The title of this report is explained by the dual purpose of the study: to discover what words and what inflectional endings occur most frequently in Latin literature.
The Latin vocabulary now taught to beginners is based on a word-count of selections from six Latin authors of the Golden Age. The wider reading now required by the "reading method" is not adequately sampled by this count, and no count of other Latin authors commonly read in college is available, except a few indices verborum of single authors.
This count Is based on 202,158 words in selections from over two hundred Latin authors, from Ennius to Erasmus. It yields a "recommended basic vocabulary" of 1,471 vords, arranged in groups of related words., which enable one to recognize 83.6 per cent of the words in the literature examined in this study. The Latin Word List now in use has 1,791 words which enable one to recognize only 81.2 per cent of the words in this literature.
The inflectional system now taught to beginners is unnecessarily cumbersome and difficult to apply in reading Latin. The average first-year Latin textbook published since the Classical Investigation requires the memorization of 1,572 inflected forms in paradigms. Learning them in this way may cause the beginner to think through as many as fifty forms in paradigms before he can locate all possible interpretations of a given ending. The traditional interpretations of these endings (such as dative or ablative plural) do not indicate directly the function of the word in a sentence, but leave the beginner to choose among as many as thirty possibilities. Some evidence in presented to indicate that this inflectional system is rarely mastered by beginners in this country, and is poorly organized for reading purposes.
This study indicates that eighteen "common endings" carry about 90 per cent of the burden of Latin grammar. They occur on 66.5 per cent of Latin words; and additional 23.4 per cent have no endings, and 2.6 per cent consist of Irregular pronoun forms which must be learned as vocabulary under any system. If the beginner knows these eighteen "common endings" and learns the pronoun forms as vocabulary, he can already tell the function in a sentence of 92.5 per cent of all the Latin words he will ever read. In addition he may gradually learn twenty-two "rare endings" which occur on 6.8 per cent of Latin words, and fourteen "penuitimate signs" (tense and other signs) which occur between stem and ending in 7.1 per cent of Latin words. Thus the task of learning Latin grammar is reduced to learning the significance of forty endings and fourteen penultimate signs. None of the remaining endings, which occur on 0.7 per cent of Latin words, occurs on the average more than once in 5,000 words.
Interpretations of these endings and penultimate signs, which indicate as directly as possible the function of the word in a sentence, were devised and taught over a period of three years. For example, dative and ablative endings alike were called "adverb" (or "adverbs" in the plural), since they usually occur in the equivalent of an adverbial phrase. These "adverb endings" were carefully distinguished from true adverbs. Pupils taught only by this method read on the average about four times the usual amount without any knowledge of conjugations, declensions, cases, genders, quantities, etc. The method of teaching this system is outlined, only to show how the results of this research may be utilized. There has not yet been time for controlled experimentation on a wide enough scale to establish the superiority of this method, but the data on the frequency of Latin words and their endings do not require such confirmation of their validity.
F.B.D.
The first purpose of this study was to discover which words occur most frequently in the whole range of Latin literature from Ennius to Erasmus. Previous studies had sampled only the selections commonly read in secondary schools. The standard Latin word list in use at the present time is that of the College Entrance Examination Board. It is a list of 1,791 words arranged In three groups, as the words are most useful in reading Caesar, Cicero and Vergil respectively. It is based upon the study in 1907 by Gonzalez Lodge of Teachers College, Columbia University, of 77,142 words in the following sources: Caesar's Gallic War, Books I-V; Cicero against Catiline, on Pompey's command, and for the poet Archias; and the first half of Vergil's Aeneid. His list was extended by Stephen A. Hurlbut of St. Albans School, Washington, D.C., to cover Caesar's Gallic and Civil Wars complete, four additional speeches of Cicero, and his De Senectute and De Amicitia, the remainder of Vergil's Aeneid, his Bucolics and Georgics, and three other authors commonly read in secondary schools: Nepos, Sallust's Catiline and Jugurthine War, and six books of the Metamorphoses of Ovid. The Latin word list now in use, therefore, is based upon six Latin authors: Caesar, Cicero, Vergil, Ovid, Sallust, and Nepos.
While this word list admirably covers the selections commonly read in secondary schools, it offers no proof that the words which occur most frequently In these selections will also occur most frequently in the further reading of Latin. Lodge writes in his preface:
Furthermore these words are the essential words in the Latin language; for examination of a relatively equal amount of material selected from Caesar's Civil War, Cicero's Orations, other than those read in schools, and Ovid, showed the occurrence of more than nine-tenths of these words.
No other proof is offered, and this hardly establishes the case, for the Latin language is not sampled adequately by Caesar, Cicero, Vergil and Ovid, nor even by the later addition of Sallust and Nepos. These are all standard "high school authors"; beyond them lie Plautus, Terence, Catullus, Lucretius, Horace, Propertius, Tibullus, Lucan, Statius, Martial, Livy, Seneca, Quintilian, Tacitus, Pliny, Suctonius, Petronius and Apuleius, to mention only the major classical authors, plus the whole wide domain of medieval Latin. Furthermore no data are offered as to how frequently the words in Lodge's list occurred in the other selections which are mentioned; we are only told that they occurred. Without further validation of Lodge's list, even as extended by Hurlbut, against a wider sampling of Latin literature, we can only say with confidence that we know which words occur most frequently In the practice materials which are now widely used to develop reading skills, but not in the materials to which those skills will ultimately be applied. It is as though we wished to know what elements of mathematics would be most useful, and we had available only a summary of the elements which occurred most frequently in beginning textbooks in mathematics. Our practice materials in Latin should presumably be selected to develop the vocabulary of the widest ultimate usefulness in reading Latin, and it is a precarious assumption that this vocabulary will be identical with that of the practice materials now in use.
These things are not said in disparagement of the contributions of Lodge and Hurlbut, whose careful, exacting labor the present investigator knows too well, having struggled in their footsteps. Their lists exactly met the needs of our secondary schools when the materials to be read in Latin were limited, standardized and universally accepted. The situation has been changed, however, by the introduction of the "reading method" in the teaching of foreign languages, which calls for a much wider selection of practice materials than formerly, and seeks definitely to establish the habit of reading in the foreign language as a leisure time activity. This method calls for the construction of easy reading materials in a "controlled vocabulary," in which new words are introduced very gradually: usually only one new word in every fifty or sixty running words. The labor of writing such materials is so great, to say nothing of the pupils' labor in learning the new words, that we begrudge every word which is not of the widest ultimate usefulness. Hence even in secondary schools the present Latin word lists are no longer completely satisfactory, and no study has previously been made of the vocabulary of the Latin authors read in college. Furthermore, with the steady decline in the proportion of pupils studying Latin in secondary schools, from 54 per cent in 1910 to 15 per cent in 1936, as attested by the Biennial Reports of the United States Office of Education, and with trends in educational theory pointing toward a still further decline, and perhaps to the virtual elimination of Latin as a subject of study in secondary schools, 'we must make some provision for the mature reader, either in college or in later life, who may wish to "pick up" a reading knowledge of Latin In the shortest possible time. Such a reader is not likely to confine his reading to the standard "high school authors"; he is more likely to want to read widely in such authors as Catullus, Lucretius, Terence, Vergil, Horace, Ovid, Martial, Cicero's Letters, Livy, Tacitus, Petronius, and Apuleius. Such a reader is likely to turn to the admirable Loeb Classical Library which provides, on alternate pages, an unencumbered text of practically all the extant Latin and Greek authors and a clear, unpretentious, faithful English prose translation. It is frankly the objective closest to the heart of the present investigator to enable such a mature reader within a single year to read a Loeb translation intelligently: I.e., to follow the Latin text with understanding, enjoyment, and speed comparable to the reading of a modern language, referring to the English translation only for the meaning of unusual words, and for the resolution of unusually difficult syntax. The materials to develop such skill in reading Latin are not included in the present study, but the fundamental research necessary for the production of such materials is here reported in what we hope will be the most immediately usable form.
In order to secure a representative sampling of the vocabulary of the whole range of Latin literature, three standard anthologies were chosen: The Oxford Book of Latin Verse, Avery's Latin Prose Literature, and Beeson's Primer of Medieval Latin. These three anthologies contain selections from practically every important author from Ennius to Erasmus, except Plautus and Terence, in about the proportion in which our mature reader would want to read them - that being the function of a successful anthology. Among the poets included in the present word count, to mention only the most important, were Lucretius, Catullus, Vergil, Horace, Tibullus, Propertius, Manilius, Ovid, Seneca, Calpurnius Siculus, Lucan, Statius, Martial, Lactantius, Ausonius, and Claudianus. Some eighty poets were represented, besides twenty-nine anonymous selections, totalling 75,323 words. Eight pages of early Latin verse prior to Ennius were omitted from this count, as having too special a vocabulary for the purposes of this study. The prose included five to forty pages apiece from each of the following authors: Cato, Varro, Cicero, Caesar, Sallust, Nepos, Livy, the two Senecas, Velleius Paterculus, Quintus Curtius, Petronius, Quintilian, Tacitus, Pliny and Suetonius - a total of 49,363 words. The names represented in Beeson's Primer of Medieval Latin are not so well known, but this anthology is probably the most popular introduction to medieval Latin, and includes a representative collection of prose and verse from the fall of Rome to the Renaissance, totalling 77,472 words. We shall discuss the inclusion of medieval Latin after we have presented data validating the results of this study against classical authors alone. All told, 202,158 words from more than 200 Latin authors were counted. This is probably a more representative sampling of the vocabulary of Latin literature than the 77,142 words from the standard "high school authors" studied by Lodge and more than doubled, but from the same narrow range of authors (six in all) by Hurlbut. The count admittedly and, we believe, justifiably falls far short of the enormous vocabulary sampled in English word counts. Both the vocabulary and the literature to be sampled in Latin are far less extensive than in English, and the technical difficulties involved are much greater, so that the task must be done by scholars rather than by clerks. All the multitudinous forms of Latin words must be sorted unerringly at high speed. Tulerunt must be thrown upon fare; sublata upon tollo, egisti upon ago, and so on. Feriam, strike, must be distinguished from feram, carry, from feram, savage, and from ferias, holidays. Malo may mean "I prefer," or it may be some form of the Latin word for "bad," for "apple," or for "mast." Such difficulties make anything like the wholesale methods of Dr. Thorndike in counting English words impossible to carry through in Latin. However, this relatively small sample enables us to identify practically all of the words which are important enough for the beginner to memorize, as we shall presently demonstrate.
