MASTER NEGA TIVE NO. 93-81401 MICROFILMED 1993 COLUMBIA UNIVERSITY LIBRARIES/NEW YORK as part of the "Foundations of Western Civilization Preservation Project" Funded by the NATIONAL ENDOWMENT FOR THE HUMANITIES Reproductions may not be made without permission from Columbia University Library COPYRIGHT STATEMENT The copyright law of the United States - Title 17, United States Code - concerns the making of photocopies or other reproductions of copyrighted material. Under certain conditions specified in the law, libraries and archives are authorized to furnish a photocopy or other reproduction. One of these specified conditions is that the photocopy or other reproduction is not to be "used for any purpose other than private study, scholarship, or fesearch." If a user makes a request for, or later uses, a photocopy or reproduction for purposes in excess of fair use," that user may be liable for copyright infringement. This institution reserves the right to refuse to accept a copy order if. In its judgement, fulfillment of the order would involve violation of the copyright law. A UTHOR: JONES, THOMAS MADISON TITLE: CASE CONSTRUCTIONS OFSIMILISANDITS... PLACE' BALTIMORE, MD. DATE: 1903 Restrictions on Use: COLUMBIA UNIVERSITY LIBRARIES PRESERVATION DEPARTMENT BIBLIOGRAPHIC MICROFORM TARGET Original Material as Filmed - Existing Bibliographic Record '1 * • i 'i 1 877 ,56 Z8 ▼A it :i Jones, Thomas Madison, 1660- Case constructions of slmlHs and Its compounds • •• by Thomas Madison Jones ... Baltimore, the Lord Baltimore press «1903> hS p. 2U cm. Thesis. (Ph. D J Johns Hopkins, Volume of theses. * ^w * Master Negative # iMiH TECHNICAL MICROFORM DATA REDUCTION RATIO:__iZ^. FILM SIZE: 3 S ^J^'' __ IMAGE PLACEMENT: lA "iR IB IIB DATE FILMED: __^:__13 INITIALS____^£1_ HLMEDBY: RESEARCH PUBLICATIONS. INC WQQDBRIDGE. CIV D Association for information and image iNanagement 1100 Wayne Avenue. Suite 1100 Silver Spring. Maryland 20910 301/587-8202 Centimeter I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 mm iiliiiiliiiil Inches 1 1.0 1^ 1 ^-3 1 2-^ 1^ JUm 12.9 I.I Urn *^ 140 •- w k.k.1. 1.4 1 ^-^ 1.8 1.6 1.25 II I I I I I I I II I 5 MnNUFfiCTURED TO fillM STPNDflRDS BY PPPLIED IMfiGE, INC. CASE CONSTRUCTIONS OF SIMILIS AND ITS COMPOUNDS A DISSERTATION SUBMITTED TO THE BOARD OF UNIVERSITY STUDIES OF THE JOHNS HOPKINS UNIVERSITY IN CONFORMITY WITH THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY BY THOMAS MADISON JONES Profesaor of Greek in Randolph- Macon College 1903 THS FRIBDENWALD COMPANY BALTIMORE, MD., U. S. A. '.AMi LITERATURE OF THE SUBJECT. The literature of the subject of the case constructions ofsimilis and its compounds is not extensive. Most of the grammars dis- miss the matter in a few words. Kiihner, 2-328, and Bennett, App. to Latin Grammar 323 (cf. remarks on page 139 of Teach- ing of Latin and Greek in the Secondary Schools, by Bennett and Bristol), have notes of some length. Haase, Note 550 to Reisig treats the subject with considerable fullness and in his Vorle- sungen iiber lateinische Sprachwissenschaft 2, 134-142 a still more extended discussion is found. Madvig, Cic. De Fin. 5-12, deals with the subject for Cicero mainly, and RitschI, Op. 2-570 (Rh. M. 7-583) and 579 (Rh. M. 8-159), discusses it for Plautus. Also in Op. 3-261 (cf. Suet. Reliqq. Reifferscheid 522) he treats it for the fragments of the early poets. References of less im- portance are given as occasion requires. TEXTS USED. In this investigation, while for some authors several editions were consulted, the texts mainly or exclusively used are the fol- lowing. For the fragments of the early poets, Ribbeck's Scaeni- cae Romanorum Poesis Fragmenta with Miiller's Q. Enni Carmi- num Reliquiae and Bahrens' Fragmenta Poetarum Romanorum. For Plautus, the Ritschl edition of Lowe, Gotz, and Scholl with constant reference to other recent texts, and for Terence, Umpf en- bach, Dziatzko, Fleckeisen, and Tyrrell. For Cornificius, Marx ; for Varro's Lingua Latina, Muller, and for his De Re Rustica, Keil. For Cicero, Miiller; for Lucretius, Brieger with Lach- mann and Munro. For Vergil, Ribbeck ; for Livy, Weissenborn ; for Lucan, Hosius ; for Silius Italicus, Bauer; for Martial, Gil- bert and Lindsay ; for Quintilian, Bonnell ; for Juvenal, Fried- lander ; for Tacitus, Halm ; for Suetonius, Roth ; for Lactan- tius, Brandt ; for Firmicus, KroU and Skutsch ; and for the Vul- gate, the edition of Turin, 1851. TABLE OF CONTENTS. Method Pursued 9 Rule OF Old Latin Grammarians — Discussion 10 Rule or Madvig — Discussion 12 Rule of Haase— Discussion 13 View that Similis with Genitive is a Substantive — Discussion. . 23 Table of Statistics 28 Remarks on Usage of Individual Authors. Dative in Plautus AND Terence 33 Results Yielded by Statistics. Origin, Growth, and Decline OF the Genitive with Similis 40 Construction Not Influenced by Position 44 Construction of Compounds of Similis ... 45 Rare Use of Other Constructions than Genitive and Dative . . 45 CASE CONSTRUCTIONS OF SIMILIS AND ITS COMPOUNDS. The familiar construction of similis and its compounds with both the genitive and the dative case is the subject with which this paper is concerned. One way of approaching this subject is through the fundamental meaning of the cases. Thus Weis- senborn, in accordance with the original signification of the cases, as accepted by him, conceived of the genitive as giving rise to the similarity, and of the dative as receiving the similarity from without.* This method of approach is avoided as both difficult and uncertain. Unsatisfactory likewise is the treatment which, as, for instance, that of Haase,' merges the double case construc- tion of similis with the same double case construction of such a group of adjectives as vicinus, socius, amicus, and others, so as to make the same reasoning apply to all of them. For, however much these adjectives may have in common, it is not at all ob- vious that the same explanation applies to the genitive with ami- cus and the genitive with similis. A better method of treatment, it would seem, is to observe carefully the case constructions of similis in a large range of Latin authors of various periods and departments, and, by the usage of these authors, not only to test the various distinctions which have been set up between the genitive and the dative, but also, if possible, to arrive at the real diflference between the two constructions. This, therefore, is the method pursued in the present investigation, a study which finds ample ground in the following brief outline of views on the question in hand. The authorities on the subject of the double case construction of similis fall into these classes : 1. Those who note the double construction without remark. 2. Those who note the double construction denying diflference in meaning but explaining variation, (a) as a matter of period. > See Haase, note 550 to Reisig. Cf. Haase, page 14 of this paper. Cf . also Gossrau, Lat. Sprachlehre, p. 810, and Kuhnast, Liv. Syn., p. 134. ' Vorlesungen iiber lateinische Sprachwissenschaf t, 2-135. Cf. p. 13 of this paper. 10 Case Coxstructions of Similis and its Compounds (b) as a matter of objects involved in the comparison. (c) as a matter of euphony.' 3. Those who note the double construction and affirm a differ- ence in meaning to the effect, (a) that the genitive denotes inner likeness, the dative outer likeness. (b) that sitnilis with the genitive denotes *image' {Eben- Uldy Ahhild), with the dative simple comparison. (c) that similis with the genitive is a substantive, with the dative an adjective. (d) that, the explanation being found in the theory of the cases,^ the genitive is (1) objective,* or (2) partitive* or (3) genitive of origin. ' The records of the discussion go back to Flavins Caper,® a gram- marian of the second century, who states his rule thus: Illius similis ad mores refertuvy ilU similis ad vultum, Diomedes, under uses of the genitive, writes, Sim,ilis sum tiii morihis\ and under uses of the dative, Similis sum iihifigura^^ Charisius" notes the double construction without explanation. Beda" in one place repeats the rule of Caper and elsewhere states the matter thus: Similis sum tui morihus, similis tihi facie}^ Other references to the construction in the early grammarians add nothing to the statement already given, which has found metrical expression in this form : Ille tui similis, mores qui servat eosdem ; Ille tihi similis, faciem qui servat eandemM This old view is not without adherents in later and even in recent years. Weissenborn,^* Keisig,** Zumpt, " Schmitz, Midden- 3W61flain, quoted by Kiibnast, Liv. Syn., note p. 125. 4 See note 1, p. 9 of this paper. 5 Haase, Vorl. 2-135, and Hime, Intr. to Lat. Lang., 2-234. •Reisig, Syntaxis, 566. ' Weissenborn, quoted by Haase, note 550 to Reisig. 8 Gram. Lat., Keil, 7-97. Cf. Drager, 1-445, and Drakenborch, Llv. 6-18-3. »Gram. Lat. K. 1-311. Wid. 1-313. i«id. 1-108. "id. 7-276. i»id. 7-288. "Haase, Vorl. 2-134. '5 Cited by Haase, note 550 to Reisig. "Vorl. iiber lat. Spraebwissenschaft § 376. The exact references to the following grammars need hardly be given. »' Cic. Verr. 3-160. •' Case Constructions of Similis and its Compounds 11 dorf-Griiter, and Feldman (1882) repeat it in their grammars, Ruddiman, Gruber, Heraeus (1885) give it qualified approval, and there is a distinct trace of it in no less an authoritv than Kiihnast, who says, "Of likeness perceived by the senses Livy seems to use the dative without exception,"^® though he holds elsewhere (p. 124) that between the genitive and the dative with similis no sharp lines can be drawn. Antoine, too, may be cited here, who says, whenever similis in Vergil is construed with the dative, ^Ule externa taiitum similitudine agitnr," " This traditional distinction was first denied by Vossius^'' about 1600, and he is cited with approval by Drakenborch." Others who deny it in their grammars are Otto Schulz, Meiring, Gossrau, Ferdinand Schulz, Drager, and of course the advocates of other views to be mentioned hereafter. How groundless this distinction of genitive of inner likeness, dative of outer likeness is, a short exhibit will conclusively show» In making it only instances that seemed certain were included, for sometimes it is not entirely clear whether internal or external likeness is under consideration, and sometimes both are involved. The estimates are for similis and its compounds in all degrees of comparison. The genitive of the personal pronoun and verum is excluded. Genitive. Dative. Varro, Cornificius, and Lucretius. Cicero. Inner likeness 3 Outer likeness 11 Ratio of cases of inner to outer ^£ likeness with genitive 308 Ratio of cases of inner to outer ^8^ I likeness vvitb dative 308 f Inner likeness 73 Outer likeness 29 Ratio of cases of inner to outer 803 "{ likeness with genitive 319 Ratio of cases of inner to outer 1131 likeness with dative 319 8 28 39 11 As the ratios show, in the first group the dative as compared with the genitive shows a slight preference for inner likeness, "Liv. Syn., p. 125. Llvj uses the genitive very little, but the genitive of outer likeness is found. See table p. 29. »»De Casuum Syntaxi Vergiliana, p. 138. This remark, however, is just as true of Vergil's use of the genitive. He uses it only once (Aen. 5-594), but this time it is used of outer likeness. For other references see Haase, note 550 to Reisig. »o Drager, 1-445. 