mr^m (^ '^ ΓΙ Ni m^Mm fan κ l•^. 1 I Κ ^-■^:^ '& ^ J 1 r^ I LIBRARY OF COXGRRSS. | i (ϋΙ,αρ. fK i 4 (c) Qualis, intensive ib. (d) Quantus, intensive ib. (e) Qualis, with substantive in its own clause in similes ib. (/) Q^^j with force of qualis * 55 CONTENTS. Vll PAGE {g) Quantus after comparatives 56 (h) Quantus and qiialis with demonstratives expressed ib. (i) Quantus and qualis, interrogative 57 {k) Tantus, talis, followed by ut and qui ib. § 34. Uses of οίος ._ 59 (a) Attraction with οίος, όττοϊος 60 (6) Interchange of relative forms 61. (c) Intensive and interjectional uses ib. (d) Redundancy ib. § 35. Uses of όσος ib. § 36. Origin of relative conjunctions and adverbs 62 CHAPTER VII. DOUBLE RELATIVE USES. 37. (a) Compound relative words 70 (0) Collocation, or mere accumulation of relative forms 71 (c) In admirative passages, or such as are distinguished by marked rhetorical emphasis 72 (d) In similes or comparisons 73 {e) Double Interrogatives ib. CHAPTER VIII. THE RELATIVE PRONOUN IN THE CONSTRUCTION OF THE DEMONSTRATIVE. 38. Examples treated of under the head of ^Anacoluthon ' 75 39. This opinion shown to be inaccurate, and that the relative is equivalent to conjunction and copula 77 (a) The relative followed by a clause commencing with si, nisi, cum, ^c. . . ib. (6) Where the relative commences a period generally 80 (c) In the construction of the ablative absolute, &c ib. (d) In imitation of certain demonstrative constructions 81 (e) Where the relative has an adjective or possessive pronoun in connexion with it ib. (/) When the subject has been taken out of its own clause and put as object to the preceding verb ib. {g) With the subjunctive in an optative sense ib. (h) With another i-elative in its own member ib. (i) With another demonstrative in its own member 82 (_;) In the construction of the accusative and infinitive ib. CHAPTER IX. CHANGE OF POSITION OF THE WHOLE RELATIVE CLAUSE. 40. Relative preceding the demonstrative 84 41. Emphatically followed hy is, hie ib. 42. The same with relative adverbs ib. 43. Peculiar instance of this sort 85 44. This practice productive of difficulty ib. 45. In passages of oratorical force ib. 46. Separation of the demonstrative and antecedent substantive, by the inter- position of the relative clause ib. 47. Certain Greek peculiarities 86 VIU CONTENTS. CHAPTER X. USE OF QUIS, QUI:, FOR ALIQUIS. PAGE § 48. Examples of this practice 87 § 49. Si qui, si quis, equivalent to quicunque 89 § 50. Si quis has the properties of a simple relative 90 (a) With ellipsis preceding it ib. (b) With transposition into its clause ib. (c) As a partitive 91 (d) As the subject of the ablative absolute ib. (e) With its case altered by attraction ib. § 51. Two forms of declining si qui, si quis, preserved ib. § 52. Similar adverbial arrangements ib. CHAPTER XI. FALSE CONCORD OF THE RELATIVE. § 53. (a) Construction according to sense 92 (b) Agreement with a particular Avord rather than the whole notion ib. (c) Neuter gender rather than any particular subject of agreement 93 (d) Agreement with the person rather than the thing ib. § 54. Relative in agreement with the predicate ib. § 55. Relative in the singular referring to a plural antecedent 94 CHAPTER XII. MISCELLANEOUS. § 56. Relative as a partitive . . . . r 95 § 57. Relative itself in the genitive plural, after other words used partitively. ... 96 § 58. Relative agreeing with an implied antecedent 97 § 59. With change of person ib. § 60. Agreement detei-mined by the sense ib. § 61. The subject of comparison put only in the second member 98 § 62. Construction of verbs of asking ib. § 63. Relative as subject and object 99 § 64. Use of the interrogative in preference to the indefinite ib. § 65. English examples , ib. § 66. Greek peculiax'ities ib. § 67. Difference bet\veen Latin and EngUsh construction ib. § 68. Difference between Latin and Greek construction 100 § 69. Greek peculiarities ib. § 70. 'όστις used for τις ib. CONTENTS. IX SECOND PART. PAGE Preliminary Remarks , . . . 