which it should be compared. Naturally one thinks first of Oertel's Lectures on the Study of Language; but that is written for a different audience, and consequently on a different plan. For a parallel the author himself looks back to Whitney's Language and the Study of Language (1867); but the study of language, it is hardly necessary to say, has been revolutionized in the half century that intervenes. Midway, to be sure, lies Strong, Logeman and Wheeler, Introduction to the Study of the History of Language (1891), a book valuable in its day, but no longer adapted to modern needs. It deserves, however, to be mentioned here because of its relation to Paul's Prinzipien der Sprachgeschichte; which, though much closer, recalls that existing between Wundt's Völkerpsychologie, Volumes 1-2, Die Sprache, and the present work. The book thus aims at doing a service which is being done by no other work. The importance of this service cannot be overestimated. The teaching of various languages bulks large in the education given to the youth of the country. An important part of our intellectual life consists of the study of the cultural tradition of various peoples, and in this philological work are involved many questions of a linguistic nature. Both teachers and philologians need a knowledge of the principles of linguistic science, for it would seem axiomatic that no one can reflect profitably upon the phenomena of any language, unless he first knows what language itself is. Whitney. found that pupils who had enjoyed "the ordinary training in the classical or the modern languages or in both" were still capable of forming "views respecting the nature of language and the relation of languages of a wholly crude and fantastic character". Professor Bloomfield writes somewhat differently: While questions of a linguistic nature are everywhere a frequent subject of discussion, it is surprising how little even educated people are in touch with the scientific study of language. It is in Germany a subject of reproach dass der Philologe oft noch zu sehr an der altüberkommenen Betrachtungsweise hängt, die von einem mehr naiven als wissenschaftlichen Nachdenken über das Wesen der Sprache hervorgerufen wurde. The quotation is from the Preface of Brugmann's Griechische Grammatik2 (1889), which was considered worth reprinting in 1899 and 1913, in spite of a certain improvement recognized in his Kurze Vergleichende Grammatik (1904), pages V and 30. This improvement was not sufficient to keep Hirt, Handbuch der Griechischen Laut- und Formenlehre2 (1912), 57, from writing in the same strain, "die einfachsten Tatsachen sind unbekannt"; nor Kretschmer from speaking, Einleitung in die Altertumswissenschaft, 12 (1912), 463, of "eine bedauerliche Entfremdung zwischen der Sprachwissenschaft und der klassischen Philologie". Now I have no wish to enter upon the question whether there is a similar condition of affairs among our philologians; still less to seek to parallel Brugmann's citation of Meisterhans and Blass with American names. But I do wish to suggest that there are symptoms which should lead us to reflect upon this question. Why should the author of a Greek Grammar in 1915 feel put upon the defensive for "making use of the principle of Analogy"? That is a straw, but it may serve to show the direction in which the wind is blowing. More serious symptoms are the lack of books already noted; the fact that persons who have been taught from one to five languages are (to put it mildly) "surprisingly out of touch with the scientific study of language"; and that our classical philology is very largely inspired by German philology, which is itself infected with this neglect of linguistics. Our first need is to base our teaching of the classical languages squarely and fairly upon the principles of linguistic science. To form crude and fantastic ideas about the nature of language ought to be made impossible for any one who has studied Latin. That result cannot be attained without making it a great deal easier for the student to acquire that power of reading Latin which is the key to the enjoyment of its literature and the appreciation of the relation between ancient and modern civilisation. Prerequisite to this is a truer understanding of the nature of language on the part of our philologians and of our teachers themselves1. Under these circumstances it seems to the reviewer that one question alone is of prime importance: Can the work under review render the service it has undertaken to render? This question may be answered without hesitation in the affirmative. Among students of linguistics, there is a general consensus of opinion about the fundamental principles on which their work is based, and such agreement extends frequently even to matters of detail. Professor Bloomfield's book is limited avowedly to the presentation of this 'accepted doctrine', and the non-linguist may use it without fear of being misled upon such questions. To set forth this The necessity of a knowledge of