21 Livy, 6-13-3. 12 Case Constructions of Similis and its Compounds Case Constructions of Similis and its Compounds 13 while in Cicero the preference of the dative for inner likeness is distinctly marked. Madvig's" treatment of the construction may now be taken up. Following Wesenberg,'* whom he credits with first having given the correct rnle, he not only rejects what he calls the worthless rule of the old grammarians about likeness in character and figure, but sets up a new distinction. Noting the varying usage at different periods, he observes that the older writers and Cicero in comparing living beings, especially men and gods, used the genitive with similis, the dative in such cases being so rare in Cicero as hardly to escape suspicion; that in comparing things, however, genitive and dative were indiscriminately used, with the exception that hoc simile illi, ei, super iori {mutro genere) seemed always to be used by Cicero; that after the time of Livy, the dative, almost exclusively used by the Augustan poets, became more and more prevalent in comparing persons. Koby " may be compared for a statement very similar to that of Madvig. Munro*^ and Mayor'® seem to be in accord with him, and Drager " quotes him with modified approval. As apparently following Madvig more or less closely in their grammars, some of them even making the genitive exclusive with persons, are Gill- hausen, Koziol, Meissner, Menge, Putsche-Schottmuller, Gold- bacher, and Deecke. Madvig^s remark that the periods of the language are import- ant in studying the construction of similis is well worth atten- tion. However, his observation on the prevalence of the genitive in the older writers in comparing living beings disregards the usage of Varro," not to mention constructions in Plautus and Terence to be noted hereafter. And his statement concerning the prevalence of the dative in later Latin, while in general accord with the facts, is scarcely strong enough. For Cicero he makes the correct statement that, while in comparing things the genitive and dative are interchangeable, the genitive is almost exclusive in comparing persons, though, as he implies, even in "Cic. De. Fin. 5-12. "Cf. Kuhner, Cic. Tusc. 1-15-34. «* Gram. 1317. » Lucr. 4-1211. " Cic. Nat. Deo. 2-149. «' 1-445. *s3ee table p. 29. Varro uses the genitive only once in comparing per- sons, tlie dative seven times. Madvig observes, however, (De. Fin., ed. of 1876) that he had not carefnlly noticed the usage of Varro. this special sphere of the genitive examples of the dative are found." Madvig, then, apart from rendering the important service of indicating the general range of the constructions with similiSy hardly does more than to give the ordinary usage of Cicero, and that from an apparently arbitrary point of view. Indeed Haase^ rejects Madvig's conclusions as unsatisfactory, and Seyffert" may be compared for a similar opinion. We reach now a view apparently foreshadowed by Ramshorn,'* but first clearly announced by Haase,^ who, rejecting, as has just been said, the conclusions of Madvig, sets up a new distinction. His main results, reached about the same time and independently by Seyffert," were followed by Kiihner '^ and accepted by Schmalz and Landgraf.** Likewise the grammars of Meiring, Menge, Berger, Harre, Deecke, Holzweissig, and Lane are in greater or less accord with the statements of Haase. Since his distinctions are so minute, they must be given with considerable fullness, especially as they are comparatively recent and are sustained by so much authority. Speaking of such adjectives as vicinus, propinquus, socius^fami- liar is f amicus, aequalis, par, and similis, Haase says : " " If these words are construed with the genitive, it is evident that two persons (or things) are considered as belonging together in a pair, and the adjective merely supplies the ground upon which the relation as a pair rests, or the way in which it arises; consequently, with the genitive it is not the intention to express the quality that one object has with reference to another, but the connection of the two which arises from this quality, i. e. their relation as a pair of which the quality is the condition .... On the contrary, with the dative the existence of the quality is really asserted as a fact not previously present to the mind, and it is afi&rmed that one subject has it with reference to another without drawing the conclusion that by this means both are joined in a pair for which the supposed quality constitutes the ground. If, for example, one lives in my neighborhood, and I have reason to assert this as 29Drager (1-445) says there are seven such cases in Cic. The table p. 29 shows nine. To these add De Or. 3-47, mihi .... simillimnm. *>Vorl. 2, 135. 31 Cic. Lael. 488. WLat. Gram. 2-320 (1830). »3Note 550 to Reisig and Vorl. 2, 134-142. "Cic. Lael., p. 488 (2 ed. 1876). 35 Gram. 2-328. »« Note to Reisig 3-621. 8' Vorl. 2-135. 14 Case Constructions of Similis and its Compounds something hitherto unknown, the statement is, vicimis mihi est ; the conclusion follows that the distance between us is not great, etc. On the contrary, if I wish to say that, on the ground of our living near each other, we are in the relationship of neighbors, acquaintances, and friends, the statement is, vichuis metis est . . . . Consequently, the dative expresses the perception of a quality, the genitive expresses the mutual relation arising from the quality." " The same is true for similis and par. If I wish merely to designate relativity and to say that on account of similarity two objects belong together, form a pair, the one being a copy of the other, the genitive is used; on the contrary, if I just at the moment perceive the similarity, and that, too, not in such a way that the two things appear united in a pair, but for the present only in such a way that I can assert the actual similarity, the approximate likeness, the dative is used." The pair conception expressed by the genitive was the original one, Haase thinks; afterward, with a different meaning, appeared the dative, which later extended its sphere. But if the question is under what circumstances, after the difference between the genitive and the dative developed, each case may stand, " it is clear," says Haase, " that with the genitive the similarity must be an evident, actual one which binds the two subjects together in the relation of original and copy (^Urbildund Ahbihl); with the dative this is not necessary; here the similarity may be partial, limited to a single point, conditional, problematic, approximate. At any rate it is a similarity which just at the moment is asserted or perceived as a fact, and from which the conclusion is not drawn that the two similar objects are related as a pair; consequently, it is a more general and comprehensive method of expression." The fundamental ground of Haase's rule with some examples of its application are given in another passage, which it may be well to quote.^ " The explanation founded on the nature of the cases and the only correct one seems to me to be that one which I have already applied to propriusJ"^ To begin with, if the geni- tive is thought of alone, for example, h(ymo est dei, to a correct feeling for language it is undeniable that a real belonging, an actual dependence, of one object upon the other is expressed, WNote to Reisig 3-617. »The remarks on propHua arc found In note 529 to Reisig. r Case Constructions of Similis and its Compounds 15 which, if not more definitely determined, can, most naturally, at least in this case, convey the idea of possession, while, in other relations of the objects, other ideas can be conveyed. If now that belonging and dependence is more definitely determined by the idea of similarity,'" from the two taken together no other relation can arise than that of copy to original, or vice versa ; one object, as it were, is repeated in another, and the comparison of the two rests upon inner necessity. Ho7no est dei similis consequently means *man is the image of God.' " The kind of objects compared when the relation of model and copy may be aflBrmed, Haase defines as follows:" "If it is a question as regards the kind of similarity, it is understood, as a matter of course, that the copy must really be of the same kind as the original. This similarity is more likely to be that of character. But it is not at all necessary that the similarity should be incorporeal ; one object of sense may be the image of another, if one repeats the nature of the other. Objects of dif- ferent kinds cannot be compared in this way, for in this case one is not repeated in the other, but only in this or that respect may similarity be affirmed." Still when the language is figurative the model and copy relation may, he says, be affirmed of objects that are unlike, as in Cicero dialectics is called the likeness of the fist." As regards the extent of the likeness when the genitive is used, Haase notes that it should exist not only in certain specified re- pects, but should characterize throughout the objects compared. Still in comparison of character the model and copy relation may be affirmed of partial likeness.*" For example, lascivia socor- diaque gladiatorum magis quam ducum similes.**' To summarize : According to this theory the genitive is used mostly of objects of the same kind, but may be used of objects of different kinds; it is used mostly of inner likeness, but may be used of outer likeness ; it is used mostly of complete likeness, but may be used of partial likeness. Moreover, the dative is arbitrary ** and at the will of the author invades the sphere of *o««The adjective merely supplies the ground," etc. See first quotation from Haase, p. 13 of this paper. ♦» Note to Reisig 3-618. "Cic. De Fin. 2-17. The passage \fill be cited in the discussion of Haase's theory. *' Note to Reisig 3-619. *< Tac. Hist. 3-76. ^^Note to Reisig 3-619. 16 Case Coxstkuctioxs of Similis and its Compounds the genitive except in a very limited range.** Still further, the distinction thus limited, holds mainly for Cicero only, for in earlier Latin the genitive prevails and in later Latin the dative is dominant/' With all these limitations the question naturally arises whether the distinction is worth making. As if in anticipation of such a question, Haase concedes that the distinction he draws is not so readily understood as that of the old grammarians or of Madvig. But in its favor he mentions the fact that for certain phenomena it offers a ready explanation. For instance, (1). It explains the inf requency of the genitive of outer like- ness, since here the similarity must rarely be of such a character as to justify the conception of a pair. On the contrary, since inner likeness is not apparent to the eye, but depends upon the judgment, it is much easier to consider two persons as forming a pair, and this, in connection with the fact that inner likenesses are much more frequently spoken of in literature,*" accounts for the more frequent use of the genitive in such comparisons. The facts in Cicero are as follows : (1). There is no infre- quency, but a marked frequency, of the genitive, as compared with the dative, in the expression of outer likeness. (2). The genitive, as compared with the dative, is relatively more frequent in the expression not of inner, but of outer, likeness. Excluding the genitive of the personal pronoun and verum and eliminating all doubtful cases, the figures are: *' Genitive Dative. . Onter Inner likeness, likeness. 