103 CHAPTER XIII. OF THE RELATIVE AND VERB. § 71• General Remarks ib. § 72. Common principle involved in the use of subjunctive and potential 104 § 73. Difference between the Latin and Greek objective clause 105 § 74. Objective relative sentence 106 § 75. Relative clause not objective 108 (cr) The presence or absence of the demonstrative pronoun does not affect the case » ib. (6) Indicative in particular phrases 10.9 (c) Clause when objective, and when explanatory ib. {d) Predicate clauses sometimes objective 110 (e) The subjunctive not employed after some parenthetic verbs 112 § 76. Marks of the objective relative clause ib. (a) How the Latins mark an objective clause 113 (6) Distinction of cases 114 § 77. Other general principles ib. § 78. The Greek objective clause with the indicative in it 115 (a) Use of the optative after on, ώς ib. (6) Irregularities in the use of the optative and conjunctive in Greek ib. § 79. Direct interrogative sentence not regarded as objective in Latin 116 CHAPTER XIV. SPECIAL RULES FOR THE USE OF THE SUBJUNCTIVE, AFTER WORDS OF THE RELATIVE CLASS. § 80. When the relative clause contains the sentiments of another than the writer 117 § 81. In the oblique oration c 118 § 82. When the relative clause contains a defining circumstance of great import- ance to the sense 119 § 83. After qui causal 120 § 84. Generally Avhere the relative clause is essential to the sense of the Avhole proposition ib. § 85. After an absolute negative proposition . . 121 § 86. With the indefinite phrases, sunt qui, reperti sunt ib. § 87. After the interrogative quis est qui ? and its cases 122 § 88. In the phrases, idoneus est qui, <|-c ib. § 89. After is put for talis, and after tarn 1 23 § 90. After comparatives followed by quam ib. § 91 . With quippe qui ib. § 92, When qui is used for Mi , 124 § 93. With quod, quin ib. § 94. When the relative clause depends on a preceding verb in the subjunctive.. 126 § 95. General remark ib. § 90. Peculiar cases of atti'action ib. § 97. The subjunctive in modest assertion 127 § 98. Where an action is represented as done repeatedly ib. § 99. The direct rhetorical question in the oblique oration ib. a χ CONTENTS. CHAPTER ΧΛ^ CONCLUDING REMARKS. PAGE §100. Relative with the subjunctive, the type of the use of the subjunctive in other cases 127 § 101. Greek analogy 128 § 1 02. Explanation of occasional irregularities ib. Appendix 129 Index of Authors •. 137 Index of Matters 146 FIRST PART, SYNTAX RELATIVE PRONOUN. Besides its ordinary grammatical construction, the relative pro- noun has, in most languages, a number of uses, which, from their variety and frequency, require to be classed under new rules, rather than to be accounted exceptions. Some of these are intimately connected with, and tend to throw light on, the history and pro- gress of the language to which they belong; constituting, more- over, at one time a remarkably well-defined ground of distinction between the syntactical laws of different tongues, at another, a strong mark of their common connexion. These are best studied by an examination of examples chosen indiscriminately from Greek and Latin authors, illustrated, where convenient, from the practice of our own or any kindred language ; and also collected at large from writers of all periods without reference to style. It is to be regretted that the former of these methods, viz. of exhibiting the parallel Latin and Greek construction, has not been carried further in grammatical and syntactical works ; and more especially that it has not been followed out in books of exercises for composition, as used in the higher classes of schools. As regards the latter, viz. the unreserved selection of examples from authors of any age or style, but Uttle apology is necessary. The aim of a good syn- tactical treatise is, in general, at least threefold : first, to illustrate the origin of particular forms of construction, and the way in which diversities of expression have sprung from a common source; secondly, to furnish a key to the actual difficulties of construction to be found in any author, ancient or modern, or in whatever dialect composed, whose works are of value ; and thirdly, to supply 2 SYNTAX OF THE RELATIVE PRONOUN. rules for composition^ based on the practice of the best and purest writers in that particular tongue. In general, the examples here furnished will speak for themselves ; those chosen from writers of the purest period will be of greatest authority, and the frequency with which those of a given kind occur will supply another test of their value. No one at all conversant with the ordinary rules for composition will be in danger of attempting to engraft Homeric or Pindaric forms on pure Attic prose, or