the principles of linguistic science is not to be confused with the desirability of a study of comparative grammar. I should strongly advise any student who desires to fit himself to be a teacher of Latin or of Greek to include comparative grammar, in his preparation. I should not advise any one to 'adapt' Hirt's Handbuch for a text-book in our Schools. It seems worth while to say this because we have recently (American Journal of Philology 26.242 f.) been told that Hirt "modestly suggests that the gymnasia would do better by Greek, if they ceased to afflict students with a modicum of Xenophon and Homer and taught instead his handbook". This is a misrepresentation of Hirt's views upon a very important question. The problem confronting him at Leipzig may be restated in terms of American life: What shall we do with graduate students desirous of becoming teachers of Latin who come to the University ignorant of Greek? That problem is already not unheard of in America and there is danger that it may become acute. At present we seem to have three possibilities: (1) to treat the aspirations of such students as we do undesired kittens; (2) to allow these students to persist in their ignorance, and turn them out with Masters' degrees upon an unsuspecting public; (3) to put into their hands a Beginners' Book, written for School children, that they may get a modicum of Xenophon and Homer. Hirt believes that for such students a survey of the laws of the structure of the Greek language is better than this modicum of text, claims to have had the best results in so teaching them, and has made his Handbuch suited to their needs. In this he may be right or wrong-that is another question, but the problem is apt to become pressing, and we really should be able to devise some solution better than any of the three mentioned above. doctrine with sufficient wealth of illustration, in a form that is small in compass and yet such as may be read with ease and pleasure, was no light task; and the skill with which it has been accomplished is deserving of high praise. On the other hand, the very nature of this task renders it inevitable that another should find points at which he might wish for a different treatment. The broadest criticism I should offer is the wish that more space had been given to the processes of linguistic change. Room for this in part might have been gained by the exclusion of the phonetics of the second chapter, the subject being one that is usually handled separately. I must add, however, that the section is in itself most admirable, and that I should be loath to lose it. One idea that runs through the book is open to such serious objection as to require separate notice. At times reflective examination of a language may show that certain differences of sounds are distributed according to conditions which may readily be observed and stated. Our English vowels, for instance, are longer in final position and before voiced sounds than before unvoiced, longer in bid than in bit, in bee, bead than in beat. In such cases Professor Bloomfield speaks of "automatic sound-variation". Now this term suggests very strongly an idea, which Professor Bloomfield would no doubt disclaim, that such changes have no sufficient causes but just happen of themselves-automatically. But even worse than this is another suggestion, that each speaker continually makes these variations (according to the conditions involved) in each production of the sound; that we, for instance, start always with the short vowels and automatically leave them unchanged when we say bit, beat, but automatically lengthen them when we say bid, bee, or bead. Such a position hardly requires refutation. A phonetic change is a historical event or a series of such events occupying a definite portion of time; the final result is then transmitted by tradition, and it is a mistake to suppose that the process is being continually repeated. Such changes are due to complexes of causes that in their totality are unknown. Sometimes one (or more) of the elements of the complex can be ascertained; we then speak of 'conditioned' phonetic changes. Now, after the change is an accomplished fact, it is obviously a matter of indifference whether such 'conditions' are perpetual or not. The new sound goes on its own path, and what happens to it is another chapter in its history. It is surprising to find that Professor Bloomfield (221) maintains on the contrary that the process is being repeated automatically as long as these 'conditions' are undisturbed. The pre-Germanic spirant-voicing after unaccented vowel, for instance, left such automatic variations as *wása 'I was': *wēzumún 'we were' ; when, however, the stress was later shifted everywhere to the first syllable, the variation was of course no longer automatic, but purely traditional, as still in the modern forms, was: were. So, by a pre-English vowel assimilation *fotiz, the nominative plural . of *fot 'foot', became *fētiz, a variation whose automatism was destroyed