There has been some criticism in recent years of the determination of the vocabulary to be taught to beginners on the basis of frequency of occurrence in literature (see Bibliography). This criticism loses its force when applied to Latin. It is said that frequency of occurrence in literature is not a valid measure of usefulness in conversation. Probably It is not, but our pupils do not need to speak or understand spoken Latin. Latin exists only as literature, except for specialized uses in medicine and law and Latin phrases used in English, for which special counts already exist. It is also said that in attempting to speak or write a foreign language.. the student will not be able to express very many ideas in the vocabulary yielded by a frequency count. Granted., but again our students have no need for this ability. Writers of Latin textbooks may have to use a few words which do not occur frequently in Latin literature, but the words set for mastery can have no other valid basis than their usefulness in reading Latin. A few exceptions may be made for Latin words which have a large number of important English derivatives, but in the main these words also occur frequently in Latin literature. Lindsay's Etymological Study of the Ten Thousand Words In Thorndike's "Teachers Word Book" yielded only eighty-eight Latin words which had ten or more derivatives among the ten thousand most common English 'words. All but four of them (carrus, for, plico, stringo) are in our Recommended Basic Vocabulary, although ten are not in the form which Lindsay selected an the root, but in closely related forms which equally well explain the meaning of the derivatives: e.g., fides for fido, defendo for fendo, plenus for pleo, probo for probus, and so on. The only other criticism of the usual technique of word counting which applies to Latin is that it does not reveal in what senses each word is used most frequently. This lack will be supplied for Latin when the monumental Thesaurus Linguae Latinae is completed by the collaboration of seven German universities. The present study could not hope to compete with them.
The method of counting the words was as follows: first we secured nine reams of mimeograph paper in three colors: pink for classical prose, yellow for classical poetry, and white for medieval Latin. Next we ran this paper through a mimeograph machine, marking off each page into 44 little squares: four to the line, eleven to the column. Then we copied the entire contents of the three anthologies (except the pre-Ennian poetry) on these pages, one word to a square, in the exact form and order in which the words appeared in the printed text. When a word appeared in other than the common sense, such as mala meaning "apples," we wrote the meaning directly beneath the word in the same square. Occasionally this precaution was overlooked. While reading along in a context in vhich feram obviously meant "I shall carry," we sometimes forgot that in another context it might mean "savage" This left us with a relatively small number of words which might be assigned to entirely different roots, depending on the context - and the context was lost, once the sheets were cut up and the words were alphabetized. In all such cases many other forms of these same words occurred which could be distinguished: for example, in the case of faro and ferus there were large numbers of forms such as tuli or ferum which could belong only to one or the other, in the proportion of about three and one-half fero to one ferus. The ambiguous forms were then assigned to one or the other in this same proportion. While this assumption was a bit precarious, it was the only way out of an embarrassing, unforeseen situation. It affected so few words, however, and there were always such clear indications of the probable frequency of these words anyway, that the accuracy of the count was only slightly impaired.
After all the words in the three anthologies were copied in their little squares, a printer cut up the pages, whole reams at a time, into their component squares, and we came home with 202,158 little slips of paper, each containing one Latin word. First these words were alphabetized as to their initial letters. Then all the a's were put together and alphabetized to the second letter, then to the third, then to the fourth if necessary, and finally to the separate words. This process finally assembled all the scattered occurrences of a given word, from Ennius to Erasmus, in one pile of pink, yellow and white slips of paper. When this stage was reached, all the words which occurred five or more times were carefully counted, listed, tabulated as to frequency in each of the three anthologies, and total frequency, checked, and rechecked.
This detailed description of the method of counting is given, first, because the method of an investigation bears upon the accuracy of the result, and second, as a suggested technique for other investigators in this field. On the first point we feel that the method yielded highly accurate results, with the one exception mentioned above, because we had tangible slips of paper to show for every figure printed in our tables, which could be reexamined whenever any question arose about them. Any tally method loses this opportunity for later checking. On the second point, we thought the method would be somewhat faster than the only alternative we could think of: namely, writing out all the words in the dictionary and Putting a tally after each word every time it occurred. We tried the latter method on a small sample, putting tallies in the margin of our dictionary, but too much time was consumed in turning pages to find the words. After trying the method we finally selected, however? we hesitate to recommend it to other investigators, especially in view of our difficulties with ambiguous forms. We should recommend that other investigators assume that the three hundred words of highest frequency in our list will be important in any subsequent vocabulary study, and these be crossed out in the materials to be examined. This can be done very quickly, and will save about 58 per cent of the total number of words to be counted. Then a tally list might be made in alphabetical order of the 2,200 words which are next highest in frequency in the present study; i.e., of the words which occurred from 10 to 99 times. Using a double-column page, single spaced on the typewriter, this should take only 22 pages, and with the aid of index tabs. words should be very easy to locate. This tally list should account for about a third of the words in the literature to be examined. The remaining words, not more than 10 per cent of the total, none of which will occur very often, can either be written out and later alphabetized and counted as in the present study, or tallied in the margin of a dictionary.
In the present study we counted all forms of a word as one word, except the highly irregular forms of the common pronouns and the comparative and superlative of irregular adjectives which are usually taught separately. Ego is listed separately from mihi, bonus from optimus, and so on. But any form of a verb which is regularly formed from any of its principal parts is counted with that verb, since pupils are expected to learn the principal parts in learning the verb. For example.. tulisset is counted with fero; posita with pono, and so on. The forms of sum are all counted with sum, even when used as auxiliary verbs, except that some of its Irregular forms were first counted separately, to discover which of them it would be necessary to learn. Later, however, they were all combined under sum. We did not think it necessary to list the number of different authors in whose works each word appeared. This is important only when a small number of sources is used, and there is danger that a word may occur very frequently in one source and so attain a fictitious importance, even though it rarely or never occurs in other sources. We saw no possible way for this to occur in the brief selections from various works of over two hundred different authors included in the present study. None of these authors could possibly overwork a favorite word to an extent which would falsify the total count, any more than a single individual in a ten-minute conversation could alter the basic vocabulary of his community. We listed separately the number of times each word occurred in each anthology, as well as the total frequency, to give an idea of its relative importance in classical prose, classical poetry, and medieval Latin, but a further analysis of the number of authors in each anthology who used each word would have quadrupled the work and confused the result. When the range (as the number of sources is called) is given, it is a question which figure to use, the range or the frequency. A compromise is usually adopted: an index figure which means nothing in itself, but which is derived from both the range and the frequency in accordance with the investigator's idea of the relative importance of these factors. We prefer a single figure which means what it says, and we avoid the whole difficulty which the range is designed to correct by using anthologies rather than a few extensive sources.
We have not thought It necessary or feasible to follow Lodge in mentioning where each word first occurs. This is important in learning vocabulary in connection with a fixed course in Latin, but no one is likely to read the particular selections we examined in the order in which they occur in the anthologies. There would be no point in knowing that a given word occurred first on page 9 of the Oxford Book of Latin Verse if one never intended to read page 9, but had already encountered the word in a selection which was printed in this anthology on page 441.
These procedures yielded an alphabetical list of 3,800 different words which occurred five times or more in the literature examined, and which accounted for 96.4 per cent of the words in the three anthologies. Each word in the list vas followed by four numbers, representing how many times it occurred in the classical prose, the classical poetry, the medieval Latin, and the total of the three respectively. This list will be useful chiefly in writing practice materials in a "controlled vocabulary," for when the writer Is in doubt about the importance in further reading of any word he proposes to include, he may look it up in this list and check its frequency of occurrence. If it is not in this list, it is safe to assume that it occurs very infrequently.
With these data in hand we tried to locate a "point of diminishing returns" for the general reader in acquiring a usable vocabulary. To this end we divided the complete word list into word groups according to frequency of occurrence, and found the number of words in each group, the total number of times the words In each group occurred in the three anthologies combined, and the percentage this represented of the total number of words studied, excluding proper names (194,387 vords). In the following table these results are cumulated: i.e., group 2 represents the sum of group 1 plus group 2; group 3 represents the sum of groups 1, 2, and 3, and so on. In this way one can most easily see the net effect of learning the successive groups of words.
Word Frequency Number Number of Per Cent of All of Group of Occurrence of Words Occurrences Words Studied*
(1).... Over 1,000 times 17 45,478 23.3 (2).... Over 500 times 42 62,303 32.0 (3).... Over 200 times 127 87,476 45.0 (4).... Over 100 times 307 112,888 58.0 (5).... Over 50 times 661 137,608 70.7 (6).... Over 20 times 1,556 165,521 85.1 (7).... Over 10 times 2,583 179,499 92.3 (8).... Over 5 times 13,798 187,507 96.4
*Excluding proper names, which occurred 7,771 times, leaving 194,387 words.
This table is presented graphically on the next page, showing the percentage of the vocabulary of Latin literature which can be recognized after learning any given number of Latin words in order of frequency. The results of learning the first few words in order of frequency are astounding. After a pupil has learned the following 21 words he already knows the meaning of one-fourth of all the words he will ever read in Latin: -que, qui, et, sum, in, is, hic, non, cum, ad, ille, omnis, a or ab, suus, dico, nec, de, sed, ipse, si, The first three of these words alone account for 10 per cent of the Latin language.