to regard some rare poetic combinations, even of the golden age of literature, as the best model for the same species of writing. In each particular style of composition, the leading writers in that style may of course be safely imitated, where examples are tolerably frequent. It would appear that rules respecting construction have in many cases, and by niany authors, been laid down too stringently ; and that these are best inferred from a sufficiently large and well-chosen body of examples set before the student. I have here merely to add, that a great number of cases of parallel Greek and Latin construction will be quoted, in which the latter appear to be a direct imitation of the former, not properly belonging to the language in which they are found, either graces or affectation of style insensibly acquired and adopted, from the study of, or devotion to Greek originals, or pedantically intended to display an acquaintance with the same. These will be carefully noticed as we proceed. I shall here say nothing of the relative as employed ordinarily, since the present work can only be of use to such as, having made some elementary progress in the classics, are desirous of pushing their inquiries beyond the sphere of a rudimentary grammar, though by no means requiring previously high attainments for its profitable use. The whole may be regarded rather as a classifi- cation of important and singular uses of the relative ; some of them, however, common enough, but still bearing on the illustra- tion of those that are less frequent, and standing at the basis of the subject. Out of these will arise observations towards the deduction of a complete system and theory of the relative and its cognates, rather than a formal statement of the same, which, in the present stage of the inquiry, it would be unadvisable to attempt. The materials for such a complete plan are as yet by no means brought together into one place. Let us hope that it admits of accomplishment, and in the mean time let this collection of exam- ples be regarded only as a contribution to so desirable a result. ELLIPSIS BEFORE THE RELATIVE. CHAPTER I. ELLIPSIS BEFORE THE RELATIVE. Under this head I shall collect together a variety of instances in which the English language usually requires a fuller mode of expression^ and other languages commonly admit of it. It will thus appear, that in using the word ellipsis, I do so less with a view to insist upon any theory about the origin of such modes of expression, than to adopt a convenient term for classification. It is possible that, in many instances, the so-called elliptical form of speaking may have preceded the customary use of the plena locutio, or may, to say the least, have been in much more general employ- ment. The examples here brought together are collected merely in the rough, some being of more and some of less frequent occur- rence, and others again being borrowed from the practice of a language different from that in which they are quoted. The careful perusal of a body of examples such as these will not be unattended with advantage to the student, as they are of exceed- ingly frequent occurrence in the classical writers, and are in many instances difficult of translation to the inexperienced scholar. But in another point of view they are still more useful, as shedding light upon the origin and use of the many relative adverbial forms which everywhere abound, and of which we shall have to speak more at large elsewhere. Such cases are very usual where the pronoun is may be supplied. Elliptical Nominative. § 1. — This happens where is, in Latin, or the corresponding pronoun in Greek, would form the principal subject : — scspe velut qui curi'ebat fugiens host em; perscepe velut qui Junonis sacra ferre t, Horat. Sat. i. 3, 9, ' he often ran as he would run who was flying from an enemy, and often walked at a funereal pace :' cui placet obliviscitur, cui dolet me mi nit, Cicer. pro Muren. 20, 42, ^he to whom it pleases forgets, and he who is annoyed remembers:^ sed super ent quibus, Virg. -^n. v. 195, 'let them to whom,^ &c. φόνιος εκ θεών, ο ς τάδ' r\v 6 πράζας, Eurip. Phoen. 1045, 'was he Avho effected these things.