29 78 11 89 1181 Ratio of gen. to dat. in expression of outer likeness -^29 808 Ratio of gen. to dat. in expression of inner likeness -^^ Haase's theory, then, explains a fact that does not exist. (2). It explains the combination veri simile, since the pro- bable is the copy of the true. This hardly explains. For in say- ing that the probable is the copy of the true Haase seems to com- 4« Vorl. 2-137. The limitation is that the genitive of the personal pro- noun is said always to be used. (Note to Reisig 3-619. Cf. Charisius K. 1- 108.) But even here the dative is found. Cic. De Or. 3-47. Cf. p. 41. 41 Haase, Vorl. 2-136. *« Haase, Vorl. 2-188. *»Cf. table on p. 11 of this paper. f Case Constructions of Similis and its Compounds 17 pare the combination veri simile with verum, and to leave out of consideration veri in veri simile, which is the only thing he is seeking to explain. Besides the model and copy relation of Haase is scarcely in harmony with the distinction drawn between verum and veri simile, Cic. De In v. 1-82 : si res aut vera aut veri similis permittet. Acad. 2-66 : Qui enim possum non cupere verum invenire, quum gaudeam, si simile veri quid invenerim ? Nor is his explanation in better accord with Acad. 2-49 : Si tale visum ob- jectum est a deo dormienti, ut probabile sit, cur non etiam ut valde veri simile? cur deinde non ut difficiliter a vero internos- catur, deinde ut ne internoscatur quidem? postremo ut nihil in- ter hoc et illud intersit ? Might it not also be asked why, on Haase's grounds, the comparative and superlative do not show in Cicero more attachment for veri than the positive ? (3). It explains the constant ^ use of the genitive of the per- sonal pronoun with similis, for each one is his own image.^^ Each one is his own image, to be sure, but the genitive of the personal pronoun occurs when one person '^^ or even thing ^ is compared with another. This, then, is Haase's theory, with its grounds, range of appli- cation, and claims, as stated by himself. Before proceeding to compare his theory with the facts of the language, three general observations may be made with reference to it. (1) It would seem strange that in the early period the Latins in u%\xig similis should have had only the conception of two objects as a pair,^ that only in a later period they should have developed the idea of general similarity and found an expression for it in similis and the dative, and that in still later times they should have allowed the construction with the dative to obliterate the very distinction it was intended to preserve. However that may be, it is certain that, without the device of the double case con- struction, they had the means at hand, in the degrees of com- parison, for the adequate expression of both these ideas, and it is worthy of remark that Varro^ testifies that the exact conception which Haase contends is expressed by similis and the genitive is given by similis in the superlative. " Itaque qui plura habent eadem, dicuntur similiores: qui proxume accedunt ad id, ut omnia habeant eadem, vocantur gemini, simillimi." w Not absolutely constant even for Cicero. Cf . De Or. 3-47. »> Note to Reisig 3-620. " Cic. Lael. 82. ML. L. 10-4. »»Cic. Tusc. 1-43. M Haase, Vorl. 2-136. 18 Case Constructions of Similis and its Compounds (2) Varro may be supposed to be a competent witness as to the force and meaning of the word similis, and in L. L. 10-3, 4, a passage in which genitive and dative alternate in a significant way, in striking contrast with Haase's fine distinction, he gives just such force and meaning to similis as one would naturally ascribe to it. " Simile est, quod res plerasque habere videtur eas- dem, quas illud quoius quid simile. Dissimile est, quod videtur esse contrarium huius .... Sic dicitur similis homo homini, equos equo, et dissimilis homo equo . . . . Eo porro similiores sunt, qui facie quoque paene eadem, habitu corporis et filo. Itaque qui plura habent eadem, dicuntur similiores: qui proxumeacce- dunt ad id, ut omnia habeant eadem, vocantur gemini, simillimi." In this quotation from Varro attention is specially called to three things, (a) He uses the genitive with similis where the likeness is slight, (b) he changes from genitive to dative though the thought allows no difference in meaning, (c) as is most worthy of note, he uses the genitive in the general comparison (ein Satz allgemeiner Gultigkeit), which is just what Haase says should not be done,** and the dative in the particular ones. (3) The distinction is highly subjective. All that it is possi- ble to say in the great majority of cases is, that here the concep- tion of model and copy might have been in the author's mind, not that it certainly was there. This subjective character of the distinction finds illustration in the confusing use of terms by those who adhere to it. For instance, with the dative, accord- ing to Haase (2-137), the method of expression is more general and comprehensive (die Ausdrucksweise ist allgemeiner und umfassender), while, according to Seyffert (Cic. Lael. p. 488), with the genitive one object is similar to the other in a general and comprehensive way (Was alicuiue simile ist, ist dieses in allgemeiner und umfassender Bezeichnung), whereas, according to Krebs s. ^milis, Seyffert means that with the genitive the similarity itself is general and comprehensive. {Similis mit Genitiv steht da, wo die Aehnlichkeit eine allgemeine und umfas- sende ist)." This is less clear than sunlight. We come now to compare Haase's theory with the facts of the language. 5«Note to Reisig 3-620, where Haase quotes Cic. Nat. Deo. 1-90, a paa- sage in which there is again interchange of cases, but here the dative is used in the general comparison, the genitive in the particular one. 37 Cf. Gildersleeve, Lat. Gram. 359, note 4. t Case Constructions of Similis and its Compounds 19 (I). In doing so we first present instances of simple interchange between genitive and dative without apparent reason for the variation. A good example is found in Lucr. 4-1208. Et commiscendo quom semine forte virili femina vim vicit subita vi corripuitque tum similis matrum materno semine fiunt ut patribus patrio. A similar shift is found in Cic. Nat. Deo. 2-149. Itaque plec- tri similem linguam nostri solent dicere, chordarum dentep, nares cornibus iis, qui ad nervos resonant in cantibus. In this case Kiihner** and Haase^ describe the change as arbitrary (willkiir- lich), but Haase" regards the last comparison as being less simple and clear, as the relative clause shows, and accounts for the geni- tive in this way. This is hardly an adequate explanation, as Cic. De Or. 2-265 shows, (dicebat) nostros homines similes esse Syrorum venalium: ut quisque optime Graece sciret, ita esse nequissimum. For here the necessary explanatory clause is joined to the genitive." On the supposition that there is a dif- ference in conception between the genitive and the dative the shift in case in these passages is not justified. Cf. Varro as above. (II). While an arbitrary shift might have taken place in the preceding passages, some examples may now be cited in which the argument requires that the meaning should not shift, and yet both genitive and dative are used. Cic. Nat. Deo. 1-90 : Nee vero intelligo, cur maluerit Epicurus deos hominum similes dicere quam homines deorum. Quaeres quid intersit. Si enim hoc illi simile sit, esse illud huic. Here Haase ^^ justifies the dative in the last sentence on the ground that only by its use does the statement secure universal application, and Kiihner" finds that in this place the genitive is used of likeness in a more definite way, while the dative indicates likeness in quite a general manner. With this example and the explanations given of the variations by Haase and Kiihner it is interesting to compare the following statement in Varro, L. L. 10-4: Minimum ex duobus constat omne simile, item dissimile, quod nihil potest esse simile, quin alicuias sit simile, item nihil dicitur dissimile, quin adda- tur, quoius sit dissimile. Sic dicitur similis homo homini. For here the genitive is used in the more general statements and the » 2-828. wVorl. 2-189. « Note to Reisig 8-620. MNote to Reisig 3-619. •> For a similar example see Cic. De Off. 1-89. «3 2-328. 20 Case Constructions of Similis and its Compounds dative in the particular one, which is exactly the reverse of the usage for which Haase found an easy explanation in the quotation from Cicero.** Another example is found in Cicero, De Leg. 1-29 : Nihil est enim unum uni tam simile, tarn par quam omnes inter nos- met ipsos sumus. Quod si depravatio consuetudinum, si opinio- num vanitas non imbecillitatem animorum torqueret et iiecteret quocumque coepisset, sui nemo ipse tam similis esset quam omnes sunt omnium. It is to be noted, too, that what Cicero here speaks of as the highest likeness is expressed by the dative. Cic Tusc. 3-23 may also be cited. Aegris enim corporibus simillima animi est aegritudo; at non similis aegrotationis est libido, non immoderata laetitia, quae est voluptas animi elata et gestiens. Ipse etiam metus non est morbi admodum similis. In Cic. Tusc. 1-92 speaking of death it is said : Quam qui leviorem faciunt, somni simillimam volunt esse. But id. 1-97 the language is : Quam ob rem, sive sensus extinguitur morsque ei somno similis est, qui non numquam etiam sine visis somniorum placatissimam quietem aflfert, di boni, quid lucri est emori. The superlative, of course, does not account for the genitive in the first sentence, for cf. Cic. Verr. 2-2-99 : Itaque fecit, ut exitus principio simillimus reperiretur.** Under this head of variation in case where the argument re- quires that there be no difference in meaning one other example may be given, Cic. Acad. 2-50 : Quo modo autem sumis ut, si quid cui simile esse possit, sequatur ut etiam difficiliter inter- nosci possit? deinde, ut ne internosci quidem? postremo, ut eadem sint ? . . . Et quidem honestis similia sunt quaedam non honesta et bonis non bona et artificiosis minime artificiosa. Here the dative is found throughout, but in Acad. 2-54, where exactly the same subject is under discussion, the genitive occurs. Sed si satis est ad tollendamcognitionem similia esse multamultorum, cur eo non estis contenti, praesertim concedentibus nobis ?