by the phonetic change which dropped the second syllable of *fētis, giving Old English fet The examples really prove the contrary. Old English fet shows that at the time of the loss of the final syllable *fētiz was already established as the traditional pronunciation. Had the form still been *fōtiz varying automatically to *fētiz because of the following vowel, a form *föt should have resulted when that vowel was suppressed. The whole concept of sound-automatism appears to me, I confess, as the introduction of some mystic power for which there is no place in our explanation of language. Here also it is entirely needless. My opinion about the division of syllables and words differs also from that of Professor Bloomfield to some extent, as may be seen by a comparison of the American Journal of Philology, 33.403 f., 34. 157 f. On page 152 examples of Umlaut and Ablaut are given where we have been led to expect definitions. On page 154 it would be well to explain that in distinguishing between sound-variation and affixation there are two points of view, one historical, the other descriptive. From the former, the difference in vowel quantity between amās, amat (from *amāt) is sound-variation; from the latter, we may either view it in this fashion, or analyse am-ās, am-at. The historical point of view is not always possible, but the descriptive must not be confused with it. The same applies to affixation and infixation (155) and the Indo-European nasal present would illustrate the impossibility of coming to a decision on historical grounds. On pages 204 f. the discussion of 'phonetic law' might be improved, in what direction can be seen from Wundt's article in Philologische Studien 3.196 ff. Separate mention must be made of the last two chapters. Of these the first, The Teaching of Language, is written from the standpoint of modern languages; only mutatis mutandis can it be applied to the teaching of the Classics, but anyone who is teaching the latter can surely gain from reading and reflecting upon it. The second, The Study of Language, contains very sound and sane advice for the student who is planning to devote himself to the study of language. To the books mentioned in it I should like to add Otto Jespersen, Phonetische Grundfragen (Leipzig, 1904), and A Modern English Grammar on Historical Principles (Heidelberg, 1909-1914); P. Kretschmer, Sprache, in Einleitung in die Altertumswissenschaft herausgegeben von A. Gercke und E. Norden2, 463-564 (Leipzig, 1912: especially valuable for the classicist); and L. Sütterlin, Das Wesen der Sprachlichen Gebilde. Kritische Bemerkungen zu Wilhelm Wundts Sprachpsychologie (Heidelberg, 1902). Professor Bloomfield has put a valuable tool within the reach of teachers of language and philologians. It is to be hoped that the classicists will be among the first to make use of it. OHIO STATE UNIVERSITY. GEORGE MELVILLE BOLLING. Entered as second-class matter November 18, 1907, at the Post Office, New York, N. Y., under the Act of Congress of March 1, 1879 MAPS ILLUSTRATING ANCIENT HISTORY AND ANCIENT GEOGRAPHY Editorial The problem of maps one finds troublesome in classroom and private study both, particularly in private study. There are, to be sure, several very useful Atlases illustrating the geography of the ancient world. However, the maps in these are small, too small to be of service when hung on a wall. On the other hand, the Kiepert and the Rand-McNally maps, fine as they are for the class-room, are too expensive for the individual, and they take up more wall space than the ordinary private study, or even School or College office, in these days, can supply. Hence, I was much interested in an advertisement I chanced to see of a new series of maps, 44 x 32 inches, edited by Professors J. H. Breasted and C. F. Huth, Jr., of the University of Chicago (DenoyerGeppert Co., 460 East Ohio Street, Chicago). Inquiry showed that the maps could be obtained, in one form, at $1.40 each. The maps are all backed with durable muslin; the cheapest form, that at $1.40 per map, has a small wooden rod at top and bottom. From the elaborate Prospectus, sent at request by the publishers, it appears that the Series is to contain 16 maps in all. Since, on examination of five maps I have myself purchased, I find that the Series is one of great value, it seems worth while to give here, as briefly as possible, an account of the Series. Matter enclosed within quotation-marks is taken verbatim from the Prospectus. Special attention should be given to the scale of the various maps. B1-Ancient World, covering essentially the northern half of the eastern hemisphere. B2-Ancient Orient and Palestine (80 miles to the inch). B3-Egypt and Early Babylonia (32 miles to the inch). Inset maps show Crete on a large scale (12 miles to the inch), and give plans of Egyptian Thebes, Babylon, and Ninevah. B4-Oriental Empires (160 miles to the inch). This is to show four successive stages in the development of the Oriental World down to the Persian