There was a strong temptation to locate the "point of diminishing
returns" at the words which occurred fifty times or more. There, for
the trouble of learning 661 words, one would be able to recognize 70.7
per cent of the words in Latin literature, while the next group of 895
additional words gave only an additional 14.4 per cent, or 85.1 per
cent in all. After trying both of these lists with students, however,
over a period of three years, it was decided to draw the line after
the latter group of words. The reasons lying behind this decision
were as follows. We found that in order to read Latin with sustained
interest it as necessary for pupils to be able to recognize or guess
the meaning of about 98 per cent of the words in the literature they
were reading. This subjective estimate was corroborated by the
opinion of Michael West, a pioneer in the ?treading method," that not
more than one new word should be introduced in every fifty running
words. Knowing 85 per cent of the words outright, students were able
to guess the meaning of about 10 per cent more as derivatives or
compounds of words already known, or as obvious roots of known English
derivatives. In addition, they were able to guess up to 3 per cent on
the basis of the context, leaving only about 2 per cent to be looked
up in a dictionary or on the facing page of a Loeb translation.
Nothing like this result was attainable on the basis of knowing only
70 per cent of the words outright.
On the other hand, the next group of words in order of frequency,
which occurred from 10 to 19 times, yielded only an additional 7.2 per
cent of the vocabulary of Latin literature in return for the trouble
of learning 1,026 additional words. This seemed clearly to mark a
point of diminishing returns for the general reader in the burden of
acquiring a usable vocabulary. Remember that in round numbers the
first 1,500 words yielded 85 per cent of the vocabulary of Latin
literature; the next 1,000 words only 7 per cent. Add in the fact
that pupils who knew only the 1,500 words 'were already reading
extraordinary amounts of classical Latin 'with apparent relish and
without overworking the dictionary. Add in also the experience of
Michael West that when he worked up to a vocabulary of 1,500 words, he
found no difficulty in writing whatever he wanted to express within
that vocabulary. Finally there was the obvious consideration that
the words which occurred twenty times or more in the three anthologies
would be more likely to be important in other Latin authors than the
words which occurred less than twenty times. All these considerations
led us to recommend
approximately the first fifteen hundred words in order of frequency as
a basic vocabulary for the general reader. Later we deleted 85 words
which are important chiefly in medieval Latin, as will presently be
explained. This left 1,471 words, accounting for 83.6 per cent of the
vocabulary of the three anthologies combined. The Latin specialist
will, of course, extend his vocabulary beyond these limIts as far as
he can.
Our next task was to find some objective method of validating our
results: i.e., of proving that the frequency of these words in our
three anthologies would correspond closely with their frequency in a
random sampling of Latin authors. For example, the first 300 words in
order of frequency accounted for 58 per cent of the words in our three
anthologies, and 'we made the assumption or prediction that they would
also account for about 58 per cent of the words in Latin literature.
This assumption required proof. We already had subjective proof that
the lists worked well in the actual teaching of Latin, but this
impression might have been due to our natural bias in their favor. We
therefore selected ten short passages of 100 words apiece (which had
not occurred in our three anthologies) from representative classical
authors, and looked up each word in these passages to find out how
many times it occurred in our three anthologies. The following
percentage of words in these passages occurred in the six indicated
ranges of frequency in our anthologies:
This table is read as follows: of the 100 words at the beginning of
Caesar's Gallic War, Book 11, 60 appeared in the list of words which
occurred 100 times or more in the three anthologies, 10 appeared in
the list of words which occurred from 50 to 99 times, and so on. For
our present purposes the last four lines of the table are most
significant. In our study of the three anthologies we found that the
words which occurred 100 times or more comprised 58 Per cent of the
total number of words in these anthologies. On this basis we
predicted that 58 per cent of the words in Latin literature might be
found in this list. Actually 57.6 per cent of the words in these ten
short passages, not included in the anthologies, and selected quite at
random; did occur in this list. We predicted that 12.7 per cent would
appear in the next list of words which had occurred from 50 to 99
times In the anthologies. Actually 13.8 per cent so appeared,
slightly better than prediction. We predicted that 14.4 per cent
would appear in the next list, and found 13 per cent. Since these
three lists taken together constitute our basic vocabulary, with a
predicted coverage of 85.1 per cent of the vocabulary of Latin
literature, we are naturally most anxious to find out what per cent of
the words In these ten representative passages were so covered; we
find it to be 84.4 per cent. Statistical calculations (based on the
theory of chi-square and 5 per cent fiducial limits) indicate that in
19 out of 20 similar samples of 1,000 words each, 85.1 + 2.2 per cent
of the words ah6uld be found in the basic vocabulary. The obtained
figures were as close to prediction as could reasonably be expected in
a small sample of this kind. This calculation should not be confused
with the more familiar concept of Probable Error, which would mean
that in repeated samplings of Latin literature of 202,158 words each,
half of the time our present basic vocabulary would include 85.1 + 0.1
per cent of the words sampled.
These statements are included for the benefit of the statisticians.
For Latin teachers the best proof of the validity of our results is,
first, the inherent probability that an accurate count of 202,158
words from such a wide range of Latin authors would correspond pretty
closely with any further sampling of the vocabulary of Latin
literature; second, the fact that our lists were found to include just
about the percentage of words we said they would include In ten short
passages from classical authors. It is obvious from our citations
that we did not search among Latin we found ten passages which would
give this result. These were usually the first hundred words in
representative works of these authors unless they happened to have
been used in our anthologies As shown in the above table, the
correspondence between our prediction and our findings, even in this
small sample, was close enough for any practical purpose.
An incidental result of this validating procedure was a technique
for determining very accurately the relative vocabulary difficulty of
various Latin authors, on the assumption that the easier authors will
use a higher percentage of the more common words. Compare Livy and
Lucretius with Catullus and Horace in our table, for example. As a
single index figure representing vocabulary difficulty we recommend
the percentage of a sample of about 1,000 words from representative
passages of the author in question which occur in our Recommended
Basic Vocabulary: i.e., words which occurred 20 times or more In the
literature examined.
Our next task was to find out what percentage of the words in these
three anthologies the Latin Word List of the College Entrance
Examination Board would enable one to recognize. For this purpose we
starred every word in our list which also occurred in the College
Board List, found the total number of times these starred words
occurred In each anthology and in the three taken together, and then
reduced these figures to percentages. We found that a student who had
mastered the 1,791 words in the College Board List would be able to
recognize 84.5 per cent of the classical prose, 80.7 per cent of the
classical poetry, and 79.6 per cent of the medieval Latin, or 81.2 per
cent of the words in the three anthologies taken together, excluding
proper names. This last figure is represented by a vertical line in
the graph on page 10. While this showing was surprisingly good, in
view of the narrow range of authors sampled by the College Board List,
it was not quite as efficient a word list as could be compiled if the
objective were to acquire a vocabulary of the widest possible
usefulness. The basic vocabulary of 1,471 words recommended in this
report enables one to recognize 83.6 per cent of the words in the
three anthologies. If one were content to be able to recognize 81.2
per cent of the vocabulary of Latin literature (the outcome of the
College Board List), one could attain this result by learning
approximately 1,200 of the most important words in the present
list-600 fewer than are necessary to attain the same result in the
College Board List.
This does not mean that the present list is precisely onethird more
efficient than the College Board List. It is a quest- ion of purpose.
If one wishes chiefly to learn the words which appear most frequently
in the authors commonly read in secondary schools, the College Board
List is preferable. If one wishes to read a wider variety of Latin
authors either during the high school course, in college, or in later
life, the present list is preferable.
The immediate purpose of these findings Is to furnish additional
evidence on the question of the extent to which the vocabulary of the
Latin authors traditionally read in high school really functions in
the later, or wider, reading of Latin. As we indicated in our first
few paragraphs, no convincing evidence on this point had previously
been offered. Lodge writes in his preface: "A student who has at his
command these 2,000 words will have the vocabulary of fully
nine-tenths of all the ordinary Latin that he is ever likely to come
into contact with." Our findings indicate that Lodge was a bit
optimistic, although perfectly honest, since these words did comprise
above 90 per cent of the literature he examined. Although about 200
of his words are eliminated in the present College Board List, it is
unlikely that his list vould give one much more than 81.2 per cent of
the vocabulary of the wide range of authors sampled in this study, and
this result is attained at the expense of learning far more words than
are really necessary to attain it.
This part of our study also yielded a list of 69 words which appear
in the College Board List, but which did not appear five times or more
in the literature examined In this study, and were therefore omitted,
This list may be used either to supplement the present list, since
presumably these words are important in certain of the authors read in
secondary schools, or to eliminate them from the College Board List,
if one wishes to learn only the words in it which are likely to have
the widest ultimate usefulness. All such lists appear later in our
report, and may be found by consulting the Table of Contents.
We are now in a position to discuss the inclusion of medieval Latin
in the present study, as was promised earlier. This was included
because of the growing recognition of its worth and charm as easy and
interesting reading for beginners. The writer used it extensively in
his classes, especially during the second year, and could not fail to
be impressed by its popularity with students. This impression was
confirmed by the writer's contact, in the course of an experimental
study, with Latin teachers in all parts of the country who had used
medieval Latin with great success, and wanted more of it. Such
authentic, interesting reading materials for beginners are desperately
needed for the "reading method," since we have in Latin almost no
other source of easy reading that is worth reading at all. Caesar,
the first classical author commonly read, is so difficult in
vocabulary, syntax, and maturity of content that fewer pupils than
Helvetians survive. On the other hand, most of the modern Latin
written for beginners is so puerile in content and so barbarous in
style that there is no point in reading it at all. Even the illusion
that it affords practice in reading Latin is shattered vhen one
encounters real Latin, which seems to have been put together on
utterly different principles. Every teacher knows the sensation, as
of striking deep sand, when his class, after romping through "made"
Latin, first encounters real Latin. Where the pupils have been
reading pages, now they read lines.