«» (III). Passing now from the cases in which the genitive and dative interchange in an arbitrary way and from those that allow no shift in meaning but admit at the same time the shift in con- Btruction, we take up those examples of similis with the genitive •* Cf. note on page 18 of this paper. •sCf. also Cic. Dc. Or. 8-47, mihl . . . simillimum. ••Cf. Cic. Nat. Deo. 2-41. The expression varies, the thought does not. 1 Case Constructions of Similis and its Compounds 21 in which the EhenUld idea of Haase is (1) denied and (2) those in which it is practically impossible. (1) Few examples of the first kind could be expected, but Cic. Or. 220 seems to present one: Multum interest utrum numerosa sit, id est, similis numerorum, an plane e numeris con- stetoratio; alterum si fit, intolerabile vitium est, alterum nisi fit, dissipata et inculta et fluens est oratio. To say nothing of the meaning of numerosa, which similis numerorum explains, and of the sharp contrast between similis numerorum and plane e numeris constet oratio, it is certain that Cicero did not mean to commend a style of oratory so nearly resembling meter as to make it necessary that the two should be conceived of as model and copy. Another example of the same kind, though, perhaps, not quite so clear, is found in Cic. Tusc. 1-81 : Quaererem ex eo, cuius Buorum similis fuisset Africani fratris nepos, facie vel patris, vita omnium perditorum ita similis, ut esset facile deterrimus. To disregard the difficulty involved in being the Ebenhild of all bad men, since they must have been numerous and of various kinds, a difference is given in deterrimus which excludes the strict Ebenhild idea. (2) Some passages may now be cited in which the EhenUld idea is practically impossible. (a) The comparison is between persons and things, as in Cic. De Off. 1-89 : Ilia vero omnibus in rebus repudianda est optan- dumque, ut ii, qui praesunt rei publicae, legum similes sint, quae ad puniendum non iracundia, sed aequitate ducuntur. Plautus shows several examples. (b) Unlike things are compared, as in De Fin. 4-28 : Cuius- cumque enim modi animal constitueris, necesse est, etiamsi id sine corpore sit, ut fingimus tamen esse in animo quaedam simi- lia eorum, quae sunt in corpore. Nat. Deo. 2-29 : omnem enim naturam necesse est . . . habere aliquem in se principatum, ut in homine mentem, in belua quiddam simile mentis. Cicero can hardly intend to affirm that there is in the brute the EhenUld of the human intellect. A similar example is found in De Fin. 5- 38: Sunt autem bestiae quaedam, in quibus inest aliquid simile virtutis.*' (c) Here too seem to belong such indefinite expressions as portenti, monstri, ostenti, prodigii simile, never the dative in early Latin or in Cicero. Considering the necessarily indefinite nature •'Cf. Cic. Ad. Fam. 9-16-8. 22 Case Constructions of Similis and its Compounds of the monstrum, etc. and the range of subjects, trifling or serious, of which monstri simile, etc. is said, the conditions are not met for the Ehenbild conception of Haase. Cf. pro monstro . . . est, PL Asin. 289, and portento similis, Livy 10-47-6. In comparisons of unlike persons and things Haase says he finds the dative to be regularly used though the genitive is found when the comparison is a figurative one. The explanation of the genitive in the examples given above (a and b) in which unlike things are compared is not found in the figurative lan- guage, for to find here figures available as explanations is to press too far the mere presence of a word of likeness. Moreover, the pas- sage in which Haase explains the genitive by the figure, when compared with a similar place, may serve to show that the expla- nation is hardly adequate. The genitive that he explains by the figure is found in De Fin. 2-17: Rhetoricam palmae, dialecticam pugni similem esse dicebat. The passage to be compared with this occurs in Or. 113: Cum compresserat digitos pugnumque fecerat, dialecticam aiebat eiusmodi esse; cum autem diduxerat et manum dilataverat, palmae illius similem eloquentiam esse dicebat. Unlike things are compared here, but the parallelism, dialecticam pugni similem esse and dialecticam eiusmodi esse re- duces the figure to a point so low that it can scarcely be evoked as an explanation. (d) One object is compared with more than one. Haase recog- nizes the difficulty involved in this kind of comparison from his own point of view, when he says that veri simile is natural, but veris similia, as sometimes found, is readily explained, because in cases of undetermined plurality the idea of a pair is not to be ex- pected. But in other cases as well, such as the following, the idea of a pair is hardly to be found. De Fin. 4-32 : Nemo enim est, qui aliter dixerit, quin omnium naturarum simile esset id, ad quod omnia referrentur. Nat. Deo. 2-36: Neque enim, si stirpium similis (natura) sit aut etiam bestiarum, optima putanda sit potius quam deterrima. Tusc. 2-36: Illi autem voluerunt nihil horum simile esse apnd Lacaenas virgines qnibus magis palaestra, Eurotas, sol, pulvis, labor, militia studio est, quam fertilitas barbara. (IV). Having shown how the dative interchanges with the genitive not only where a slight change in meaning would be of no great consequence, but also where the argument does not T Case Constructions of Similis and its Compounds 23 admit difference of meaning, and having noted that the genitive is found where the Ebenhild idea is either denied or practically impossible, attention is called to an example of the dative where, according to Haase, the genitive certainly ought to stand. The pair exists and the likeness is so striking «« that the plot of one of the plays of Plautus rests upon it. Not even the wife with the keen eye of suspicion can distinguish her Menaechmus from his brother. The sentence is taken from Varro L. L. 10-38: Nam ut in geminis quom similem dicimus esse Menaechmum Menaechmo, de uno dicimus. In concluding one can hardly forbear introducing a quotation from Quintilian, 5-11-30: Scio quosdam inani diligentia per minutissimas ista partes secuisse, et esse aliquid minus simile, ut simia homini et marmora def ormata prima manu, aliquid plus, ut illud, Non ovum tam simile ovo. Quintilian, to be sure, was not thinking of modern grammarians, but he evidently knew nothing of the distinction Haase makes. If he had, he would have used the dative in the first place and the genitive in the second, for in that the highest degree of like- ness is expressed. Haase's theory, then, discounted largely in practical value by the restrictions he places upon it, open to objection in view of a character highly subjective and a range of ap- plication limited mainly to Cicero, and out of accord, further- more, with the general facts of the language, cannot be said to be satisfactory, especially since the phenomena it claims to explain either vanish upon examination or are not adequately explained. Hence the fact that it is not accepted by such authorities as Drager** and Krebs'« was to be expected. So Bennett declares: " In" point of meaning absolutely no distinction between the two (cases) can be discovered." " In reading Haase's discussion of the subject it is readily noticed how carefully he refrains from calling similis with the genitive a substantive. Indeed, in one place," he apparently rejects this view, as Madvig" and Wilkins'* do by implication, and elsewhere'* he dismisses the suggestion as useless and calls similis so used an adjective, as does also Landgraf. It is equally easy, however, to notice that, while Haase and others who follow him avoid calling wCf. Men. 1088-1090. 69i«445. " The teaching of Latin and Greek, p. 139. "Clc. De Fin. 5-12. ''«Cic. De Or. 3-47. '0 Antibarbarns 2-525. 78 Note to Reisig 3-618. 75 Vorl. 2-134. 24 Case Constructions of Similis and its Compounds similis with the genitive a substantive, they translate it as such without reserve.'® Others still not only use the substantive translation, Bild, AUild, Ehenbild, image, counterparty like, likenessy but state with more or less caution that similis with the genitive is a substantive. For instance, Kuhner," who evidently has Haase's distinction in mind, speaking with some reserve, says of adjectives of similarity and dissimilarity and their opposites; " The genitive is used especially when the adjective idea approaches the substantive idea." But later he remarks without qualifica- tion : '' ''Similia with the genitive is used by Cicero and the older authors, when it assumes the substantive idea of copy, image:' This brings us to the view, which, without regard to Haase's fine distinction, accounts for the two constructions with similis by assigning it to a group of words which are used both as sub- stantives with the genitive and as adjectives with the dative. Thus Drager ^ says that many adjectives become substantives and then are used with the genitive or with a personal pronoun. Such adjectives denote friendship, relationship, association, and their opposites. The words aequalis, affiniSy amicusy cognatusy contrarms, and joar are members of the group, and among them similis is thus placed by Gildersleeve: "^ " similiSy like (* we ne'er shall look upon his like again')." This view of the double con- struction must, therefore, be considered. The well-known substantivizing of adjectives needs no dis- cussion here, the double use and corresponding construction of the words of the group just referred to are not a matter of dis- pute, nor need it be denied that similis with the genitive is some- times a substantive. The question at issue is whether or not similis with the genitive is always a substantive. The considera- tion of this question may begin with the citation of some of the better examples of similis as a substantive without a case. The instances are far less common than might be supposed. A good example is found in Cic. Verr. 2-3-155 : Volo, mi frater, frater- culo tuo credas. Consorti quidem in lucris atque (in) furtis, gemi- no et simillimo nequitia, improbitate, audacia. The same use is found in Cic. Verr. 2-3-162: Quid isto fore festivius arbitramur, si est tuus natura filius, consuetudine discipulus, voluntate simi- lis. A notable example occurs in Juv. 2-6 : Si quis Aristotelen similem vel Pittacon emit. The neuter plural of similis without 7« Vorl. 2-135. Cf. Grammars of Holzweistig, Menge, Bergcr, and Lane. "2-327. "Jw 2-328. 79 1-444. » 359, Rem. 1. 11 i Case Constructions of Similis and its Compounds 25 a case is not so common as might have been expected. As an example Cic. Nat. Deo. 1-105 may be cited : ®^ Neque deficiat umquam ex infinitis corporibus similium accessio. There are, however, many examples in Quintilian. The neuter singular without a case has not been observed as a substantive except in the well-known meaning, likeness, simile, comparison, parallel case. A good example is found in Plautus, Amph. 446 : Nil hoc similist similius. Another of many similar examples occurs in Cicero, De Fin. 3-46 : Utunturque simili.