Empire. B5-Eastern Mediterranean. Two maps "illustrating the earliest age of Greek History" (56 miles to the inch). "The first depicts the situation in the Aegean Era, about 1500 B. C., the second shows the changes after the coming of the Greeks. It also gives the Greek and Phoenician Colonies in the Aegean". B6-Greek and Phoenician Colonization, showing the Mediterranean and its Hinterland, about 550 B. C. (80 miles to the inch). No. 22 B7-Boeotia and Attica (4 miles to the inch). "Insets provide plans of Athens and the Piraeus, of. Syracuse, of Alexandria and of Priene". B8-Athens (25 miles to the inch), "giving all essential localities and public structures. The latter are differentiated by color by the chief periods down to Hadrian". B9-Sequence Map of Greece, representing four stages of Greek development. "The periods covered are Greece and the Persians, Greece in the Age of Pericles, Greece at the time of Alexander, and Greece in the latter half of the third Century, B. C". BIO-Alexander's Empire, about 280 B. С. BII-Ancient Italy (32 miles to the inch). Two parallel maps. "The first shows the peoples and tribes of Italy, the second shows Rome's road system in Italy, the more important Roman and Latin colonies and the citizen and allied territories of Italy prior to the Social War". B12-Growth of Roman Power in Italy. Here Italy is again shown twice; in each case two stages in the growth of Roman power in Italy are represented. Thus the map as a whole gives a survey of the territorial development of Rome from about 500 B. C. to 265 В. С. B13-Rome (10 inches to the mile). The Republican and the Imperial City are differentiated in color. Insets give a view of Republican Rome and the Fora of the Emperors (20 inches to the mile). B14-Conquest of the Mediterranean. This map shows the expansion of Rome outside of Italy down to the time of Caesar. Four maps give the development during the first two wars with Carthage; three others show subsequent acquisitions of territory, in the East and the West, down to 44 B. C. The dates of the maps are B. C. 500, 304, 290, 265, 133, 63, and 44. The scale is 160 miles to the inch. B15-Caesar's Gaul (32 miles to the inch). This shows "the conquests of Caesar, and the relations of the Gallo-Roman frontier to the center of Roman power in the late Republic and the Early Empire. Caesar's campaigns are clearly shown. Insets give Plans of Roman Camps". B16-Roman Empire (80 miles to the inch). This gives the Empire at the death of Augustus. It "shows provinces and vassal states, the former differentiated into Imperial and Senatorial Provinces. Also all the important land and water trade routes are given. Two insets (scale 240 miles to the inch) represent the growth and shrinkage of Rome's territories from Augustus to the latter part of the third century, and the division of 337 into Prefectures and Dioceses the re in Empire, respectively". It remains to point out that the maps can be got in other forms, more convenient to handle, but also more expensive. Thus, they may be got on spring roller and board, with dust-proof cover, at $2.40 each; in steel spring roller case, they cost $3.00 each. The whole set, with the maps eyeletted, and put into a loose leaf chart head, so that a map may be removed and put back at will, and mounted on a special tripod stand, costs $24.00. In a drop front spring roller case, the set costs $31.00. For the earliest times, we may doubt whether an author could count on any financial gain. Literature was not so indispensable to the early Roman that he would have made any very great effort to get books. Furthermore, there was no regularly organized book A manual for teachers is to accompany and to explain trade, and it is hard to see how any very great profit the maps. The five maps I have myself bought I find satisfactory. These are B II (names on this map are given partly in the English forms, partly in the Latin forms); B 12 (names are given mostly in English forms); B 13; B 14; and B 15. All these maps are very helpful, and their cost puts them within reach of the individual. B 13 and B 14 are particularly satisfactory, at least to me. One important element in the success of these maps lies in the fact that they avoid excessive detail. Further, what is given is thrown out into sharp relief. One feature of them, however, I do not like, though the authors of the Series evidently pride themselves on itthe fact that names are given largely in English. Were the names all in English, the result might be endurable. But the putting, now of Latin, now of English names on the same map, and, worse yet, the putting of both names for the same thing (e. g. for such a minor stream as the Arnus, both Arnus and Arno are given), makes a rather revolting jumble. Even more absurd is the fact that, in the Map of Gaul, where, as a matter of fact, the names are given mostly in Latin, we have standing out in bold letters a French form, "Aquitaine". In one matter more the publishers might, without great expense, help