Medieval Latin offers a bridge between these extremes. Its content
is attractive for young readers: richly musical hymns, student songs,
drinking songs, stories of knights and crusaders, stories of life in
castles and monasteries, romances, beast fables, humorous stories,
diatribes against women, portions of the Vulgate, and penetrating
comments on religious problems with which adolescents are deeply
concerned. It affords an insight into a period which we call the Dark
Ages much as we called Africa the Dark Continent because so few of us
know anything about it. It abounds in men of real religious genius,
who had something important to say about the religious problems of
young people today. It shows what happens -when a civilization really
declines, and how that civilization may find itself again. It affords
by contrast real insight into what the scientific method of thinking
means in our culture, as pupils learn how the most brilliant men
thought during the Middle Ages, proceeding logically from authority to
absurdity. It shows the curiously haphazard manner in which our
modern system of education got established, and how the vaunted
curricula of our institutions of higher learning often rest upon needs
which ceased to exist in the thirteenth century. It has in addition
the glamor of the unexplored, of a literature which few living men
know anything about, and which 4 s not available in translation.
Small wonder that pupils love It!
This literature is written in authentic Latin, by men who spoke and
thought in Latin, with the ring of genuine communication in it. Yet
it Is easy to read-not all of it, to be sure, but substantial portions
of it, which Beeson, Harrington, Waddell, and other anthologists have
culled for us. Its sentence structure, in general, is simpler; its
word order approaches that of modern languages; and where its syntax
offends the purist, it is usually in the direction of simplicity.
"Dixit quod" followed by the indicative; that is just vhat our pupils
write before they have mastered the infinitive in indirect discourse.
When a word differs at all from its classical meaning, it Is also in a
direction which is easier for the modern to understand, such as
pastor, meaning "pastor."
When a teacher shudders at such deviations from the pure
Ciceronian, he is shuddering at a ghost raised by the humanists of the
Renaissance. Then it happened to be the fashion in speaking and
writing to imitate the forms of the Golden Age of Latin literature,
and to despise later Latin as corrupt and impure. When it was a
professional asset to speak and write pure Ciceronian and a social
disaster to lapse from this standard, it was natural for pedagogues to
protect their pupils from the influence of medieval Latin. To some of
us this deliberate archaism, disregarding the later development of the
language, and substituting imitation for the impulse to communicate,
stamped out the dying embers of Latin literature. However that may
be, the need for protection against the contamination of medieval
Latin certainly ceased before the Pilgrims landed in this country.
Our pupils have no need to speak or write Ciceronian Latin. They will
enjoy reading it, but they can only be helped on their way toward it
by Latin literature which is equally authentic but simpler, and closer
to our own. The uncouthness of this literature is exaggerated by
people who have never read it. Most of what is presented for the
general reader in the popular anthologies is respectable, living
Latin-far better for every educational purpose than the artificial
Latin of modern textbooks.
In view of the need for easy and interesting material for beginners
that is worth reading, and in view of the almost inexhaustible supply
of such material in medieval Latin, it became one of the major
objectives of this study to find out whether the vocabulary of
medieval Latin differs so markedly from that of classical Latin as is
commonly supposed. We are glad to report that such is not the case,
at least in such selections as are presented in Beeson's "Primer of
Medieval Latin," which are fairly representative of the other popular
anthologies. There are four major sources of evidence on this point.
The first is simply the reading of our data on the frequency of
occurrence of our 3,800 words in the three anthologies. There are
discrepancies, to be sure, but the overwhelming majority of words
which are important in classical Latin are also important in medieval
Latin, and those which are unimportant in the former are also
unimportant in the latter. We tried to find the correlation between
the frequency of these words in medieval Latin and their frequency in
the other two anthologies combined, but the data did not lend
themselves to this form of mathematical expression. (They do not
cluster around a mean but are pyramidal in form: a tremendous number
of words in the lower frequencies, diminishing in number as rapidly as
they increase in importance in the higher frequencies. This threw
most of our tallies into the lower left hand corner of our scatter
diagram.) In the absence of any such precise mathematical expression
of the similarity of our figures on medieval and classical Latin, we
leave it to the reader to discover the similarity for himself.
The second source of evidence on this point is the validation of
our results against the ten representative classical authors, as
previousl~ described. The reader may recall that the words which
occurred twenty times or more in these three anthologies combined
(heavily weighted, therefore, with medieval Latin) comprised 85.1 per
cent of the total number of words studied. We then dared to predict
that 85.1 per cent of the words in a random sampling of classical
authors would appear in this list. We tried it on ten short passages
from representative classical authors, and 84.4 per cent did appear in
our list. As few as 82.9 per cent and as many as 87.3 per cent might
have appeared in such a small sample by the laws of chance without the
deviation from prediction being significant. If the time at our
disposal had permitted a larger sampling, a smaller deviation from
prediction might have been expected, but since this was close enough
for all practical purposes, we did not think it worth while to carry
the investigation farther. This means that the Latin words in a list
heavily weighted with medieval Latin will occur in just about the same
proportions in classical Latin alone.
The third source of evidence on this point is that derived from our
analysis of the effectiveness of the Latin Word List of the College
Entrance Examination Board. Here we have precisely the opposite
situation: a word-list derived exclusively from six classical authors
of the Golden Age, applied to a random selection of medieval Latin.
The reader may recall that this word list comprised 84.5 per cent of
the words in the classical prose, 80.7 per cent in classical poetry,
and 79.6 per cent in medieval Latin. Hence this list, never designed
for medieval Latin, came within one per cent of being as good for
medieval Latin as for classical poetry.
The fourth and last source of evidence on this point is that we
finally went over our recommended basic vocabulary of 1,556 words
which occurred twenty times or more in the three anthologies combined,
and pulled out all words which occurred less than ten times in
classical Latin alone (i.e., those which were important chiefly in
medieval Latin). There proved to be only 85 such words, excluding
proper names. Learning 85 extra words out of 1,556 would be a small
enough price to pay for the additional benefits to be derived from
medieval Latin. Even so, we put these 85 words in a separate list, so
that teachers may use them or not at their pleasure. The remainder of
our recommended basic vocabulary will work equally well in classical
or in medieval Latin. This "purified" list slightly increases the
correspondence between prediction and actuality in the validation of
our results against the ten classical authors, but not enough to tax
the reader with the results.
We believe that these four sources of evidence establish the case
for the inclusion of medieval Latin. It affords practically as good
preparation for the vocabulary of classical Latin as any similar
amount of classical Latin. In case any teacher disagrees with us,
however, he need not shun our results on that account. We have listed
the frequency of occurrence of each word in the three anthologies
separately. In using this list he may disregard the third column if
he chooses. We still have sampled 124,686 words from 145 authors of
classical Latin, as against 77,142 words of three authors in Lodge's
word count. Our recommended basic vocabulary (now of 1,471 words) has
been purged of all words which are important chiefly in medieval
Latin, and comprises 83.6 per cent of the words in Latin literature,
while the Latin Word List now in use has 1,791 words, and comprises,
according to our findings,
only 81.2 per cent of these words. If one wishes to read other
authors besides Caesar, Cicero, Vergils Ovid, Nepos and Sallust, he
cannot go wrong in making use of our list, even if he abhors medieval
Latin. In using our list, it is not expected that a student, however
mature, will memorize the words one by one, out of context. Rather it
is expected that the list will be used chiefly to write or rewrite
materials for beginners in a controlled vocabulary, or to distinguish
the words in further reading which are important to remember. For
these purposes the list has been arranged in groups of related words
which can most easily be learned together, as well as in alphabetical
order. This will also assist the learner in reviewing the words he
has learned, and in getting them in good order in his mind. We hope
this list will advance us farther toward the goal of enabling the
mature reader in a single year, or the secondary-school pupil in two
years, to read a Loeb translation intelligently, and to utilize the
resources of medieval Latin in his progress.
The second purpose of this study was to discover the relative
frequency of the Latin word endings. The writer's pupils were balked
in their efforts to read Latin by the colossal apparatus of the
traditional Latin grammar. We who have long since mastered this
apparatus do not often realize just how colossal it is. To regain
this awareness, graduate students in the classes of Professor Carr of
Teachers College, Columbia University, make an annual count of the
number of inflected forms to be memorized in high school Latin
textbooks published since the Classical Investigation. It is one of
their most illuminating exercises in methods of teaching. At the time
of the writer's participation in this devastating inquiry six years
ago the average number of inflected forms to be memorized in eighteen
of the most popular first year Latin textbooks was 1,572. The range
was from 780 to 2,800. This did not include the inflected forms
printed In the grammatical appendix; it was restricted to the forms
printed in bold-face type in the body of the text with the directions
expressed or implied, "Learn this by heart."
These forms, of course,
are never mastered during the first year, or anything like it. The
Classical Investigation revealed the fact-as witnessed by the table on
page 92 of its General Report-that pupils on the average throughout
the country learn about half of the forms they are expected to learn,
even as measured by standardized tests of the most common forms. Even
when the exceptional pupil attains a respectable mastery of these
forms, they are organized in the worst possible way for reading. When
a pupil encounters the ending a, for example, he has to think through
five declensions of ten forms each before he can locate the possible
interpretations of this ending. With practice he can short-circuit
this process, of course, but the rote learning of forms in paradigms
creates precisely this difficulty. This is quite natural inasmuch as
the traditional conjugations and declensions were never organized for
reading purposes, but for speaking and writing Latin. The traditional
definition of grammar is "ars scribendi et dicendi". Since pupils no
longer have occasion to speak and write Latin, it is high time that we
evolved a new Latin grammar: a system for the interpretation of word
endings that may properly be called "ars legendi." Such a system would
associate directly with the word endings commonly encountered in
reading all possible interpretations of these endings. These
interpretations would not include such meaningless concepts as
"dative" or "ablative." The grammars commonly list more than ten uses
of the dative and more than twenty uses of the ablative. Always in
the plural and about half the time in the singular the dative and
ablative endings are identical, so that the pupil cannot tell which is
which. This means that when a pupil has finally located the dative
and ablative among the possible interpretations of an ending like is,
he must in theory think through more than thirty possible uses of
these cases before he can tell the function of this word in the
sentence. The only way to distinguish among several of these possible
uses is first to discover by intuitive divination what the sentence
means. With this in mind it is possible to diagnose the datives and
ablatives. The meaning affords a clue to the grammar, rather than
vice versa. It is obvious that such a colossally inefficient
instrument of interpretation is almost never used by beginners in
learning to read Latin. They learn an acceptable smattering of the
facts of Latin grammar on the side, as one of the mysterious
visitations of Providence upon them for the sins of their elders, and
pick up whatever skill in reading Latin they can through native wit.