®' This neuter form, petrified as a substantive with a distinct meaning, would seem to indicate a strongly marked tendency in the adjective similis to become a substantive. We pass now to some instances of similis with the genitive used as a substantive. Beginning with the substantive similis used with the genitive of the personal pronoun, Plautus, Persa 698 furnishes the earliest example noted : Videor vidisse hie forma persimilem tuL Other examples are, Cic. Ad. Att. 8-9-2 : Quanto autem fero- cius ille causae suae confidet, cum vos, cum vestri similes . . . gratulantes viderit ? Ad. Fam. 14-7-2 shows this construction with a preposition : £t tamen eiusmodi spero negotia esse, ut et vos istic commodissime sperem esse et me aliquando cum simili- bus nostri rem publicam defensuros: Cf. Ad. Att. 1-16-3: Pauci tamen boni inerant . . . qui maesti inter sui dissimiles et mae- rentes sedebant. Cf. also Cic. Phil. 10-3 : Cur semper tui dis- similes defendis?® Some examples of the substantive similis with the genitive of other pronouns referring to persons may also be given, and here again Plautus heads the list. Most. 128 : Nituntur, nt alii aibi esse illornm similis expetant. Other instances are Cic. Cluent. 158 : Sed hoc polliceor omni- bus . . . me . . . vel his judicibus vel horumsimilibusfacillime probaturum. De Fin. 4-49 : Quis igitur tibi istud dabit praeter Pyrrhonem, Aristonem eorumve similes ? Passing to the genitive of pronouns not referring to persons with the substantive similiSy Cic. Ad. Fam. 2-16-2 shows the relative : Nosti enim non modo stomachi mei, cuius tu similem quondam habebas, sed etiam oculorum . . . fastidium. And Cic. De Or. 3-208 shows the substantive similis in the neuter plural 8»Cf. Verr. 2-^-68. ^Cf. De Fin. 3-54. "Cf. Ad. Fam. 7-1-4, Ad. Att. 9-11-4, Verr. 2-3-148, Pliil. 3-18. f ^ 26 Case Constructions of Simius and its Compounds with the genitive of a pronoun : Haec enim sunt fere atque horum similia, vel plura etiam esse posaant, quae . . . orationem . . . ''^Tre lSt*of examples may close with one in which the sub- stantiye simiKs is used with the genitive of a noun Ojc- Q°;« • Rose. 55: Siniillima enim et maxime gemma societas hereditatis est Of. Verr. 2-3-162. , ^^ , The foregoing, as already said, are some of the better examples of the substontive use of simUis both without a case and with the genitive. The words « better examples ' are used advisedly, for in a matter so subjective what seems a very clear substantive use to one may appear to another in a very diflferent H^-f^^ as a matter of fact the dative is found with tmilts in examples in which its substantive use seems otherwise as distinct as in the cases just cited for its substantive use with the genitive. Com- pare e. g. Cic. Tusc. 5-97: Atque his similia ad victum etiam transferuntur.* , _,. * v,^ Most of the examples are naturally taken from Cicero, for he nsed the substantive siinilis with the genitive much more than others. Even in Cicero, however, the examples are by no means 80 numerous as might be expected. Exclusive of ven swnU which is omitted because of its fixed character, simihs with the genitive occurs about 240 times in Cicero. Of these 240 cases 29 have been observed in which the substantive use of nmilis seems clear «• In these twenty-nine examples of the genitive the per- sonal pronoun occurs twenty times, and of these twenty occur- rences nine are in the orations and six in the letters. In Plautus only the two examples quoted above of the substantive stmthi with the genitive were found and in Terence none. While this small number of cases of the substantive stmihs might be increased indefinitely by another investigator, since the question is largely a subjective one, there are cases in which it is practically impossible that similis with the genitive is a substan- tive. Such are cases in which simUis with the genitive is modi- fied by an adverb : MCf. Cic. Tnsc. 1-22, Lael. 50, Phil. 3-22. M In Juv. 5-132 and elsewhere similU with the dative looks mnch like a "'.ror3-208, Brut. 249. Q. Rose. 55, Verr. «;3;-^^3' ^^^^^^^^^^ Caec. 102, 103, Clnent. 158. Leg. Agr. 2-77, 2-97, Flac. 104 Phil. 2-2,3-18, 10-3, 13.48, Ad Fam. 5-8-3, 7-1-4, 14-7-2, Ad Att., 1-16-3 8 9_2 9-11-4, Acad. 2.91, De Fin. 4-49, Tnsc. 1-22, 1-43, Nat. D. 2.81, 3-23, Lael. 50. Case Constructions of Similis and its Compounds 27 PL Amph. 442 : Nimis similest mei. PL Mil. 519: Itast ista huius similis nostrai tua. Ter. Heaut. 1020 : Tui similis est probe. Ter. Phor. 501 : Qaam uterque est similis sui. Cic. Quinct. 38: Quis tam tui, Sexte, dissimilis. Cic. Cat. 1-5: Cum jam nemo .... tam tui similis inveniri poterit. Cic. Tusc. 1-81 : Vita omnium perditorum ita similis. Cic. Tusc. 3-23: Metus non est morbi admodum similis. Cic. Legg. 1-29: Sui nemo ipse tam similis esset, which may be compared with the dative in same construction in the same flection : Nihil est enim unum uni tam simile. Scarcely less clear are cases like the following, in which sinii- Us with the genitive is parallel with another adjective, the noun being present. Cic. Div. 1-88: Amphiaraus et Tiresias, non humiles et obscuri neque eorum similes . . . . sed clari et prae- stantes viri, qui ... . futura dicebant. Cic. Brut. 51 : Rhodii saniores et Atticorum similiores. Cic. Phil. 2-66: Incredibile ac simile portenti est. Cic. De Or. 1-184 : Haec igitur et horum similia jura suae civitatis ignorantem .... prope cunctis civi- bus lucem ingenii et consilii sui porrigentem .... nonne in- primis flagitiosum putandum est? Leg. Agr. 3-5: Omnium legum iniquissimam dissimillimamque legis esse arbitror eam, quam, etc. Other cases in which si7niUs with the genitive does not seem to be a substantive are such as the following: Cic. Tusc. 5-45: Videamus ne, ut acervus ex sui generis granis, sic beata vita ex flui similibus partibus effici debeat. Phil. 1-5: Nam cum .... magis magisque perditi homines cum sui similibus servis tectis ac templis urbis minarentur. Cluent. 79 : Hanc deinde suspi- tionem augeret Staieni improbitas et non nullorum eius similium judicum turpitude. Many other examples might be given to illustrate the use of similis with the genitive as an adjective. For, as has been said, relatively very few of them make the impression that similis is used as a substantive. Other examples, however, would not make the case more clear than those already cited ; so the list need not be extended.®' 87 The following passages may be compared : Cic. De Or. 1-189, Or. 39, Verr. 2-3-163, 2-4-16, De Dom. 83, Phil. 2-66, Ad Fam. 11-3-1, Ad Att. 14-18-2, Tusc. 1-43, Div. 2-37, Fat. 3, De Off. 1-81. 28 Case Coxstkcctioks of Simius and its Compounds It thus appears that none of the views proposed with reference to the doable case construction of simiUs and its compounds ,8 ratisfLtory. It remains, therefore, to present in tabulated form T^lZt a considerable range of Latin authors as to the con- struct'o;s in question and to draw the conclusion which these statistics yield. Fragments of the Early Poets. Since the passages are so few, they may be cited. Genitive. Naevlus, com. fr., Ribbeck 60. Pol bant parasitomm alionim [bic] similest. Dative. Pacnvins, Ribbeck 374. Id magis veri simile. Enniiifl, Satires, Babrem 490, Mul- ler p. 86. Simla quam similis, turplisima bestia, nobis ! Lncillns, Babrcna 232, Muller p. 40. Qnod pnero similis. Accins, Ribbeck 404. Silvani melo Af ranins, Ribbeck 29. Consimilem ad anrls cantnm et Terenti nnmne similem dicent anditnm refert. qnempiam ? Af ranins, Ribbeck 397. Ubi qnid repentino bnins con- simile accidit. Novins, Ribbeck 62. Tn pncri pansilli simile es. Laberins, Ribbeck 124. Sepnlcri similis nil nisi nomen retineo. Besides, Titinius, Eibbeck 34, has persimilis, but the case is not clear (formicae). , ,. . ». A word may be added in regard to some of these fragments. For instance, there is no reason to doubt that Cicero, Nat. Deo. 1-97, quoted Ennius correctly.'' In writing the dative here for an original genitive, to say nothing of the common usage of the older poets, he would have violated his own rule. For he has only one instance of similis with the dative of the personal pro- noun (De Or. 3-47). »Cf., howeTer, MuUer, Q. Ennlut, pp. 1«9 .nd 271, for carelesinei. of Cicero in quoting. f Case Constructions of Similis and its Compounds 29 In the Accins passage (Cic. Nat. Deo. 2-89) rmlo is cited by Georges (Lex. der Lat. Wort. s. v.) as ablatiye. Ussing, how- ever, (PI. Amph. 595), Ritschl (Suet. Rehq. Reiff. p. 523), and Loch- take it as dative. Cf. Nene 1-503 (1902). In a fragment qnoted in Cicero, Tnsc. 2-36, Ribbeck - finds the genitive with simiHs, while Baiter and Keyser and Mtiller assign the genitive not to the fragment but to Cicero. ^.. . In Afranius (Ribbeck 29), quoted in the Suetonian Life of Terence, the MSS do not agree, but the genitive is read m the best MS.*^ table I. Use of Similis and its Compounds in Prose. Table .bowing timilis and Its compounds as used in Cornificius, Varro Cicero, Livy « Quintilian,»» Tacitus, Suetonius, Lactantius. Firmicus. and the Vulgate. The genitive of the personal pronoun and verum is excluded ; also elliptical expressions. Genitive. S3 I- 00 3 4J 5 > 3|d 00 d a o « 3 OB 3 3 s DATrV'E. ao 31* ±5 3 O s « a •» 00 3 3 si 5 n 3 u ' s similis dissimilis. • . • • adsimilis consimilis persimilis absimills • • ^ Persons compared with persons, 8 Things compared with things..! i Animals compared with animalSj . . Persons compared with things..].. Persons compared with animals i Things compared with animals External likeness Internal likeness General likeness •* 6 6144 ll 13 1 1 • ■ 65 1 1 2 1 2 29 73 67 6 1 2 7 6 22166 43 63 1 2 9 i ■ • 9 13 69 31 3 .. 3 1 .. .. 1 2211 2 39 .26 6 12 61 4 2 i ■ • 17 102410114 • . 3l 1 . ■ { • • 9 33 62 3 316 3 6 12 36 27|17 I 8 3 .14 611 • • • • • • 61 32 9 16 7 • • • 24 60 41 «»De genltlvl apud priscos scriptores Latinos usu, Bartenstein prog. 1880, » Tr»g. Rom. Fragmenta, ex Incerti. Incertorum Fabulis, 806. Cf. Loch aa cited. »» Ritschl, Suet. Reliq. Rciflf. p. 482. •"Livy, Books I-X and XXXI-XL. »» Quintilian, Books I-VI. •*Thl8 class, of no consequence in this study but added for completeness, includes an cases not clearly involving external or internal likeness exclu- sively. There are many such. Cf. p. 11 of this paper. 30 Case Constructions of Similis and its Compounds TABLE II. Use of Similis and its Compounds in Pobtrt. Table showing similis and its compounds as used in Plautus, Terence, Lucretius, Vergil, Lucan, Silius Italicus, Martial and Juvenal. The geni- tive of the personal pronoun and verum is excluded ; also elliptical expres- sions. Genitive. Dative. 3 3 • 1 • • 3 £ 3 • ^ ! U c • « 3 O • i « : e » o is • 00 3 3 .2 8 e 8 3 £ 3 1^ • • 3 i2 9 o 1 • OQ • 1 >• similis dissimilis adsimilis. . . consimilis persimllis 16 • • 1 2 • • 6 6 7 • • • • 1 • • 8 5 8 • • • ■ 8 • • 2 4 1. • • • • • • • • • • • • • • • • • • • • ft ■ • • • • • • • • • • • • • « • ■ • • • • • • • • • 4 • • • • • • i 8 « • • • • • 17 1 1 • • • • 7 ft 8 2 2 8 1 • • • • • • 1 8 • • • • • • 18 2 • • • • • • 9 7 • • 2 2 18 • • • • • • • • 8 8 1 • • 1 8 2 • • • • • • 8 x^ersons compareci wrim ucxbuub. • 1 ft 8 jLnings njiiipcivcu n iiu tuiueo. . .•...•..••••« • • 1>o vfiTkn o r^rtmnn 'TPrl witVi t.hlTlirS. ....... > 6 2 • • • ■ • • • • • • • 1. • • • • • • • • • • • 3 '^)«&v.ar%T«ci #«^%TviT\aif. A/1 n7 1 f Vk a n 1 m A 1ft ...... 1 ^^l«4v\rve! #%^XTV1 Yka 1*0^ ^ITI^h fk D 1 TTin 1 S ... .