^ Β 2 4 ELLIPTICAL NOMINATIVE. [CH. I. Ohs. 1. In the first of these examples is may be regarded as wanting, with the verb belonging to it assumed from that in the relative clause, for the relative is in the first case the subject of currebat, and in the second of ferret. In the second and third instances the relative is no longer subjective, but objective. The Greek example is otherwise peculiar, by reason of the expression ήν 6 ττράξας for the simple tense of the verb. 2. On the other hand, the neuter form of the pronoun id is very often found where it appears to be redundant, or is to be regarded as emphatic : intelligo te id quod omnes fortes ac honi viri facere debent, ^c. Cic. Epist. ad Divers, v. 19: in tanta tamque corrupta civitate, Catilina id quod factu facillimum erat, <^c. Sail, in Cat. c. 14: dicere id quod res habet ; in luctu mortem, 8fc. ibid. Cat. 51 : ceterum id quod non timebant per dolum ac proditionem prope libertas amissa est, Liv. ii. 3 : id quod in civili causa hodierno die primum videmus, unum atque idem sentientem, Cic. Cat. iv. 9 : but in all these and similar instances the relative is made to precede the clause to which it refers, which is often, as in the last example, a condition of marked and emphatic importance. Sometimes when there are two members, id is omitted with the first, whilst illud is inserted with the second; as, imperium gravius quod fit quam illud quod amicitia adjungitur, Ter. Adelph. i. 1, 41. 3. It was noticed above that we are not, in examples where is is want- ing, to infer that this was in all cases the intended construction. The following case seems to prove the contrary ; ad quas non est facile inventu qui descendat, Cic. de Amic. 17. Here if we suppose is, we require facilis, as the construction non est facile inventu eum would hardly be allowable. It is obvious that the impersonal est facile inventu is here followed by qui descendat as a direct object. 4. Examples Of the ellipsis of the principal subject are much rarer when the relative is in one of the oblique cases, as in the following : δ u γαρ άττέστειΧεν 6 θεός, τα ρήματα τοϋ θεοϋ λαλεί, John, c. iii. v. 34 : δ ν ψίλεΊς, ασθενεΊ, ibid. xi. 3: turn erupit e senatu, triumphans gaudio, quern omnino vivum illinc exire non oportuerat, Cic. pro Muren. 25 : but this last instance is perhaps not in point, while the subject of erupit is well known, and the relative clause only contains an emphatic addition not necessary to the sense — * a man such that he ought never to have left it alive.' When, however, the elliptical word is only the object of the preceding verb, the relative is frequently in one of the oblique cases, as : υ ν 6 Ώατηρ ηγίασε, καϊ άττέστειλεν εΙς τον κόσμον, νμεΊ,ς λέγετε, John χ. 36. Elliptical Genitive, § 2. — Sometimes the genitive ejus, or in Greek αυτού, or some other genitive, is vi^anting. In the following example there is an ellipsis of sententia ejus, unless any one prefers understanding aliquis respondent : — minus supplicii, quam App, Herdonius, meruit ? plus hercule aliquanto, qui vere rem mstimare velit, Liv. iii. 19, ^ in the view of him who wishes to estimate the thing aright :' ΜΙνως yap παλαίτατος^ ώ ν άκο^ ίσμεν, Thucyd. i. 4 : ώς οντε ώ ν, πννθά- νονταί ατταζ,ίονντων το ίργον, οί ς τ επίμεΧίς ε'ΐΎ] ύ^ίναι, ουκ ονει^ίζόντων, ibid. i. 5, ^as neither despising the employment of § 2.] ELLIPTICAL GENITIVE. 5 those whom they interrogate, nor those to whom it ought to be specially a matter of cognizance reproaching them for it :' ου ^ηθ'' ος ην yap μοι, συ προΰθηκας σττοδον, Soph. Electr. 1179, ' for you have brought the ashes ο f h i m who was left to me :' avavdpia γαρ, το ττλέον όστις αποΧίσας τοΰΧασσον ίΧαβε, Eurip. Phoen. 519, ^it is pusillanimity, on the part of one who has lost the greater portion, to put up with the less :' δώ^ο' άτΓοαιρύσθαί, όστις σεθει» άντίον ε'/ττ^/? Hom. II. α, 230, ^of him who may thwart you :' hie, qui forte velint rapido conten- dere cursu, invitat pretiis animos, et prcemia ponit, Virg. ^n. V. 291, 'of those who :^ καΧ των Θανόντων, \ώπ 6 σο ι Ζ,ωμ&ν, πίρι, Soph. Antig. 214, 'both with respect to the dead and as many of us as live:' ύς οφιν ηκεις ονπερ εξτ/ύχου πάλαι, ^sch. Choeph. 204 : ουκ ίτ epyov εγκαθευδείν 6 στ ι ς ίστ ελεύ- θερος, Aristoph. Lysist. 615, 'it is not the business of him who is free :' κα\ μην παρησω γ' ούδεν, ώς 6pyης εχω, α ττ ε ρ ζυνίημι, Soph. CEdip. Tyr. 340 : δίος '/σχετε μηΒΐν δ σ' αύδώ, CEdip. Colon. 