As a step in the direction of organizing a system for the
interpretation of the Latin word endings commonly encountered in
reading, the writer selected a thousand running words from each of ten
classical authors of prose and verse, and classified them according to
their endings. The following classes of endings were found to occur
in the following proportions:
These findings were checked by 5,000 words of Latin prose and 5,000
words of Latin verse selected at random from the word count previously
described. They have since been checked many times over by students
in the writer's classes in methods of teaching Latin with
approximately the same results. In the table just given the 18
"common" endings included all endings which occurred on one per cent
or more of the words studied. The 22 "rare" endings included all
endings which occurred more than four times but on less than one per
cent of the 10,000 words studied. "All other endings," none of which
occurred more than four times among these 10,000 words were rejected
as of negligible importance in reading, even though they included such
endings as the second person passive, singular and plural, and the
genitive plural in -uum. The latter in Its rare occurrences has
regularly been mistaken by the writer's pupils for the genitive in
-um, so that it is interpreted correctly in spite of its omission.
The 18 common endings with their interpretations and the proportion
in which each is found to occur in classical prose and verse are as
follows:
*One family of Latin words, called the third declension, ends in
almost every conceivable way in the subject form. Such endfngs always
indicate or refer to the subject except in neuter words, when they may
also indicate or refer to the object. If any leg-ular ending does not
fit the interpretation given above, look it up in a dictionary to see
whether it may not be the subject form of such a word.
**Differs from this interpretation In reporting what someone says,
thinks, knows or perceives. Occasionally in poetry re means subjects
acted or you, acted upon.
At first glance this table looks impossibly complicated to one
trained in the traditional conjugations or declensions. Try, however,
to put yourself in the place of a student who has not been so trained,
who has been reading easy Latin for about three months without any
systematic instruction in what the endings mean, and who has gradually
become aware that Latin words assume a bewildering variety of forms to
indicate their function in a sentence. He is now grappling with
really difficult sentences which cannot be understood without knowing
what the endings mean. He is left to flounder just long enough to
gather the impression that the number of endings is infinite, and that
they mean so many different things in different contexts that he will
never be able to learn them. Then he is told that to master all this
intricate, bewildering system he has only to perform five definite,
limited tasks:
1. To learn by heart 18 common endings and their interpretations,
which will enable him to tell, within certain limits, the function in
a sentence of 92.5 per cent of all the words he will ever read in
Latin. (These endings occur on 66.5 per cent of the words, 23.4 per
cent have no endings, and 2.6 per cent are irregular pronoun forms
which he will learn as vocabulary.)
2. To learn 8 rules governing variable endings which will tell
him, in so far as any system of grammar is able to tell him, which
interpretation of an ending which means several different things is
most likely to apply in a given context.
3. To learn 22 rare endings and their interpretations, which will
enable him to tell, within certain limits, the function in a sentence
of 6.8 per cent more of all the words he will ever read in Latin, or
99.3 per cent in all. None of the few remaining endings will occur on
the average more than once in 5,000 words.
4. To learn 14 "penuitimate signs" which occur between stem and
ending in 7.1 per cent of all the words he will ever read in Latin,
which will enable him to tell a few other important things about them.
5. To apply these 40 endings, 14 penultimate signs, and 8 rules in
reading about a thousand pages of Latin, to find out by experience how
they work. No system of grammar can give more than a clue to the
meaning. It takes long practice to respond to these clues instantly,
correctly, and almost unconsciously, and to recognize at a glance the
patterns into which Latin words usually fall.
Students grasp at these findings as a drowning man grasps at a straw.
Once again it begins to seem possible to learn how to read Latin.
There are only 40 endings and 14 penultimate signs to learn, rather
than the 1,572 inflected forms which have to be memorized in the
average first-year textbook. Even this task is simplified by the fact
that we never learn these 54 inflected forms outright. After the
first three months of reading Latin without benefit of inflections we
find that it restores the students' confidence to memorize the 18
common endings, which really carry about 90 per cent of the burden of
Latin grammar. It takes only an intensive drive of one week to get
all pupils to the point at which they can give our interpretations of
any of these endings instantly, automatically, and without ever making
a mistake. This is all the grammar we learn systematically for a long
while to come. We paste the tables of rare endings and penultimate
signs and their interpretations and the rules governing variable
endings in our readers, and use them only for reference. When we
finally come to the point of learning them systematically, most of us
find that we know them already, and need only to nail them down for
good and all with a little practice. With this situation in mind, let
us return to our table of the 18 common endings and see what it means.
In this table "subject" means that the word to which this ending is
attached is the subject of a sentence or clause, or modifies the
subject. "Object" means that the word to which this ending Is
attached is (or modifies) the object of a verb or preposition, or the
subject of an infinitive (which we explain as an extension of the
object idea, as is historically true). "Adjective" means that the
word to which this ending is attached is (or modifies) the equivalent
of an ' adjective phrase in English introduced by "of," usually
modifying the nearest appropriate noun or pronoun. It should not be
confused with a true adjective, which has the same type of ending as
the word It modifies - which is why we have to keep saying "is or
modifies." We impress upon students that this -interpretation means
adjective ending, not a regular adjective. We trust that you
recognize the genitive in this guise. "Adverb" means that the word to
which this ending is attached is (or modifies) the equivalent of an
adverb phrase in English, usually modifying the nearest appropriate
verb, adjective or other adverb, and telling how, when, where, why,
etc. Again it is not a true adverb, except that e occasionally does
indicate a true adverb-less frequently, however, than it indicates
such an adverb phrase as we have described. This interpretation holds
in most cases for either the dative or the ablative. Plural
interpretations like "subjects" indicate that the word is plural. The
interpretations given of the verb-endings are probably self
explanatory.
The interpretations given for the endings were the product of a year
of experimentation. They are not completely satisfactory but are more
satisfactory in practice than the conventional interpretations as
"dative," "ablative," and the like, which may mean thirty different
things. The ideal was to attach directly to each ending all its
possible interpretations in terms which would not require further
definition. At first we looked for terms which would express the
function in a sentence of the subject, the verb, the object and the
modifiers. Many terms were tried out, such as "actor," "act,"
"recipient," and the like, but they were discarded because they proved
to be more confusing to pupils than the corresponding grammatical
terms to which they were already somewhat accustomed, and because we
wished to develop a feeling for these elements of a sentence as a
basis for the comparative study of English sentence structure. If
they did not know these grammatical terms already, we taught them.
The nominative and accusative were easy enough to translate into the
familiar English equivalents of subject and object. Words in
agreement with these elements of the sentence were accounted for by
saying that these endings indicate or m dif the subject or object.
Most genitives were found to modify the nearest appropriate noun or
pronoun like an English adjective phrase introduced by "of," so the
interpretation "adjective" was adopted. This did not account for the
irregular use of the genitive with certain verbs. We expected to
teach this usage as an exception to the rule-as it must be taught even
in traditional grammar-but pupils became so accustomed to the "of"
relationship that they usually carried it over to these verbs as they
do to the corresponding verbs in English without sensing the violation
of the rule.
We could see no point in distinguishing between the dative and the
ablative when these endings are never distinguished in form in the
plural and only about half the time in the singular. This seemed to
indicate a close similarity in function between these forms even in
the Roman mind. The logical English equivalent of both is an adverb
phrase, modifying the nearest appropriate verb, adjective, or other
adverb, telling how, when, ),here, etc. Hence the interpretation
"adverb" was adopted for dative and ablative indiscriminately. Even
the indirect object is adverbial in idea. It tells where something is
told, given, or shown. A check of all the standard grammars and the
use of this interpretation in teaching for three years has convinced
the writer that the adverbial relationship is more effective than any
other that ingenuity has been able to suggest up to the present time
in bringing home to pupils the fundamental idea of the dative and
ablative alike. It is not invariably true, but exceptions are fewer
in number and frequency of occurrence than in the traditional
interpretations.
As an alternative to teaching more than thirty different uses of these
two cases it has obvious advantages. It can really be used in reading
rather than in reasoning about sentences that have already been read.
Both the "adjective" and "adverb" interpretations have the additional
advantage that they force the pupil to put together the words that
belong together. Genitives normally go -with the nearest appropriate
noun or pronoun; datives and ablatives with the nearest appropriate
verb, adjective or other adverb. When they are thrown into this
functional relationship the meaning usually becomes clear.
Plurals were indicated by making the Interpretations plural, such as
"subjects." This 'was easier to learn and to remember than two words,
such as "subject plural." Since the objective was to build up
associations with word endings which would help and not hinder speed
in reading, the shortest possible interpretation was preferred-even at
the cost of euphony.
In deciding which Interpretation of a "variable" ending to select, we
advised pupils first to consider which-one made the best sense. If
this was not clear, we advised them to consult the dictionary to find
the dictionary forms of the word, or to recall them from memory. They
knew that the dictionary gave the subject and adjective forms of each
word that had variable endings except I as a verb-ending. Then they
consulted the following table, which was pasted in the back cover of
the dictionary or reader for ready reference:
Further distinctions among the varying interpretations of these
endings must be made, even in traditional Latin grammar, upon the
basis of the context and agreement between subject and verb, adjective
and noun, pronoun and antecedent.
These rules seem confusing in the aggregate but are clear enough
individually. They need be consulted only in those rare instances in
which sentences will not yield to direct interpretation from the
meanings of the words. The experience of the writer indicates that
each rule will be consulted even by beginners not more than once in a
hundred lines of Latin. As pupils learn through reading the patterns
into which Latin words usually fall, they gradually dispense with the
rules altogether. We have not found it necessary to require the
memorization of these rules.