*.. W-v-t-oT-nol lilronAftft ........................ 8 1 7 6 1 . 4 i 4 1 18 1 • • 416 11 1 1 4 JCtJiliKLLlai lllkCWCSO. ... .... .... ........ «....... 16 • • • • 4 1 1 /^^wKkTKil IiItatiacq .. .. ..a...... 1 » vrcuorai llKcuOBa • |-- 1 1 Case Constructions of Similis and its Compounds 31 TABLE III. Use of Similis and its Compounds in Prose. Table showing use of similis and its compounds otherwise than as in Table I in authors named in that table. Genitive. Use of Similis and its Compounds in Poetrt. Table showing use of similis and its com- pounds otherwise than as in Table II in authors named in that table. Dative. • 00 3 f^ c a O • 1 i 3 • > •J 1 c 3 • 00 2 • 3 3 2 o 3 GQ 00 3 C s • 00 3 a •^^ o i 1 3 > 00 3 3 s • 00 3 2 3 1 • 3- 1^ • 00 3 O "^ "oS l-H • OQ • i > 3 •-9 Blmilis mei 5 1 9 6 86 11 2 2 2 10 4 diRfliniilifl mpi .••••••. ftimiliR tui 1 19 8 dlHSimiliB tui sifTiilis Riii 1 1 • • • 1 1 1 ... 2 diRRimilia aui 1 fiimilis noBtri 2 difiBimiliB nostri similis vestri 8 1 1 dissimilis veatri Bimilia mihi I similis tlbi 5 10 1 2 ... 1 1 8 Bi mills sibi Bimilia nobis Bimilia vobia Similis veri similis vero Bimilis veria 9 • • • • • • • • • 77 • • • 9 2 1 • • • • • • 10 2 1 1 2 • • • 6 1 • • • • • • • • • • • • • • • • • • Bimilia inter ba 12 4 9 12 1 1 1 8 • • • • • • 2 9 • • • • • • • • • dissimilis inter se. . . . Bimilis intAF nosmAt. . • • • • • • diRsimilis intpr vfia .. > • • • • • Bimilis intt»r pob RimillH atouB 1 • • • • ... 1 2 1 8 consimilis a.taue similis ac 1 • • "i 1 Aim II in AC fti - • • • diflfiiniilis ac ai 1 Bimilis At 2 3 1 dissimilis et Similis -que Rimilinpt si • • • • « • 1 AiniiliAiit • • • 1 ftimilifl lit. Ri 1 ainiiltR rIc lit - 1 RimillR tAmniiiiiiri Ri. ftimiliR niiA . .-* --.-- 1 fiimiliR Olio --..-.. 1 1 fliiniliR oiialla 1 ftimiliR niiARi .... AdftimiliR niiflRi ....... 1 2 1 coTiRimiliR nuRRi * conRimiliR vplnt dissimilis in 1 diftRimiliR all 1 32 Case Constructions of Similis and its Compounds TABLE IV. The Usage of Cicebo. Table showing similis and its componnds as used in Cicero. The genitive of the personal pronoun and verum is excluded ; also elliptical expressions. similis disslmilis adslmilis consimllls perslmilis Persons compared with persons Things compared with thing's — Animals compared with animals Persons compared with things Persons compared with animals Things compared with animals External likeness Internal likeness General likeness Genitive. S 8S 7 24 16 9 8 23 a o O 40 2 1 26 16 4 16 21 i 16 9 6 2 12 1 I 67 4 29 27 1 1 1 2 14 88 9 Dative. I 2 16 2 17 2 6 11 a 3 g o 8 2 2 9 i 2 I ft Pu 86 8 28 3 1 1 8 26 6 Table showing use of similis and Its compounds otherwise than as above in Cicero. similis mei disslmilis mei similis tui disslmilis tul similis sui disslmilis sui similis nostri disslmilis nostri similis vestri disslmilis vestri similis mihi similis veri similis inter se disslmilis inter se . . similis inter nosmet. disslmilis inter vos. . similis inter eos similis atque similis ac si similis et similis et si similis ut si similis tamquam si.. similis quasi 1 i • a 1 5 1 o 2 2 2 1 16 6 12 1 1 8 4 18 6 6 1 • • 22 1 • • 1 • • 1 • • 8 • • 2 2 6 1 2 2 2 -a a, S o mm 0U 1 • • • • 18 2 • • 86 8 1 1 1 1 1 1 Case Constructions of Similis and its Compounds 33 Remarks on the Usage of Individual Authors. Plautus, Besides the cases shown in the table there are two instances of the genitive in elliptical expressions" and two other cases in which editors read the genitive, though the MSS show no case at Six verses deserve special mention." They are: Amph. 601, Capt. 582, Men. 1088 and 1089, Mil. 240, and True. 505, none of which are preserved in the Ambrosianus. The Palatine recension shows the dative. In all these cases, however, Ritschl read the genitive, regarding this the current construction in colloquial language " and the only one used by Plautus,"* and his readings have been followed by most succeeding editors.'^ Engelbrecht, however, (Stud. Ter. 38), Brix (Capt. 116, 1884), and Lorenz (Mil. 240, 1886) accept Ritschl's views with some qualifications, while Spengel (Phil. 1861, 565), Ott (Zs. f. d. ost. Gym. 1871- 149), Ussing (Amph. 595), Drager (Hist. Syn. 1-445), and Fabia (Ter. Eun. 334, 1895) are at variance with his conclusions. It is necessary, therefore, to examine these cases with care, and in our examination of them to keep three things well in mind: (1) that the dative is unquestioned in Ennius,^**^ (2) that the dative is the prevailing construction with the comparative of 5iw- ilis.^^ Ai>&Tt from veri similius, sui similior^ etc., which as stereo- typed expressions should not be considered, similis in the com- parative is seldom used, so that our material for comparison is limited. It occurs once in Lucretius, in Varro twice, in all three cases with the dative. In Cicero it is found five times,'*' twice with the genitive, three times with the dative. And we must re- member (3) that there is no MS evidence for the use of the geni- tive with the comparative of similis in Plautus.'®* On the other » Amph. 267 and Asin. 241. »«Poen. 613 and True. 507. *7The Bacchides fragment 19 (Gotz), as depending on the grammarians need not be here considered. * Op. 2-581. » Op. 2-572. »~ Lorenz, Most. 88 (Ed. 1866), Kuhnast, Liv. Syn. 125, Lindsay, Capt. 116, Brix-Niemeyer, Men. 1088, Loch as cited on p. 29 of this paper. »»» Cf. p. 28 of this paper. »m Gildersleeve, Gram. p. 229. i»»Brut. 51 and 148, Cluent. 88, De Fin. 4-80, Div. 1-13. In Tusc. 4-53 the form is donbtfnL *<** In Mil. 552 the better readings show no case, the inferior ones give a form that may be either genitive or dative. 34 Case Constructions of Simius and its Compounds hand in the only three instances in Plautus in which the com- parative of sirnilis occurs with a case the MSS show the dative. The instances are: Amph. 601 : Neque lac lacti magls est simile quam ille ego similest mel, and Men. 1088-9 : Nam ego hominem homini simlliorem numquam vidi alterum, Neque aqua aquae neque lactest lacti, crede mihi, usquam similius. In these last three cases, then, in view of the facts cited, that the dative with similis is not questioned in Ennius, that the MSS of Plautus show no instance of the genitive with similis in the comparative, and that later the dative with the comparative is the prevalent construction, there seems to be no good reason for changing the MS reading. Ritschl's reasons for changing the readings are singularly inconclusive.^"^ For example, in Men. 1088 he reads hominis for the MS Jiomini, because tui and hums of line 1090 demand the genitive in 1088. This reasoning not only disregards the change from the comparative with homini to the positive with tui and huius, but also the fact that ^wi is prac- tically stereotyped in such expressions as we have here, and, if any explanation of huius were necessary, nothing is more natural than that huius, a pronoun used of a person, should follow the stereo- typed personal pronoun tui. And, having secured thus, by re- vision, a dative in 1088, he changes lacti of 1089 into lactis to bnng it into agreement with 1088 and 1090. To be sure, Plautus, as Ritschl contends, will use only the genitive with similis, if all his datives are changed to genitives. It seems remarkable that in the entire discussion of these passages no one has called attention to the fact that comparatives are here dealt with. In Mil. 240 : Tam similem quam lacte lacti est, the MSS read lacti. Here Ussing reads lacti est, other accessible editors lactist. This lactist Ritschl and other editors write as standing for an original lactis est.'"" But the reasons for considering this another instance of the dative in Plautus are scarcely less conclusive than in the three verses where similis in the comparative is used. For, if we admit the dative in Plautus at all, we might readily admit it in this tam quam construction, which shows a distinct preference for the dative. Excluding examples of the genitive 103 Op. 2-571 and 580. iM Ritschl, Op. 2-570, Leo, Forscli. 260, Lorenz, Brix, and Tyrrell In their notes. Case Constructions of Similis and its Compounds 35 of the personal pronoun with similis and De Off. 1-121, where the MSS differ, there are four tam quam correlations in Cicero in which similis with a case occurs.^^ In one of these the geni- tive is used, in three the dative, the dative once when a person is involved (Brut. 204) in spite of Cicero's strong preference for the genitive in such cases.^** Besides, sifnilis with either tam or quam without the correlative shows the same preference for the dative that the correlation shows. Thus sirnilis with a case occurs in two sentences in Cicero with the exclamatory quam, and both times the dative is found.'^ Likewise, disregarding the genitive of the personal pronoun, tain similis with a case is found only once in Cicero and this once with the dative though persons are compared.'*®* The preference that sirnilis shows for the dative in these tam and quam sentences is no doubt accounted for, as in cases of similis in the comparative, by the stress laid on the ad- jective nature of similis by the comparison in the one case and by the modifying adverb in the other. How strong this preference is may be seea in the fact that the tain and quam sentences in Cicero show in round numbers a per cent of in- stances of similis with the dative of the person ten times greater than normal. It is to be noted that of these four datives under discus- sion in Plautus three are found in a comparison of milk with niilk,"° while the fourth »" one stands in immediate connection with this comparison (which appears again in Bac. fr. 19, Gotz,"* where the reading, as depending on grammarians and variously quoted in the critical notes of Gotz, Leo, and Ussing, need not here be considered). This milk comparison, therefore, is evi- dently proverbial, and it is, to say the least, interesting to note a verysimiliar proverb in Quintilian, 5-11-30: illud, Non ovum tam simile ovo. Here is the proverb (marked by illud) as in Plautus, here is the tam as in Plautus, and here is the dative as ><" With dative Brut. 204, Cato 80, Legg. 1-29 ; with genitive Brut. 285. »w In comparing persons, exclusive of personal pronouns, Cicero uses the genitive 88 times, the dative 9 times. »«» Verr. 3-4-77, Nat. Deo. 1-97, a quotation from Ennius. In Phil. 2-26 the interrogative quam is found with veri similey but veri simile is constant till after Cicero. lowiDeFin. 5-62. "0 Amph. 601, Men. 1089, Mil. 240. "» Men. 1088. »»«Cf. Peine, Diss. De Dativi Usu apud Priscos Scriptores Latinos, p. 93. i\ 36 Case Coxstructions of Similis and its Compounds in the MSS of Plautus. Moreover, the stereotyped form of pro- Terbial speech must not be forgotten."' In view of all the facts the conclusion seems warranted that in Amph. 601, Men. 1088 and 1089, and Mil. 240 the dative, which the MSS of Plautus show, should be retained. Only two other passages In Plautus need to be noted. In Capt. 582 the MSS read omnis inveniri (MSS -ire) similis tibi vis, and in True. 505 the MSS, differing in other respects, agree in reading mihi. Now, since in the four passages just discussed, the dative is to be retained, since an unquestioned dative of the personal pro- noun with similis is found in Ennius"* on the one hand and in Cicero"* on the other, there is certainly no imperative reason for departing from the reading of the MSS in Capt. 582 and True. 505."