the users of the maps, in their cheapest and so most serviceable form. At present, the identifying letter and number appear on these maps only inside, on the bottom margin. They could easily be added on the top margin, and could be set also beside the name of the map, inside. Finally, it would be easy and most helpful to put identifying letter, number and name on the back of the map, in position to show plainly when the map is rolled up. С. к. THE PROFITS OF LITERATURE IN ANCIENT ROME The fact that ancient Rome possessed a literature is proof enough that literature was profitable. Our task, then, is simply to find out what form the profits took. This is not easy to do: we have only scattered bits of evidence to piece together, and these are capable of varying and contrary interpretation. Most students of the question have not taken into account all the evidence, or have reached their conclusions first and sought support afterwards. I shall attempt to collect here most at least of the testimony of Rome to the financial relation of author and publisher. This paper was read at the Tenth Annual Meeting of The Classical Association of the Atlantic States, held at the Central High School, at Philadelphia, April 14, 1916. could have been derived from such hit or miss methods as were possible. The pirating of editions was always fatally easy, and, with no trade to control it, was in earlier days easier still. We must, however, recognize the existence of a book trade earlier than is commonly done. The usual statement, that there was none till the time of Sulla, overlooks a remark by Polybius (3.32). He says, in comparing his own universal history with the local or partial histories of others, that it is easier to buy his forty books, which are as it were in one piece. Polybius was of course thoroughly cosmopolitan, but it seems likely that he was thinking chiefly of Roman readers, who must then have had some means of getting books. In another passage (16.14) he mentions certain Greeks who wrote 'with no view to gain', implying that there were others (probably also Greeks) who did write with a view to gain. But it is not necessary from the context to believe that these men were writing for Roman readers, rather than for the markets of Athens or Alexandria, which, long before this, had developed book trades. For the early period, then, we can hardly assume that Roman writers could have made much money from their books. Livius Andronicus may have sold to his classes copies of his Odyssey, and dramatic poets may have sold their plays to the producers, but these would hardly be cases in point. The earliest publisher of whom we have definite knowledge is Atticus. Publishing was a side line with him, yet he seems to have had a retail store in Rome (Cicero was perhaps his chief customer), and connections in Athens and elsewhere in Greece. He competed successfully with his rivals for the publication of some, at least, of Cicero's works, and it is conceivable that he published Lucretius's poem, if the tradition be correct that this was edited by Cicero. He issued, among other works, the treatises on Cato by Brutus and Hirtius, and a series of valuable and excellent editions of Greek authors came from him2. Cicero had such confidence in his judgment that he left it to him to decide when to publish the Second Philippic (Ad Att. 15.13.1). The financial aspect of their relation has never been made clear. Birt was formerly of the opinion that Cicero received a percentage on the sales of his works, but now seems to have receded somewhat from this position. The chief evidence for this opinion is a letter in which Cicero says to Atticus, quoniam impensam fecimus in macrocolla (Ad Att. 13.25.3). But even the use of the first person plural would not prove that expenses were ordinarily shared and the F. W. Hall, Companion to Classical Texts, 51, 230; Th. Birt, Antike Buchwesen, 284 f. 3Buchwesen, 354; compare his discussion in his revision of Müller's Handbuch, 1. 316 ff. receipts divided pro rata. Macrocolla were large sheets used for some special purpose, possibly for a presentation copy. Therefore, we are not justified in inferring that this was the regular arrangement between Cicero and Atticus. Even if the author did help now and then to defray the preliminary expenses, it would not prove that he would receive more than reimbursement. Another passage often used to prove that Cicero made a profit from his works is the following: Ligarianam praeclare vendidisti. Posthac quicquid scripsero, tibi praeconium deferam (Ad Att. 13.12.2), but this is to be interpreted in the light of another letter (Ad Att. 13.19.2), written about a week later: Ligarianam, ut video, praeclare auctoritas tua commendavit. Vendo is common enough in the sense of 'recommend'; cf. e. g. Ad Att. 1.16.16,8.16.1; Horace Epp. 2.1.75. In Horace, Epp. 2.1.35, pretium is used in a similar figurative sense. The first edition of the Academica resulted in a loss for Atticus (Ad Att. 13.13.1), for which Cicero