The 22 rare endings with their interpretations and the proportion in
which each occurs in classical prose or verse are as follows:
The experience of the writer indicates that it is unwise to attempt
the memorization of the "rare" endings before the "common" endings
have been thoroughly mastered as a tool of interpretation which means
that they must be recognized and interpreted automatically in reading.
This does not usually occur before the latter part of the first year
or the first part of the second year. This does not mean that these
"rare" endings should be omitted in reading or in classroom
conversation. The sentences in which they occur can usually be
understood directly from the meaning of the words or by the analysis
of the "common" endings. If the meaning hinges upon the
interpretation of a "rare" ending, the writer prefers to give the
pupils the correct interpretation with the understanding that it is to
be studied later on in the course.
It is an interesting commentary upon the utility of the organization
of 1,572 inflected forms in paradigms that the writer, who is probably
not far below the average in teaching skill, is unable to get his
pupils to use more than 18 endings instantly and automatically as a
tool of interpretation in reading before the end of the first year.
It is also interesting to note in passing that this study brought to
light two endings, d and ius, more important in terms of frequency of
occurrence than most other "rare" endings, vhich have been neglected
by the traditional grammar. Pupils have never regarded d as an
ending, but as the final letter of a host of queer pronouns that they
never remember anyway; but if they become sensitive to this final
letter as indicating either subject or object, it will clear up almost
as many difficulties in interpretation as some of the "common"
endings. Similarly ius has been treated chiefly as an irregular
genitive; later as a way of forming the comparative of adjectives and
adverbs. Ask any pupil trained in traditional Latin grammar - to name
four different things which the ending ius may indicate about the
function of a word in a sentence, and he will pause a long while
before answering. Since it occurs more frequently than almost any
other "rare" ending, it is well to know immediately that it indicates
or modifies the subject or object, or an adjective phrase modifying
the nearest appropriate noun or pronoun, or is a comparative adverb
like "more ingelligently." The second person present imperative also
occurs more frequently than previous studies would lead one to expect.
The scientific study of Latin grammar as the ars legendi will not only
eliminate the study of unnecessary forms; it will bring to light ways
of interpreting the functions of words In a sentence that the
traditional grammar has disregarded.
The "penultimate signs" are explained to students somewhat as follows:
Between the stem and ending of verbs and adjectives there are
sometimes a few extra letters which give a special twist to the
meaning. We call these additional letters "penultimate signs,"
because "penultimate is a Latin word for "next-to-the-last." There are
14 of them which are important, and they occur in about 7 per cent of
Latin words. They are:
1. On the present stem of verbs:
(The present stem is approximately the first form given in the
dictionary or in our basic vocabulary.)
2. On the Perfect stem of verbs:
(The perfect stem is the second dictionary form of verbs in our basic
vocabulary except in verbs marked dep. (deponent) which have no
perfect stem.)
3. On the last dictionary form of verbs (the perfect passive
participle):
The perfect passive participle is most commonly used with some form of
the verb sum (to be) to make the perfect tenses of the passive voice;
e.g.,
It is also commonly used with an adverb ending to indicate some
circumstance attending an event in the fewest possible words; e.g...
Otherwise it Is a simple verbal adjective, past time and passive
voice; e.g.,
The only penultimate sign affixed to the stem of the perfect passive
participle is -ur-, which changes it from passive to active and gives
the idea of immediate future action; e.g.,
4. On adjectives:
A few adjectives ending in -er like acer add -rim- to form the
superlative; e.g., acerrimus, sharpest. A few others change beyond
recognition in the comparative and superlative forms - just as in
English, while most adjectives go as lon longer, longest, a few go as
good, better, best: (bonus, melior, optimus). In all such cases
the imparative and superlative forms of common irregular adjectives are
given separately in the Word List and should be learned as vocabulary.
It has already been pointed out in the list of Rare Endings that -ius
is the neuter subject and object ending of comparative adjectives and
also the ending of comparative adverbs. In these cases -ior- is
omitted before the ending; it is fortius (more bravely, or braver),
not fortiorius.
The superlative form of adverbs adds -e to the stem of the superlative
of the corresponding adjective: e.g., fortissime, most bravely.
It must be obvious to the experienced teacher that such tables are
given to pupils only as skeletal framework for organizing and
remembering a great deal of concrete experience with the language.
None of these tables would have any meaning for pupils if memorized as
it stands, without the wealth of concrete illustration and practice
which a teacher is able to provide. In other words, this system of
interpretation has to be taught, like any other system, and not simply
handed to pupils with the injunction to learn it. It is not fair,
therefore to look at any sentence in such explanations as are given of
the penultimate signs and to object that it is not clear as it stands.
Pupils find it clear enough after they have gone through the
experiences on which it is based, and it serves as the briefest
possible statement of the case for further reference. Neither Is it
fair to object that a few of the interpretations are not technically
correct from the standpoint of traditional grammar. For example, the
first example of the first "penultimate sign" is -a- in the present
tense of the first conjugation. In the traditional grammar this is
not a tense sign but a part of the stem. We prefer to treat It as
though it were a tense sign, however, to distinguish it from the
present subjunctive. Nor is it fair to bring up the few possible
exceptions to any of the interpretations given, unless evidence of the
frequency of these exceptions is also presented. It is recognized
that none of these explanations will hold in 100 per cent of the
cases, but they explain so large a percentage of the difficulties in
interpretation which pupils ordinarily encounter that they are worth
learning by heart, in the course of a large amount of reading. The
exceptions can be dealt with as they occur. Remember that we have
presented a complete new system of Latin grammar for beginners in a
very few pages. The reader must supply the necessary explanations,
Illustrations, practice exercises, exceptions to these rules, and
other details from his imagination and teaching experience.
This system of interpretation obviously applies only to reading, not
to writing or speaking Latin. Since pupils almost never have occasion
in later life to write or speak Latin, practice in these skills is
usually justified in terms of its contribution to the ability to read
Latin. No proof has ever been given that such practice contributes
enough to the ability to read Latin to justify the learning of the
traditional apparatus of Latin grammar which it entails. The usual
argument, in fact, is devoted to proving that practice in writing and
speaking Latin assists in the mastery of the traditional grammar. Of
course it does, since this grammar was devised for purposes of
speaking and writing, but this is no proof that such practice assists
reading. The only evidence on this point that the writer has seen is
an exploratory study of his own while a student at Teachers College in
1931. He tested 65 of his classmates on all the elements
traditionally included in Latin courses and found the correlation of
each element with the ability to read Latin as tested by a good
comprehension test. The highest correlation discovered was between
the ability to read Latin and the ability to interpret forms-even by
the traditional designations: .63 ñ .052. The lowest correlation
discovered -was between the ability to read Latin and the ability to
translate English into Latin: .35 ñ .077. This evidence must not be
taken too seriously since the number of cases was small and not
representative, but it offers some corroboration of what a
psychological examination of the reading process would lead one to
expect: namely, that the ability to write Latin has very little
relation to the ability to read it. It may be useful in an advanced
study of language, as in English universities, but how many American
students ever get to the point of using it for this purpose? Their
needs may be met in the later years of their Latin course, as we
explain tn the last paragraph of this introduction.
Practice in speaking Latin is often defended on the ground that words
and phrases learned through the eye and ear together make a stronger
impression and are remembered better than words and phrases learned
through the eye alone. No convincing evidence has been offered on
this point, but the writer's faith in the oral method was shaken when
he made up two artificial languages: one of nonsense syllables and one
of random designs on paper which could not be pronounced, and which
resembled nothing that could be pronounced. Both symbols and
syllables were assigned meanings it random by drawing three-letter
words out of a hat. All of the writer's pupils, numbering about 75,
from the eighth grade up to college seniors, undertook to learn as
many "words" of each language as they could in five minutes. While
they studied the nonense syllables, they were instructed to pronounce
them softly, tilizing all the resources of throat and ear to memorize
them. hen they studied the symbols they were instructed not to
pronounce even the meaning assigned to it but to associate one visual
symbol with the other. So far as the writer could tell, these
instructions were faithfully carried out. Each class was divided into
two equivalent groups, one of which studied the symbols first, the
;her the syllables first. At the end of five minutes both groups )re
given a test consisting simply of the symbols or syllables in
different order than the one in which they had been studied, and were
asked to write out the meanings which had been assigned to each. This
was followed by another learning period in which the two groups
exchanged their tasks, and by another test. Every group in the
experiment learned the symbols better than the syllables by a wide and
significant margin. This was contrary to the general expectation and
aroused considerable discussion as to what it meant. It was agreed
that the evidence was too flimsy to base a theory of learning upon it,
but it put the burden of proof upon those who argue that speaking
Latin is an invaluable aid in remembering words and phrases. The
writer still uses Latin in routine classroom conversation for comic
relief and for motivation, but he expects grammatical correctness in
the response only by imitation or good luck. He never argues that
this by-play is materially assisting the process of learning to read
Latin, and he would never think of arguing that the traditional
apparatus of Latin grammar should be taught because it is necessary in
speaking Latin, nor that practice in speaking is important because it
fixes in mind the traditional grammar.
This system of interpretation does not require pupils to distinguish
the quantity of vowels in such endings as a and is. The writer felt
that it would be unwise and unnecessary to make his pupils dependent
on macrons, since they are printed only In recent high school
textbooks. It is impossible to assemble even an adequate amount of
easy and interesting Latin with macrons to carry on the reading method
in high school, to say nothing of the continued reading of Latin for
pleasure. If macrons are not printed in the text, just how does It
help a pupil to know that if the ending -a is long, it is "adverb,"
while if it is shorts it may be "subject," "subjects," or "objects,"
depending on whether the subject form ends in -a? The pupil has to
discover from the context which of these functions the word has in the
sentence; then he can tell whether the final -a is long or short. But
if he already knows the function of the word in the sentence, why
should he want to know the quantity of the vowel? Even if macrons are
printed, they only eliminate one possibility. There are at least ten
times as many short -a endings as long -a endings. In all of these,
the pupil still has to choose among three possible interpretations: if
the word is of the first declension, it is nominative singular; if it
is not, it may be nominative or accusative plural, In our experience
pupils find it no harder to follow the rule, "if the subject form is
-a, it may be subject or adverb; if not, it may be subjects or
objects." And that rule works whether there are macrons or not.
