ago the reform movement started. In several different countries bands of zealous pioneers took up arms against the then prevailing system and sought to put an end to it. The reform prospered. The reformers have carried all before them, and the daring innovators of twenty or thirty years ago now enjoy the prestige that their efforts have earned for them.

What was the nature of this reform? What abuses has it swept away? And for what innovations have we to thank it? It would appear, on analysis, to have had a threefold object:

The first two objects have certainly been attained; phonetics is the order of the day, and both teachers and students have to use phonetic symbols whether they like it or not; moreover, the new generation does recognize that the deciphering and analysis of ancient texts is not the primary use of language.

The third object has not been so successfully pursued; indeed, we are still very far from learning the foreign tongue by the same processes as those by which we learnt our own. The chief reason for this failure was a bad diagnosis of the chief evils of the system hitherto employed. Many of the reformers and most of their disciples imagined ‘translation’ to be the root of the evil, and so translation in every shape or form was banned; there must be no bilingualism at all, and so the mother-tongue must be excluded from the course, the lessons must be conducted entirely in the foreign language.

But translation and the use of the mother-tongue, as it turns out, are perfectly harmless and in many cases positively beneficial; the evil lay in the exaggerated attention which had always been paid to grammatical construction; that was the dragon that the St Georges might well have slain had not the red herring of ‘translation’ unfortunately been drawn across the track. As it was, the red herring was duly run down and annihilated, and the dragon still lives!

The misunderstanding was natural enough; logicians would quote it as an example of the fallacy of the False Cause. The process of grammatical construction was carried out by means of a vicious form of translation exercise, and the result was utterly bad. Two important reforms might have been effected: in the first place, the vicious form of translation might have been replaced by a beneficial form; and secondly, new and more worthy uses of translation might have been found. But the act of translation itself (nay, the mere use of the mother-tongue) was made the scapegoat and so paid the penalty. It is now time for a second band of reformers to attack and to destroy the original cause of unsuccessful language-study, viz. grammatical construction, or at any rate to limit it to special cases and to appropriate occasions. It is time, too, to rehabilitate in some measure the character of the comparatively innocent process of translation, and to remove the stigma attached to those who still use the mother-tongue as a vehicular language, and by so doing proceed naturally enough from the known to the unknown.

These are no reactionary suggestions; they are made in the spirit of the nine essential principles treated in the previous chapters, and are not in contradiction to the urgent plea set forth in these pages for the recognition and fostering of our ‘spontaneous’ capacities for language-study. We can afford to ignore no necessary tool in our efforts to teach well and to produce perfect results, and translation is often a necessary tool, especially during the process of deriving constructed from memorized matter.

We suggest for the moment no tenth principle based on these considerations; we submit the problem and we more than hint at a solution. It is now time for experimental work on ‘ergonic’ lines, and the data to be obtained thereby will enable us to form our conclusions and to embody them among the principles of language-study.