tried to console him by promising him that the second edition would be far superior in every way, but there is no suggestion in this consolation that Cicero shared the loss, nor is there any sign of financial concern on Cicero's part at all. So, if there was a percentage arrangement in this case, Cicero's profit must have been guaranteed. It is easier to believe that there was no such arrangement. In another case (Ad Att. 2.1.2), Cicero directs Atticus to have the De Consulatu Suo in Athens and other cities of Greece, videtur enim posse aliquid nostris rebus lucis adferre. There is no suggestion here of financial returns. We know of at least one case where Atticus allowed a work of Cicero to get out without the author's consent (Ad Att. 13.21 A.1), and the incident proves that in this case at least there was no very binding contract between them. A somewhat similar expérience is referred to in De Oratore 1.5 (compare 1.94) regarding the De Inventione. Cicero, therefore, gives us no reason to believe that authors were financially benefited by their labors, even if he gives no absolute proof that they were not. Let us turn to Horace. The antecedents and early life of the poet would not suggest that he could afford to devote himself to literature unless it promised some reward. He tells us that poverty drove him to write verses (Epp. 2.2.49 ff.), and there must have been some prospect of financial success to draw to Rome such a steady stream of aspiring authors. Yet in another passage (Serm. 1.4.71), written very soon after his coming to Rome after Philippi, Horace disclaims any intention of publishing his works broadcast. Single poems, recited privately to a small circle of friends, could hardly have been a source of great profit. Their chief financial value was as advertising. It is true that later Horace had an arrangement with the Sosii brothers to publish his works, and from this they were to get aera, while Horace was content with longum aevum (Ars Poetica 345-346; Epp. 1.20.2). The relations of Martial and his publishers are very complicated. We may safely say that, if any Roman author made money from literature, Martial was that author. He was not likely to overlook any source of income. Yet we do not know what financial basis his dealings with his publishers had, though Birt believes he received his profits in the form of a lump sum. He mentions in his poems four publishers (or booksellers) -Secundus, Pollius, Atrectus, and Trypho. I shall mention these men in the order in which they appear in the poems. In the second epigram of the collection we now have, Martial advises travellers to buy the small handy edition of the Epigrams, an edition on parchment, which Secundus has on sale at his shop behind the temple of Peace. We might infer that Secundus was the publisher of this volume in the series, especially if we could think of this poem as having appeared separately on the pila of the bookseller, like the rhyming advertisement and (identical) title page of Lowell's Fable for Critics, with which Putnam aptly compares Martial's poem. In the same volume (1.113), we find a reference to Pollius, the publisher of the poet's youthful works. In 1.117 Lupercus, who wishes to borrow a copy of Martial's poems, is told that Atrectus, whose shop is in the Argiletum, sells the book for five denarii, a price which Lupercus naturally thinks too high. Here we have, then, three possible publishers mentioned in one book. I shall consider more fully at another time the relations of the three and the problem of different editions of Book 1, a matter somewhat freely debated in recent years. In 4.72 a borrower is referred to the bookstore of Trypho, and in 13.3 the poet says that Trypho has the Xenia for sale at four sestertii, but can sell it for two and still make money. The interpretation of this last remark is doubtful. We know that fixed prices were as rare in ancient Italy as in modern, but we are hardly prepared for such a suggestion as this. I venture to add to the number of explanations. Trypho was a man of consequence to his influence we owe the publication of Quintilian's Institutio Oratoria and in my opinion a close friend of Martial. It is, therefore, good advertising that Martial is on jesting terms with the famous publisher, and the poet can make facetious remarks about his publisher's excessive profits without fear of being misunderstood at least by his contemporaries. Obviously none of the passages I have cited allows us to determine the financial relations existing between Martial and his publishers. A percentage basis is out of the question when we remember Martial's complaint that, even though his poems are known in distant Britain, nescit sacculus ista meus (11.3.6). I do not believe that we can assume even the payment of a lump sum as a regular thing. Martial himself has led many critics astray by the phrase praemium libellorum (10.74.7). Yet the passage proves nothing. The poet complains that, worn out by calls and social formalities, |