It may be advisable to take up a few other objections of the defenders
of the teaching of quantities. "How do your pupils distinguish
between venit, present, and venit, perfect?" If the other main verbs
in the same passage are present, they naturally and unconsciously read
it as present; if not, as perfect. "How do your pupils distinguish
between ' levis, light, and levis, smooth?" By the appropriateness of
the adjective to the word which it modifies. If either would be
equally appropriate, then it makes no difference. "How do your pupils
learn pronunciation?" By imitating the teacher, by daily practice and
correction, as in learning any other language. "But how do they learn
to read verse?" They read it chiefly in the fourth year, by which time
some study of quantity is admissible without confusing the initial
attempts of the pupil to get meaning from the printed page. They have
not memorized their quantities by this time, but neither have pupils
taught by the usual method. Instead, they have had three years of
practice in pronunciation that is fairly correct by-modern standards,
not a bad foundation for the reading of verse. One other
consideration should be noted. This system of interpretation is
frankly designed for the reading method, and no Latin teacher can use
this method at the present time without recourse to large amounts of
mimeographed material. It is hard enough to get out correct copies of
this material, and the insertion of macrons makes it practically
impossible. That is the practical reason which led the writer to
devise ways to get along without them. The final test is whether the
pupils can read Latin well without any explicit knowledge of
quantities. Since the pupils taught exclusively by this method read
on the average four times the usual amount, and were tested
individually on every selection they read, the writer is satisfied
that it can be done.
An outline of the writer's method of teaching this system of
interpretation may be helpful in showing how the results of this
research may be utilized, since the system obviously calls for such a
drastic readjustment of the usual content of a course in Latin that a
teacher may 'well wonder how to teach it. This part of the report is
not a presentation of the results of research, but of teaching
experience, and as such it should be judged by different standards.
Scientific accuracy is claimed only for the data on the frequency of
Latin words and their endings.
The principal activity in the writer's
classes is the reading of new Latin, usually in small, congenial
groups seated around a table. Each group chooses with the aid of the
teacher a selection in mimeographed form from the files, or a small
book from the shelves, that is related to their interests and adapted
to their level of ability in reading. They elect a chairman for the
day who reads through each sentence sentence slowly in Latin. After
each sentence the pupils in the group ask one another for help on any
points which they have not understood. If no questions are asked, the
chairman proceeds with the next sentence. The meaning of words is
usually taken up first. Each pupil mentions the words he has not
understood. As he mentions them, if any other pupil in the group
happens to know the word, he tells the meaning, and, If possible, the
dictionary forms of the word. If no pupil knows the word, it is
underlined, and other unfamiliar words are mentioned until all the
words in the sentence which no pupil in the group knows are
underlined. The chairman then assigns one or more words to each pupil
to look up in a dictionary. All look them up at the same time, thus
dividing the labor of thumbing the dictionary by the number of pupils
in the group. After each has reported the meaning and the dictionary
forms of his word, the chairman usually reads through the sentence
again in Latin. In most cases the meaning will be clear, but if any
pupil still has difficulty in putting the words together to make
sense, he asks the group about the part that troubles him. Help is
usually forthcoming on the basis of the context or of a rule that
someone has remembered. If no member of the group can clear up the
difficulty, they resort to a formal analysis of the endings in the
part of the sentence that troubles them. This is understood to be the
"next-to-the-last resort," and is done very infrequently. The words
are marked in the order in which they occur with a system of graphic
devices. All words with "subject" endings are distinguished and
grouped together by some diacritical mark: usually a bracket under
these words. All words with "object" endings are distinguished and
grouped together by another diacritical mark: usually a bracket over
these words. The object of the verb may be distinguished from the
object of a preposition by an additional line over the bracket; in
extreme stages of sophistication the subject of an infinitive may be
distinguished from other words with "object" endings by an additional
wavy line over the bracket. All words with "adjective" endings are
distinguished and grouped together by an encircling line with an arrow
pointing to the nearest appropriate noun or pronoun. All words with
"adverb" endings are distinguished and grouped together by an
encircling line with an arrow pointing to the nearest appropriate
verb, adjective or other adverb. The verb is left to care for itself,
which it usually does very nicely, After this system is well
established, pupils gradually dispense with the diacritical marks;
they assign them mentally instead and proceed at once to the crux of
the difficulty, which usually hinges upon the interpretation of a
"variable" ending. If the sentence will not yield even to this formal
analysis, they proceed to the last resort of all and call upon the
teacher. He spends about half of each period going about from group
to group, stimulating their endeavors and answering their questions.
It is the only Latin classroom the writer has observed in which the
pupils ask the questions and the teacher answers them, but it seemed a
more normal procedure because the teacher knew how to read Latin and
the pupils did not. Almost all of the teacher's instruction in
grammar-as the art of putting words together to make sense-occurred
during these brief conferences while pupils were confronted with a
difficulty which they were unable to resolve by their own efforts, and
which interfered with their getting on with a story in which they were
interested.
Some of these difficulties, of course, required more elaborate
explanation and more practice exercises to make the point clear than
could economically be offered to one group at a time. In all such
cases the teacher told the group to bring up the difficulty fop
"clinical study" and gave them the meaning provisionally. We reserved
a part of almost every class period for such "clinical study" of
especially difficult sentences. Each group brought up its unresolved
difficulties one at a time. These were written on the blackboard, and
the whole class, with the help of the teacher, endeavored to discover
what the difficulty was and to resolve it for all time to come. When
the point was especially important to remember, the teacher made up
and dictated parallel sentences for the class to figure out as
practice exercises. This "clinical study" afforded many opportunities
for a comparative study of English and Latin ways of expressing the
same idea.
As soon as a group finished a story or a chapter in a book, they
applied for a test, which they took individually, either orally or in
writing. It usually included some test of comprehension, preferably
other than translation, questions on the interpretation of word
endings and questions on the meaning of new words which occur
frequently in reading. Any other elements might be included which
were appropriate to the selection read. The writer at first tried to
make objective tests on all of his stories, but this proved to be so
time-consuming that the test became usually an informal discussion
with the group around the teacher's desk. The ulterior motive of the
test was to correct the tendency of the less energetic members of a
group to take undue advantage of the activity of others in reading the
story, thus avoiding a valuable learning experience. The test made
each individual feel responsible for understanding the whole story at
the end, even though he might justifiably profit from the superior
talent of a classmate in coming to understand it. If all members of
the group did not do reasonably well on the test, they were told that
they had not really read the story and were asked to do a more
thorough job. This collective responsibility fostered group
co-operation, rather than individual competition. The test also
provided a means of diagnosis of individual difficulties and led
periodically to a reshuffling of groups, to bring together pupils who
were having similar difficulties in reading. The other beneficial
results of a test also accrued: to promote the organization and
retention of learning and to add zest by recognizing success.
After one unfortunate experience in assigning "credit" for test
scores, which soon led pupils to read for the sake of a series of
numbers in the teacher's record book, the writer divorced the taking
of tests from the idea of accumulating irrevocable and inviolate
"credit." For a long time the writer refused to keep any record of
test scores other than his memory of the pupils' particular
difficulties. After the idea became established that the tests had
nothing to do with marks or promotion but were intended only to reveal
and to correct difficulties in reading, an informal record of progress
was kept which proved helpful in revealing tendencies in development
over a longer period of time than the writer could keep in memory.
Usually not more than half the period was spent in this group reading.
The rest was devoted to the "clinical study" of difficult sentences,
and to investigation and discussion of interesting etymologies and
linguistic principles applicable to English, and of significant social
problems in the light of the experience of antiquity. Some time was
also given to reading aloud pupils' translations of stories which they
had especially enjoyed. The pupils could see some point in reading
their stories to one another if they had worked out a good
translation, or in paraphrasing them if they had not, because no two
groups had read exactly the same material; and the reading of each
group could usually make a significant contribution to some problem of
interest to the whole class.
The reading of stories aloud in separate small groups in the same
classroom had the disadvantage of being rather noisy, but this proved
more irritating to the teacher than to the pupils. As each group was
concentrating on Its own problem, it was usually not distracted by the
discussion of other groups. It had the advantage of the interest
aroused by working with a congenial group for a common purpose, not
dominated by the teacher. All reading of new Latin during the first
two years was done In class. "Rome work" consisted chiefly of
collateral reading in English.
The advantages of this system of interpretation over the traditional
grammar are obvious. It is common knowledge that the traditional
grammar is almost never learned by students in this country-except by
a handful of professors after years of teaching and research. Senior
high school teachers complain that their work is hampered because of
poor preparation in the junior high school. College teachers complain
that the basic forms are not mastered in the senior high school.
Graduate schools complain that even the highly selected students who
reach them after eight years of study do not know their grammar.
Teachers of summer courses in education have reason to believe that
even Latin teachers after years of experience do not know the
traditional grammar. In attempting to discover the correlation
between the ability to read and the ability to write Latin the writer
gave a battery of objective tests, adapted from Ullman and Smalley's
"Progress Tests in First I . ear Latin," to a group of 65 Latin
teachers, averaging better than four years of teaching experience, in
the summer session of 1931 at Teachers College, Columbia University.
The results were highly illuminating, quite apart from the objective
of the study. The scores on these home-made tests would be
meaningless to the reader in terms of accepted standards, but an idea
of the mastery shown may be gained from the fact that only five out of
the sixty-five were able to write the complete conjugation of capio
without an error, even disregarding the quantity of vowels.' The
number of such grossly incorrect forms as capiebo surpassed the most
cynical expectations of the instructors. This is a fair sample of the
attainments of these teachers in Latin grammar as revealed by three
separate hours of objective tests.