219: κaμoιyε που ταΰτ ίστίν, ο'ίτινΈς βάθρων εκ τώνδε μ' εξάρα ντες είτ' ελαύι/ετε, ibid. 259, Όη the part ofyou who:' τροφας κη^ομίνους, a φ^ ων τε βλάστωσίν, ά ^' ών τ ονησιν £υρωσι, Soph. Elect. 1042, ' for the provision of their parents :' non opis est nostrce. Dido, nee quidquid uhique est gentis Dar- danice, Virg. ^n. i. 601, ' it is not in our power, nor in the power of any thing that remains of the Dardanian race:' en dewtra fidesque quern secum patrios port are Penates, Virg. ^n. iv. 599 : οιόπΈρ και την τύχην, 6 σ a αν παρά λογον ζ,υμβ^, εΐώθαμεν αΐτιασ- Θαι, Thucyd. i. 140, 'on account of those things which.' Obs. 1. Many of the examples that might be quoted under this head, however, require to be distinguished. Take, for instance, the following : την ^ΰναμιν εφ' ο ν ς, Xenoph. Anab. v. 1, 8 ; and ονΒεν αν ε^ει ώ u μέλλω λέγειν^ ibid. ν. 1, 10: συ νυν άφεϊς σεαυτυν, ών λέγεις ττέρι, Soph. CEdip. Tyr. 680. In the first case there is actually an ellipsis of τούτων, dependent upon ^υναμιν ; in the two which follow, the relative clauses are more strictly objective, and supply the place of the ante- cedent or demonstrative. The first is the most rare. In Greek, from the tendency which the relative has to conform itself to the structure of the antecedent, it appears often as though it really filled its place ; and we cannot resort to the supposition of ellipsis, without, in many cases, further supposing an attraction to have taken place. In other examples it is difficult to determine whether the relative is under the influence of the verb in its own or in the antecedent clause. 2. In other passages we may explain the construction in several ways ; i. e. the explanation may be sought under this or some other rule. Thus, si chartce sileant quod bene feceris mercedem tuleris, Hor. Od. iv. 8, 21. It is here difficult to say whether we are to understand de eo after sileant, or ejus after mercedem ; and perhaps it is sufficient to suppose that the 6 ELLIPTICAL GENITIVE. [CH. I. construction is sileant id, which would be Latin. Again, όπως μη εζαρνος εσει α νυν λέγεις, Plat. Euthyd. 283, C. where the ellipsis is perhaps an ellipsis of the accusative, which is the construction of ε^αρνος in Plat. Lysis. 205, A. καΐ το εράν εξαρνος ει. 3. In other cases, in place of the elliptic genitive and relative, the Greeks change the construction, as, a. ν τ l ττράξι^ς καλόν μετά ττό^ον, δ μεν πόνος ο'ιγεταί, το ^ε καλόν μένει' αν τι ποι//σϊ]ς αίσ-χρον μεθ' ηδονής, το μεν ηΕν όΊχ^εται, το ^έ αίσχ^ρον μένει' or in such instances we may say that both the genitive and the relative clause are elliptical, and are left to be inferred from the conditional clause, or the latter may be regarded as their substitute. 4. Besides the examples cited, there are others in which we may pro- perly have recourse to another method of explanation : thus, νίφαι καθαρμω τη vl• ε την στ έγη ν, 6 σ α κενθει, CEdip. Tyr. 1204, is an instance of the well-known Greek construction, by which the subject of the relative clause is removed from its own member, and made the object of the preceding verb, for νΊφαι καθαρμω οσα ijh η στέγη κενϋει. Ι say removed, not so much to indicate my belief that this construction ori- ginated in this way, but only in deference to a popular method of con- ceiving and stating this. But we may also understand τούτων before οσα, 5. There are again many apparent cases of ellipsis occurring before όστις, where either the genitive, dative, or accusative may be supplied. In some of these όστις has a partitive force, and, like the Latin quisque, is used in the singular with reference to a plural noun. But from a colla- tion of passages, it appears rather as though the clause with 'όστις is fre- quently added in some of these cases per epexegesin, or by way of eluci- dation. Thus we have επειτ "ίσως αν δεσποτών ωμών φρένας τυχοιμ' α ν, όστις αργυρού μ ώνησεται, Eurip. Hecub. 