« Terence. There are only two passages that call for special notice. One of these is Heaut. 382, which Umpfenbach reads thus; Id cum studuisti, isti formae ut mores consimiles forent. This is like- wise the reading of all accessible editors except Fleckeisen "^ and Shuckburgh, who read as follows: Id tu quom studuisti, formae ut mores consimiles forent. The omission of isti by DEG, requiring, as it does, compensation elsewhere in the verse, does not commend itself against the testimony of the other MSS, especially since the omission would easily be explained by the preceding isti. The word must, therefore, be accounted for as it stands. There are three possible explanations. (1) It may be regarded as genitive =istius. Georges cites this place with Plautus, True. 930 (which Spengel suspects) and Cato, fr. oratt. 20 "» (Jordan), as examples of the genitive isti. So Wagner "» explained it, as also Neue 2-398 (1892). But it is to be noted that in the other two cases cited isti is used in connec- tion with modi, and Buecheler (Lat. Dek. 78) and Engelbrecht '" Cf. Ott, Zs. f. d. ost. Gym. 22-149, who regards the dative as constant in this milk proverb of Plantns. Cf. also Ott, Sprichworter der Romer p. 183, and 8utphen, Amer. Jonr. Phil. 22-144. "♦ Cic. Nat. Deo. 1-97. ii5 De Or. 8-47. "«In Capt. 582 Ussing, Sonnenschein, and Hallidie retain tibi, and Brix (Capt. 116) also accepts it. In True. 505 Ussing reads mihi and Brix (Capt. 116) accepts it. 117 Bnt Fleckeisen in ed. of 1898 returned to isti. >>8 Also Accius, fr. 136 (Ribbeck), but the MSS have istiu». »• Cf. the notes of Shuckburgh and Gray. f Case Constructions of Similis and its Compounds 37 (Stud. Terent. 38) limit this form of the genitive to the combina- tion with modi. In this place, then, isti cannot be said to be satisfactory as a genitive. (2) Dziatzko in his critical note construes isti with snores and suggests the order: isti ut mores formae. This explanation does not commend itself for two reasons. As Dziatzko's proposed transposition indicates, the position of isti favors construing it with formae. The sense, too, favors this construction, for the forma, as something clearly perceived, serves as the standard of comparison, while, as v. 384 shows, the mores is a matter of in- ference. (3) It may be dative with for7nae. This is a more satisfactory explanation, since the dative form is regular while the genitive is hardly admissible, and since neither sense nor position favors the nominative. Engelbrecht takes it to be dative, remarking that even if the genitive predominated with similis in the comic poets, specially Plautus, it need not surprise us to find the dative in the more elegant plays of Terence, and this observation has added force, since it has been made very probable that the dative is found in Plautus. Schleuter ^'^ and Peine "' agree with Engel- brecht in regarding isti as dative here. The other case is Eun. 468. Here Umpfenbach reads : Per- pulchra credo dona aut nostri similia. A reads nostri. The other MSS read nostris, as do Priscian (K. 3-34 and 115) and Donatus. Neglecting Ritschl's suggestion of nostrum = mstrO" rum *" as without MS authority, the readings nostris and nostn are to be considered. Nostris. This reading is not sustained by the best MS, but it is easy to see how the copyist might have written nostri for nostris. In view of the stereotyped genitives of the personal pro- noun with similiSy mei, tui, nostri, etc., force of habit on the part of the scribe might have induced nostri here, and besides the initial 5 of the next word might have contributed to the same result. Fabia's objection to nostris on the ground that, the re- ference being to Pamphila only, the sense requires the singular, is not of considerable weight The reference might easily be to the soldier's gifts generally."* With nostris the meaning is per- ** De Ace. et Dat. Usu Terentiano p. 36. "> De Dat. Usu apud Pris. Scriptt. Lat. p. 91. "« Engelbrecht, Stud. Ter. 88. »»Cf. Sonnenschein, Bud. 728, Brix, Men. 290, 739, 803. 38 Case Constructions of Similis and its Compounds fectly clear, while fiostri is quite misleading, as will be shown below. Furthermore, the dative of the possessive pronoun, the noun being omitted, is found elsewhere with similis. Varro, L. L. 10-71: tertia parum nostris similia. Lucan 9-514: Stat sortiger illit Juppiter, Qt memorant, sed non ant fulmina yibrans Ant similis nostro. Bentley and Klotz read iiosiriSj which is likewise accepted by TJssing, PI. Amph. 595 and Lorenz, Mil. 240. But in view of the reading of the best MS 7wstris should not be insisted upon. Nbstri. If, with the best MS, we read iwsiri, the form may be either the genitive plural of the personal pronoun, or the geni- tive singular of the possessive pronoun. Taking fiostri to be the genitive plural of the personal pronoun, there is of course ellip- sis, nostri similia, like (the gifts of) us. This ellipsis is by no means uncommon in early or later Latin. A good example is found in PL Amph. 267: Et enim vero qnoniam formam cepi huins in med ct statam, Decet et facta moresque huius habere me similes item. But the elision here would be so harsh that we cannot regard Tiostri as the genitive of the personal pronoun. It may, however, be the genitive singular of the possessive pronoun. So Papillon and Fabia regard it, making it agree with a suppressed doni. Such a construction is misleading, to say the least. For the genitive of the personal pronoun with similis is extremely common with stereotyped meaning: similis mei, tui, nostri, vestri, sui, like me, like you, like us, etc. Hence to replace the genitive of the personal pronoun with similis by a possessive of the same form is wholly unexpected and misleading. Indeed no such possessive nostri, vestri, mei, or tui, the noun being omitted, has been noticed. There is, however, one case of sui thus used. It is found iniCic. Off. 1-121: superioris filius Africani, qui hunc Paulo natum adoptavit, propter infirmitatem valetudinia non tam potuit patris similis esse, quam ille fuerat sui. Here, however, sui is disjoined from similis, and it is easy to supply the preceding patris after sui. But in our passage nostri is joined immediately to similia in a most misleading way, and do7ii with changed number and case is not easy to understand from the preceding dona. Still the reading nostri found in Umpfen- bach, Papillon, Fleckeisen, Dziatzko, Fabia, and Tyrrell, is per- f Case Constructions of Similis and its Compounds 39 haps to be retained, since it rests on the best MS and is sup- ported by the parallel construction in Cicero. Cf. also Job, 16-4 : Poteram et ego similia vestri loqui. Lucretius. On the cu7n et cum construction of 2-416 and 5-1060 and the qui et qui construction of 2-419 cf. Munro, Lucretius, 1-280. See also for ' -que ' of comparison remarks on Livy in this paper Vergil, It is to be noted that Vergil uses the genitive with similis only once (Aen. 5-594). It is also worthy of remark that of the nine- teen datives with similis seven are datives of the present parti- ciple. ^ Livy. Only eight genitives are found in the two decades examined (I-X, XXXI-XL), and six of these occur in the first decade. The percentage, too, is higher in the first decade. Of these eight genitives three are personal names, two are pronouns referring to persons, and two are nouns denoting persons. The remaining gemtiveprodigii (31-12-8) is no doubt influenced by the genitive hitherto constant in such expressions as mo7istri, portent i, prodi- gii si7nilis. On the contrary, it may be observed, as evidence of the declin- ing genitive that Livy uses portento similis twice and that simi- Ims vero ''' in two places in the first decade replaces veri constant in earlier authors. Bk. 10-28-1, haudquaquam similis pugna in dextro laevoque cornu erat, shows a -que which is no less comparative than the comparative et. Cf. Munro, Lucretius, 1-280, Kuhner, 2-636, and Drager, 2-29. Silius Italicus. The present participle in the dative occurs five times with stmilis. Cf. Vergil's use of the present participle with similis. Martial. Martial 1-109-19 is cited for similis in the sense of likeness followed by the accusative in apposition. '"But cf. Haase, Vorl. 2-141 on Cic. Ad. Fam. 13-5-1. Cf. also for i-lvj 8 nsage Haase, note to Reisig, 8-631 and Kiihnast, Llv. Syn. p. 135. 40 Case Constructions of Similis and its Compounds Tacitus. Dissimilis in is found once. Ann. 2-39 : Aetate et forma hand dissimili in dominum erat. As substantive with the dative simi- lis is found in Hist. 3-83: juxta scorta et scortis similes. On the other hand, the only genitive is found in a passage in which the comparative is used. Hist. 3-76: lascivia socordiaque gladi- atorum magis quam ducum similes. The earlier distinctions are here reversed. Juve)ial. Kesembling a substantive with the dative similis is found in 5-132 and 8-53, whereas the only genitive (sui) is used with dis- similis as an adjective. In 2-6 is found similem as a substantive in the sense of like- ness, apparently in apposition. (Mart. 1-109-19, Statius, Silv. 3-3-201 and 5-1-1 are cited as showing this later and unusual construction.) Smtonius, Of the fourteen datives two are present, three perfect parti- ciples. Lactantius. The solitary genitive is found in 2-4-2: Cum aves ipsae . . . simulacris fabre factis, id est, hominum plane similibus, insidant. The Vulgate, Dissimilis with ab and the ablative occurs once. Dan. 7-19 : Post hoc volui diligenter discere de bestia quarta, quae erat dis- similis valde ab omnibus et terribilis nimis. Some cases of interchange between genitive and dative of pro- nouns are worth noticing. Here there is no distinction made between the cases. Gen. 2-18 : Faciamus ei adjutorium simile sibi. Gen. 2-20: Adae vero non inveniebatur adjutor similis ejus. Eccli. 13-20: Omnis caro ad similem sibi conjungetur, et omnis homo simili sui sociabitur. Eccli. 45-7: Excelsum fecit Aaron fratrem ejus et similem sibi de tribu Levi. The dative of the participle occurs with similis once. ♦ ««*«» A glance at the table shows: (1) That the dative with similis runs through all periods and departments of the language. Case Constructions of Similis and its Compounds 41 (2) That the genitive with similis (except in set phrases) prac- tically belongs to the earlier periods of the language. (3) That the genitive with similis is almost wholly excluded from the higher forms of poetry."