The defenders of traditional Latin grammar will find in this situation
an explanation of the mediocre attainments of Latin pupils-as revealed
in the nationwide survey of the Classical Investigation. Get better
trained teachers, they will say, and all will be well. It would be
interesting to know how they propose to get better trained teachers.
The mastery of the elements of grammar has been the sine qua non of
Latin teachers for a thousand years. They have exhausted their
ingenuity and their patience in the attempt to drive the conjugations
and declensions into the heads of recalcitrant youth. Is there
anything in the present situation in education which would lead one to
expect a marked improvement in the near future? On the contrary, the
diminishing respect for formal drill in secondary education is likely
to have precisely the opposite effect. We may count on a diminished
rather than an increased mastery of Latin gramm r as the years roll
on. How, then, may the few pupils who really ought to keep alive the
classical tradition In our civilization take their initial steps in
learning to read Latin?
The system of interpretation of the 18 basic Latin endings presented
in this study offers a way out. If the writer were content with
traditional tests of interpretation-ability to give the possible
interpretations of isolated words in objective tests-he could
demonstrate conclusively that the 18 basic endings can be memorized to
the point of absolute mastery, of never making a mistake, by
practically every member of an average class in one week. This is not
the real test, however. In the course of their first year of study
Latin pupils should acquire a method of attacking difficult Latin
sentences-of interpreting the function in a sentence of the words in
the order in which they occur-that 'will function instantly,
automatically and almost unconsciously. The writer is unable at
present to prove that this occurred in his classes, since he has been
unable to devise a convincing test of this subconscious analysis of
Latin sentences. His case therefore must rest upon the likelihood
that a system of interpretation, easily held in mind, that attaches
directly to each ending all its possible interpretations in terms
which do not require further definition, is more effective in reading
Latin than a system which was never devised for reading Latin, which
scatters the possible interpretations of each ending among fifty or
more forms in paradigms, which interprets the function In a sentence
of these forms in such ambiguous terms as "dative" and "ablative"
which may mean thirty different things, and which is never mastered
even after years of study.
A second advantage of this system of interpretation is that it
relegates the study of grammar during those all- important first two
years to a properly subordinate place. The Classical Investigation
revealed the fact that 69 per cent of the pupils who begin the study
of Latin drop out in the course of the first two years. If they have
1,572 inflected forms to memorize during the first year and more in
the second, it is obvious that they will learn little else- and
equally obvious that they will never learn these forms. Any
substantial progress toward such broad objectives as literary culture,
language consciousness and a social outlook is out of the question.
On the other hand, if they attempt to master only 18 basic endings in
the course of the first year and 22 more in the second, they will
actually be unconscious of the fact that they are studying Latin
grammar. The writer's pupils, stimulated by their parents' concern
that they have never heard of conjugations and declensions,
continually clamor for the study of grammar. if progressive education
consisted Of giving Pupils whatever they wanted, regardless of why
they wanted it, the writer 'would long since have capitulated. As it
is, he Is at great pains to convince his pupils that they have
actually studied Latin grammar in a more carefully thought out and
effective fashion than any of their contemporaries. The time saved is
devoted to extensive reading, to the study of etymologies and
linguistic developments that really have something to do with English
(which the repetition of amo, amas, amat decidedly has not!) and to
the investigation and discussion of significant social problems in the
light of the experience of antiquity. The pupils read entirely in
class time on the average about four times as much as is recommended
by the Classical Investigation as a desirable minim=, and during one
trial period of three months the fastest pupil read fourteen times as
much as the slowest.
After the second year a carefully selected class consists chiefly of
those enthusiasts who 'wish to learn all there is to know about Latin.
These pupils may be taught more and more of the traditional forms
without harm, since their reading habits are already well established,
and by the end of the four year course, if the instructor is
interested in this outcome, they will be able to impress their college
instructors with their knowledge of "grammar," and even with their
ability to write Latin. Such knowledge, however, is appropriate only
for specialists, and can only be acquired with satisfaction and
success upon the basis of at least two years' direct experience with
the language. The first two years of a high school course in Latin
are no place for it.
PERCENTAGE OF WORDS IN TEN PASSAGES FROM CLASS1CAL AUTHORS
WHICH OCCURRED IN THE FOLLOWING RANGES OF FREQUENCY IN THE
ANTHOLOGIES
> 100 50-99 20-49 10-19 5-9 Ommited
Percentages
Caesar, Gallic-War, II, 1 60 10 11 8 7 4
Cicero, de Amicitia, I, 1 65 10 12 6 1 6
Livy, Historia, XXI, 1 70 11 10 6 3 0
Pliny, Letters, III, 1 54 18 11 10 2 5
Tacitus, Agricola, 1 62 12 14 3 3 6
Catullus, XLIV, 1 o 35 11 20 15 8 11
Lucretius, I, 149 68 15 9 4 3 1
Vergil, Aeneid, I, 1 62 18 16 3 0 1
Horace, Odes, I, 1 39 16 16 9 13 7
Ovid, Metamorphoses I, 1 61 17 11 4 2 5
Actual per cent of total 57.6 13.8 13.0 6.8 4.2 4.6
Predicted per cent of total. 58.0 12.7 14.4 7.2 4.1 3.6
Actual per cent in basic vocabulary (1st 3 groups combined) 84.4
Predicted per cent in basic vocabulary 85.1
THE COUNT OF INFLECTIONAL ENDINGS
18 "common" endings 66.5 per cent
Indeclinables (uninflected words) 23.4 per cent
22 "rare" endings 6.8 per cent
24 irregular pronoun forms 2.6 per cent
All other endings .7 per cent
100.0 per cent
The Constant Endings The Variable Endings
Per Cent Per Cent
1.8 am object 4.7 "Joker"* subject (also object if neuter)
2.6 em object 4.6 um subject object adjectives
3.8 e subject object adverb address
1.3 as objects
1.7 os objects 2.5 es subject subjects objects
3.0 us objects subjects objects adjective
1.6 ibus adverbs 8.9 a subject subjects objects adverb
4.8 o adverb
on verbs: I act 6.2 is subject adjective adverbs
2.6 ae subjects adjective adverb
6.8 t subject acts 5.2 i subjects adjective adverb
2.5 nt subjects act on verbs: pres. stem: to be acted upon**
1.9 re to act** perf. stem: I acted
Rules Governing Variable Endings
a If subject form is a, may be subject or adverb.
If subject form is not a, may be subjects or objects.
ae Quae may be subject, subjects, or objects.
Other forms in ae may be subjects, adjective or adverb.
e If subject form is e, it may be subject or object. If not, it is
adverb, except when set off by commas, to address someone.
es If subject form is es, may be subject, subjects or objects.
If subject form is not es, may be subjects or objects.
i If adjective form is not i, it is adverb.
If subject form is um, it is adjective.
Otherwise it may be-subjects or adjective (except qui and idem:
subject or subjects).
On verbs: present stem, to be acted upon;
perfect stem, I acted.
is May be subject or adjective as shown in dictionary forms.
If it is neither subject nor adjective, it is adverbs.
um If subject form is um, may be subject or object.
If adjective form is not i, it is adjectives.
Otherwise it is object.
us If adjective form is not us ' it is subject.
Otherwise it may be subject, subjects, objects, or adjective.
The Twenty-two Rare Endings
Per Cent (all less than 1 per cent) Per Cent
.29 arum adjectives .29 m I act
.11 erum adjectives .24 mus we act
.56 orum adjectives .53 s you (s.) act
.07 tis you (p.) act
.12 ebus adverb
.63 u adverb .10 r I* acted upon
.05 iter true adverb .09 mur we acted upon
.89 tur subject acted upon
.05 ei adjective adverb .51 ntur subjects acted upon
.65 d subject object
.57 ius subject object .05 isti you (s.) acted
adjective, true .20 erunt subjects acted
comparative adverb .39 ri to be acted upon
.21 isse to have acted
.17 verb stem with or without to: Act!
*On nouns or adjectives: subject (also object if neuter).
Penultimate Signs
Per
Cent Sign Example Meaning Idea
0.41 -a- laudat He praises On verbs whose second and
third forms end in -avi,
-atus; present action-
habeat (that) he have On all other verbs, present
action, subjunctive, except
that -am may also indicate
future action
1.36 -e- habet He has On verbs whose first form
ends in -eo, present action
laudet (that) he praise On verbs whose second and
third forms end in -avi,
-atus, present action-,
subjunctive
capiet He will take On all other verbs, future
action
0.18 -b- plus any vowel except a
dabunt They will give Future Action
1.00 -ba- legebat He was reading Action going on in the past
1.02 -re- legeret (If) he were Action going on in the past,
reading subjunctive
0.74 -ns- amans A loving Present Participle (a verb
(sub.) (husband) used as an adjective:
1.15 -nt- amantis (active voice) English ending -ing)
(adj.)
0.51 -nd- legendus To be read Necessity or obligation,
(passive voice) like "This book must be read."
0.51 -era- lexerat He had read Action completed in the past
0.17 -isse- lexisset He had read Action completed in the past,
(Distinguish from -isse as an subjunctive
ending)
0.27 -eri- lexerit (If) he read Past action, subjunctive
He will have (Rarely) Action completed
read in the future -
independent clauses
liber lectus est or lectus sit; the book was read
liber lectus erat or lectus esset; the book had been read
libro lecto, the book having been read, with the book
read, or when the book was read.
(The grammatical name for this is an "ablative absolute.")
liber amissus, the lost book
0.18 lecturus sum, I am just about to read, or I am on the
point of reading.
morituri salutamus, We who are about to die salute you.
Per
Cent Sign Example Meaning Idea
0.27 -ior- fortiores braver Comparative form of adjectives
0.10 -issim- fortissimi bravest Superlative form of adjectives
THE METHOD OF TEACHING
PREFACE
THE WORD COUNT
THE COUNT OF INFLECTIONAL ENDINGS
THE METHOD OF TEACHING
THE VOCABULARY OF LATIN LITERATURE
THE RECOMMENDED BASIC VOCABULARY
APPENDIXES
BIBLIOGRAPHY
VITA