356, 'next perchance it would be my lot to obtain a cruel-hearted master, whoever should chance to purchase me.' If Είσττοτών is here merely a plural put for singular, then the construction is regular enough ; but this will hardly be allowed, as the use of the plural depends rather on όστις. In a somewhat different way, ήτις, in the following, may be regarded as a partitive with πόλεων: είρηται γαρ εν αυταΐς, των 'ΈJλληvί^ωv πόλεων ήτις μη^αμοϋ ί,υμμαγ/ί, εξεΤναι, Thucyd. i. 35 ; or we may regard a dative as elliptical τινί, though it is certainly not usual to express it in such cases. In fact, a passage occurring just beloAV, in the same author, seems to prove both these points, viz. that the τις is only expressed in the relative clause, but that the dative is the implied construction : ει γαρ είρηται kv ταΊς σπον- ^αϊς, εζεΊναι παρ* οποτέρους τις των άγραφων πόλεων βούλεται έλθέΐν, ου το~ις επ\ βλάβγι ετέρων ιοϊισιν η ζννθηκη εστίν, αλλ' όστις, και όστις, Thucyd. i. 40. Here Ιονσιν is the dative, and as όστις has reference to each one collectively of a class which the Greeks likewise often express by a plural, there is nothing here but the ordinary construc- tion, while the dative implied with εξε~ιναι has for its substitute τις in the next clause. We may thus explain the syntax of αλλά τιμή τοίσι χρησ- τούς, 'όστις ευ λογίζεται, Aristoph. Equit. 1272. Elliptical Dative. § 3. — In the folloΛving examples a dative may be supplied : — καί τον Άχίλλεύς Οηκίν αίθΧιον ου hapoio, όστις ελαφρότατος § 3.] ELLIPTICAL DATIVE. 7 ΤΓοσσι κραιπνοίσι π^Χοιτο, Horn. 11. ψ'. 749 : ^οκοΰντ Ιμοί, Sokovvtu δ' δ ς κραίνΕί στρατού, Soph. Ajax^ 1023. Antigone 35 : also, yviOTov δε και ο ς μήλα νηττιός εστίν, Horn. II. η\ 401, ^for him whoever might be swiftest of foot' — Ho him who rules the host' — ^to a man of the weakest understanding:' κακόν εσσεταί, ο c κε λίπηται, Hom. II. /. 235, ^for him who may be left:' δσσα ^ι^ονσιν, οτ ις σφ" αλίτηται ομόσσας, ibid. /. 265, ^ t ο him who may swear falsely:' δόθ' τίτις Ιστί, Soph. Electr. 1102, ^to her whoever she is :' ητοί μ\ν τον νεκρον εττίτράττεθ', οΐπ ε ρ άριστοι αμψ^ αντω βεβάμεν, Horn. II. ρ\ 509, '^to those who are bravest and best:' vivitur parvo bene, cut paternum splendet, Hor. Od. ii. 16, ]3: jti77 γαρ αϊδε δαίμονες θεΐεν μ αφωνον τησ^ε της αράς ετι, δ ς μ, ώ κάκιστε, κ.τ.λ. Soph. (Edip. Col. 862, 'against you who:' όστις ημών επ\ TVpavvidi ώδ' εστάλης, Aristoph. Vesp. 503, *^from thee who hast been deputed to tyrannize over us:' viden' tu? ancillas, aurum, vest em, (sub. Antiphilie,) quam ego cum una ancillula hie reliqui, unde esse censes ? Ter. Heaut. ii. 3, 11. Ohs. 1. As remarked in the preceding paragraph, it is not always easy to determine what the ellipsis is ; and in other cases the relative clause actually fills the place of the demonstrative, if it does not take its construction, as in the quotation from Horace, where cui is used in pre- ference to cujus, in order to mark the construction due to the impersonal vivitur. In the annexed example the elliptical word may be made dative or genitive with ττερ/ : καΙ σψων δ νυν έπαινος, ον κομίζετον τοΰδ' ανΕρος, οΤς πονε'ιτον, GEdip. Col. 1406, 'the praise which you two obtain fi-om this man, for the toils which you endure on his account;' where notice also the removal of the subject out of the relative clause, and its being put as object (here latter of two substantives) to the antecedent, in accordance with the analogy of an extensive class of instances. Several curious examples occur in Homer with the relative adjective οίος, in which the same difficulty occurs; thus, ώμ^ άποταμνόμενον κρεα ε^μεναι, οΐά μ εοργας, Hom. χ'. 347, ' to eat thy uncooked flesh in revenge for such evils as thou hast done tome:' ώκΰμορος οη μοι, τέκος, εσσεαι, oV αγο- ρεύεις, ibid, σ'. 95, 'by reason of the things which you state.' The other passage quoted by Trollope as belonging to the same class, oJov εμόν γε μένος nal χε~ιρες άαπτοι, ibid. θ'. 450, is, I think, different, and contains a common construction of oToc, and its Latin representative, qualis, which is similarly used, as we shall elsewhere see. It is true, some would regard them all as adverbs here, and translate, ' so basely have you treated me ' — ' so rashly do you talk ' — ' such is my invincible prowess ;' but even then it is necessary to account for the origin of such modes of speech, many so-called adverbial forms being only remnants of declinable ones, which have in progress of time become fixed and per- manent. 