* (4) That the genitive with similis prevails very largely in the comic poets, falls behind the dative by a half in Lucretius and by nearly three-fourths in Varro, but in Cornificius is equal to the dative, and in Cicero,"« who uses the genitive far more than any other writer except the comic poets, is relatively much more frequent in those writings which have much in common with conversational language (i. e. in the warm, personal orations and letters) than in the rhetorical and philosophical works. The conclusion would seem to be that, while similis with the da- tive IS a natural construction in all periods of the language, the genitive is used mainly in periods and departments which mark it as an inheritance from the common speech that passed into the literary language, suffered an early decline, and finally, except in the significant formula noted below, fell into disuse. The origin of the construction of similis with the genitive is probably to be sought in the familiar combination of si7nilis v^ith the genitive of the personal pronoun. For it is a most note- worthy fact that, while the dative supplants the genitive every- where else (a partial exception being found in veri simile ^"), the combination of similis with the genitive of the personal pronoun persists through every period and range of the language, the Vul- gate, however, showing similis with the dative of the personal pro- noun quite f requently."« In the nature of the case, similis with the personal pronoun is an easy, familiar, colloquial form of speech. And so, exclusive of similis with sui (which in view of its fre- quent reference to non.personal objects is manifestly less strictly personal than the other persons of the personal pronoun), simi- lis with the genitive of the personal pronoun occurs sixteen times in the orations of Cicero and thirteen times in the letters, while '25 Tables are very incomplete for poetry, to be sure. Cf., however, Wil- kins on Clc. De Or. 3-47, who finds only two examples of the genitive in the Augustan poets, Verg. Aen. 5-594 and Hor. Sat. 2-1-3. Cf. also Madvig Uc. De Fin. 5-12, Ritschl, Op. 2-581, Engelbrecht, Stud. Ter. p. 38. '"For Cicero as "antiquated" see Teuffel I. p. 250. >" In this combination constantly in use the familiar genitive with similis became stereotyped. '^Cf. Bennett and Bristol, the Teaching of Latin and Greek, p. 139, for the crowding out of the genitive by the dative. 42 Case Constructions of Similis and its Compounds in all the other writings of Cicero it is found only five times. Likewise, in the comic poets the construction is found eight times, while in all the other poetical authors examined it occurs once only, and that in a superheated personal passage, Lucan, 6-244. Moreover, in many of these combinations similis is used with- out a substantive, that is, similis is used as a substantive. In- deed in Cicero, of the twenty-nine examples of the substantive similis with the genitive, twenty are genitives of personal pro- nouns.^^ Used as a substantive similis would naturally be con- strued with the genitive, as in Plautus, Persa 698 : Videor vidisse hie forma peraimilem tni.^^o ** the very image of you." Here, then, we probably have the origin of similis with the genitive. It began in combinations of a substantive similis with the genitive of a personal pro- noun. The substantive similis thus taking first the genitive of the personal pronoun, would then easily take the genitive of other pronouns referring to persons and the genitive of personal names, then the genitive of names of things, and meantime the construction of the non- substantive similis with the genitive would be a further easy extension. And, as a matter of fact, it is found that in the twenty-nine examples of the genitive with the substantive similis in Cicero the genitive of the personal pro- noun occurs twenty times, of other pronouns six times, of names of persons twice, and of the name of a thing once. This view that similis with the genitive is colloquial "* finds striking confirmation in certain passages of Cicero in which, ac- cording to Landgraf,^^^ marks of the common speech are found. For, considering the relative infrequency with which both the colloquialisms of Landgraf and similis with the genitive occur, it will be observed that they are found together with surprising frequency. A few instances may be given. Landgraf and Wolfflin "* assign diminutives in -cuius to the common speech. Cf. with this statement Cic. Nat. Deo. 1-123: Neque enim tam desipiens fuisset, ut homunculi similem deum >»Cf. p. 26 of this paper. 1^ Cf. for imago so used Plautas, Cas. 515, Nunc amicine anne iDimici sis imago, Alcesime, mihi sciam. »3» Cf. Ritschl, Op. 2-581. i3> Blatter f. d. Bayerische Gymnasial- and Real-Schulwesen, 1880. »MPhil. 35-153. Cf. Teuf. I. 214-9. Case Constructioxs of Similis axd its Compounds 43 fingeret. Or. 67: Nisi quod versiculi sunt, nihil est aliud cotidiani dissimile sermonis, and Verr. 2-3-155 : Volo, mi frater fraterculo tuo credas. Consorti quidem in lucris atque (in) fur- tis, gemmo et simillimo nequitia, improbitate, audacia. In the last example, which shows a substantive similis, is likewise asyn- deton, which Landgraf (324) mentions as a mark of the common speech. Landgraf (322) assigns 7mmis minusque to the common speech. Cf. with this statement Cic. Phil. 1-5: et cotidie magis magis- que perditi homines cum sui similibus servis tectis ac templis urbis minarentur. Thie expression tela texere was not noticed in Landgraf's list of colloquialisms. It is found in Plautus, however, (Pseud. 400) and surely has the tone of common speech. With this compare Cic. De Or. 3-226 where similis is substantive: quamquam ea tela texitur . . . . ut eorum civium, quos nostri patres non tule- runt, jam similes habere cupiamus. In tela texitur is the alii- teration, too, which Landgraf (329) likewise assigns to the com- mon speech. These passages may serve as an indication of the company similis with the genitive keeps. Finally, from another stand-point, a still more striking con- firmation of the correctness of the view that simUis with the geni- tive is an extension of a colloquial use of which similis with the genitive of the personal pronoun is the germ, is found in the fact that this view is easily seen to explain the partial truth contained m all the other theories that have been advanced on this subject. First, there is the theory of the old Latin grammarians that the genitive is used of inner, the dative of outer likeness. Taking into account the genitive of the personal pronoun with similis this IS true, the reason being that the comparison would usually be made in view of likeness in character. So in our language * the like of you,' etc. is used of likeness in character. Then, too, Madvig's distinction for Cicero, that the genitive is mainly used of persons, is just what would be expected in view of the fact that the germ of the construction was the genitive of the personal pronoun. Haase's explanation, also, that similis with the genitive means EhenUld, Ahhild, image, is largely true, if limited to this con- Btruction in its original form of similis with the genitive of the 44 Case Constructions of Similis and its Compounds personal pronoun. Used with such a genitive similis is often equivalent to EhenbiW^^ Likewise, the single remaining theory, that similis with the genitive is a substantive, finds its basis in the original construc- tion of similis with the genitive of the personal pronoun, in which, when the substantive was omitted, similis itself was used as a substantive. Cf. the example in Plautus, Persa 698 : Videor yidisse bic forma perslmilem tni. To summarize, therefore : the statement of the facts with refer- ence to the double case construction of similis may be made thus. The genitiye with similis very probably had its origin in the col- loquial use of the substantive similis with the genitive of the per- sonal pronoun. This easily extended to the genitive of other pro- nouns referring to persons and to the genitive of the names of persons, and finally to the genitive of the names of things, while, in the process of extension, the genitive came to be used with the adjective as well as with the substantive similis. The construc- tion with the genitive reached its highest point in those works of Cicero, which in their warm, personal quality stand nearest to the common speech. After Cicero, however, the dative, which, as the normal construction, was used from the earliest times, re- placed, under the influence of poetic usage,^** the old and collo- quial genitive. There was, however, a period, specially represented by Cicero, in which both genitive and dative with similis were in common use. Within this period, when, on the one hand, the substantive nature of similis is prominent, the preference for the genitive practically excluded the dative, and, on the other hand, when the adjective nature of similis is stressed, as by the comparative de- gree, the tarn . . . qiiaryi correlation, or the exclamatory j'waw, the dative is very distinctly preferred. Between these extremes there is a wide middle ground in which genitive and dative are used with no perceptible difference in meaning. It may be worth while to add that a careful examination of the orations and philosophical works of Cicero serves to show that the relative position of words has no influence on the case con- i»*Cf. p. 26 of this paper. iwFor the influence of the poets in a general way in producing the liter- ary language from the common speech cf . Wolfflin, Phil. 34-149. Cf. Teuf. I. pp. 250 and 411 and II. pp. 4 and 7. Case Constructions of Similis and its Compounds 45 struction. Since similis^ the verb, where there is one, and the two objects compared are all concerned in the arrangement, the order varies very much, but the arrangements more frequently occurring are common to both genitive and dative. Special treatment of the compounds of sifnilis is not necessary. The tables, pages 29, 32, show that the compounds follow the construction of the simple adjective. A glance at the table on page 31 will show how very rarely similis is used with other constructions than the genitive and the dative. Several of these constructions, as similis ac si, et si, ut si, tamquam si, and others, are found only once. LIFE. . Thomas Madison Jones was born near Doe Hill, Highland Co., Va., August 4, 1860. He attended the public schools of his county for several years, but most of his early education is due to James W. Johnson, under whose care for nine months he received invalu- able training. In 1885 he became a member of the Baltimore Conference of the Methodist Episcopal Church, South, and in 1891 went to Randolph- Macon College as a student at the age of thirty-one. He graduated with the degree of A. B. in 1894, and the following year taught in Randolph-Macon Academy at Bedford City, Va. He was a graduate student of Latin, Greek and German at the Johns Hopkins University for the years 1895-1898, where it was his privilege to be under the instruction of Professors Warren, Smith, Gildersleeve, and Wood. He held a University scholarship in Latin in the last year of his residence and was recommended by the department of Latin for appoint- ment to a fellowship for the following year. For the next two years, 1898-1900, he was Professor of Ancient Languages in Emory and Henry College, attended the University of Chicago in the summer of 1900, and since June, 1900, has been Professor of Greek in Randolph-Macon College.