2. In some examples we may resort to different suppositions, while the same words admit equally of different constructions. Some impersonals are followed either by the dative, or the accusative and infinitive. Thus, operce pretium est audir'e, qui omnia jprce divitiis humaiia spernunt, neque 8 ELLIPTICAL ACCUSATIVE. [CH. I. honori magno locum, neque virtuti putant esse, nisi, ^c. Liv. iii. 26, ' it is worth while for those who despise to hear.' The mstSince paulo qui est tolerabiUs, Ter. Heaut. i. 2, 31, is of doubtful class, the commentators hesitating whether to understand ei or eum, or whether qui is the sin- gular put for the plural, which is certainly the preferable supposition. In this last case the construction is exactly that of the Greek 'όστις, men- tioned § 2. obs. 5. Elliptical Accusative, § 4. — An accusative is wanting in the examples which are sub- joined : — et qua vectus Abas et qua grandcevus Alethes vicit hyems, Virg. /En. i. 125 : integras vero tenere possessiones, qui se debere fateantur, Caes. B. C. iii. 20: non sibi defuisse, cui nupta diceretur, Liv. i. 47^ ^ that a husband was not wanting to her :' defuisse, qui se regno dignum putaret, qui meminisset se esse Prisci Tarquinii filium ; qui habere quam sperare regnum mallet, ibid. ^ that there was wanting, a m a η who felt himself worthy of empire, who remembered his illustrious parentage, and who preferred the possession to the hope of power :^ censendo enim, quoscunque magistratus esse, qui senatum haberent, Liv. iii. 40, ^ for by making this vote or proposition, he expressed his opinion, that those who held the senate were magistrates of some kind or other :^ — dignos enim esse, qui annis cepissent, Liv. iv. 49, *^that those were worthy who had taken by arms,^ &c. : si quidem licere dicimus quod legibus, quod more majorum insti- tutisque conceditur, Cic. Or. Philipp. xiii. 6, ^if we call that lawful which is permitted by the laws, by the custom of our ancestors, by our institutes :' profecta deinde cohors ad sepeliendos qui ceciderant, Liv. iii. 43, ^ t h ο s e who had fallen :' mirari se non sacrilegorum numero haberi, qui supplicibus eorum nocuissent, aut non gravioribus poenis affici, qui religionem minuerent, quam qui fana spoliarent, Corn. Nep. in Ages. 4, ^that he wondered that those who injured their suppliants were not reckoned among the sacrilegious, or that those who treated religion disdain- fully were not affected with heavier punishments than those who despoiled sanctuaries :' ac satis habere coegit, si liceret Africam obtinere, qui jam complures annos possessionem Sicilice tenebant, Nep. in Timol. 2, ^ and compelled those who for many years past continued to hold possession of Sicily, to think them- selves fortunate if they got safe to Africa (their own shores) :^ minime mirror qui insanire occipiunt ex injuria, Ter. Adelph. ii. 1, 43. The practice of the Greeks, in this particular, is exactly similar, but is carried much further. A few examples are here subjoined. See also Matth. Gr. Gr. § 293. YJbgoq l\ ϊχων § 5.] ELLIPTICAL ACCUSATIVE. 9 ους εΤτΓον, Xenoph. Anab. i. 2, 'Cyrus having those whom I mentioned:^ ^ουΧωσάμενοι μίν ω κακία "φνχης ενεγίγνετο, ελευθε- ρώσαντες δε φ αρετή, Plat. Phaedr. 256, Β^ 'that part of the soul in which the baser qualities were^ and that in which virtue was situated :' αλλ' avayKaaai θεούς αν μη Θύλωσιν, Soph. CEdip. Tyr. 375, ' for a man to compel the gods to do the things which they are determined not to do is impracticable:^ κα\ εποιησε σοφον 6 σ a δίδασκάλων αγαθών είχ^ετο, Plat. Men. 93, D, ' and had him well instructed, so far as instruction was to be had from good teachers :' ουκονν και ούτος τον νίον τον αντου Αυσί- μαχον, δ σ α μεν ^ί^ασκαΧων εί-χ^ετο, κάΧλίστα ^Αθηναίων επαί^ενσεν, ibid. 94, Α. In both these examples, we may understand an accu- sative after a verb of teaching. Again, κα\ a μεν αν επινοησαντες μη εττεζύΧθωσίν, οΙκείων στερεσθαι ηγουνταί, Thucyd. i. 70, 'they deem themselves deprived of their own property, where they do not succeed in realizing those things which they have once designed to obtain.' Obs. 1. In this last and similar examples, the elliptical word appears to be the Greek accusative absolute, 'as to those things.' But different persons will regard the construction differently. Some would call it anacoluthon, some ellipsis, and others again neither of these. We may rather say, perhaps, that the whole relative clause occupies the place of the accusativus prcepositivus so common in Greek. 2. Many other cases of this ellipsis might be quoted, as the following: — ττροαγορενόντωρ ημών α εμέλΧομεν νττο 'Αθηναίων βλάπτεσθαι, Thucyd. i. 68 : but here an attraction of the relative into the case of the antecedent takes place, which examples I have reserved for another place. Again, in the use of εί τις in Greek, and si quis in Latin, an accusative is often wanting, which may be conceived of as transferred into the relative clause, and this usage belongs to a different class. Thus, fca<, μ civr tg εί τις ε'ίη παραγγείλας τταρεΊναί, Xenoph. Anab. vi. 4, 15, 'bidding each who was a priest to be present:' lex erat Thehis qiice morte multabat, si quis imperium diutius retinuisset, Nep. in Epam. 7. It is not, however, merely an accusative that is wanting, but the relative clause stands in other con- nexions. Thus, s i mala condiderit in quern quis carmina, jus est judi- ciumque, Hor. Sat. ii. 1, 82, where we may substitute in eum qui for si quis. This suppressed sense is not uncommon before the conditional clause introduced by si in other cases ; as, bacchatur vates magnum s i pectore ρ ο s sit excussisse Deum, Virg. ^n. vi. 78, 'with a view to do so.' 3. In the subjoined example, vetabo qui Cereris sacrum vulgarit arcance, sub isdem sit trabibus, fragilemve mecum s olv at phaselum, Hor. Od. iii. 2, 26, the construction may be vetabo eum ut solvat, or vetabo ut is solvat, the ut in this construction being also frequently omitted. See my Accus. and Iniin. § 4. art. 3. Elliptical Ablative. § 5. — In the instances I am now going to quote, an ablative is wanting, or what is the same thing, the whole relative clause 10 ELLIPTICAL ABLATIVE. [CH. I. stands for the ablative, in the structure of the ablative absolute : — parto, quod av eb as, Hor. Sat. i. 1, 94^ ^ what you desired having been obtained:^ novos ritus sacrificandi vaticinando, i nfe r entibus in domos, quibus qumstui sunt capti superstitione aniini, Liv. iv. 30, 'those to whom the superstition of others is a matter of gain, introducing new rites of sacrifice, by assuming the prophetic character :' auctoribus qui aderantut sequeretur, Liv.iii.44, 'those who were present suggesting:^ insuper increpante qui vulneraver at, Liv. iv. 50, 'he w^ h ο had wounded him a d d i ng thereto reproach :^ qui Atheniensium rebus studuissent ej ectis, Nep. in Lysand. 1, 'the partizans of the Athenians having been banished:' emptis per commercia, quorum habitus et crines in captivorum speciem for- marentur, Tac. Agric. 39, 'persons whose dress, &c. had been made to represent that of captives, having been bought in the way of trade :' remissis qui in prcesidio erant, Sail. Bell. Jug. 58: revocantibus qui rescindebant, Liv. ii. 10: except ο quod non simul esses, ccetera Icetus, Hor. Ep. i. 10, 50, ' in every respect happy, save in your absence:' sed mihi opus erat, ut aperte tibi nunc fabuler, aliquantulmn quce affe r r e t, qui dissolverem qu(B debeo, Ter. Phorm. iv. 3, 50, ' but that, to talk openly with you, I w^anted a wife who ΛνοηΜ bring a little ready money, to enable me to pay my debts.' Ohs. The last example, it will be observed, differs from those which precede it. All the rest are cases of the construction of the ' ablative absolute,' which appear to be of the most frequent occurrence. Thelast quoted is one of the ablative after opus, and belongs therefore to another rule in syntax. Elliptical Preposition, ^c. § 6. — Ellipsis occurs with other words besides the demonstrative pronoun, and it is sometimes necessary to supply a preposition in addition. It thus happens that a Λ^erb will sometimes appear to take as an object one which does not properly belong to it ; as stertimus, indomitum quod despumare Falernmn subficiat, Pers. Sat. iii. 3, ' Ave snore long enough to dissipate the fumes and ferment of strong Falernian,' where we must understand per tempus : again, palleat infelix, quod proxima nesciat uxor ? ibid. iii. 43, ' and unhappy should grow pale with terror, on account of that W'hich the wife of his bosom may not know,' where we must understand propter id: veruntamen illis imperatoribus laus est tribuenda quod egerunt; venia danda quod relique- r u η t, Cic. pro leg. Manil. 3, where supply de eo : dJ\άίζeτaι δε § 9.] ELLIPTICAL PREPOSITION, &C. 11 και μονσίκας Iv αώτω οία τταίζομίν, Find. Olymp. α'. 14, ^ie is adorned with tlie choicest flowers of music, in such strains as we sing/ where Dissen also quotes Find. Fragm. (103.) and Donaldson quotes Fyth. iii. 17. § 7. — With relative adverbs, in like manner, it is often necessary to supply various ellipses mentally ; thus, όμως δ' Ίν '^σταμεν Xpdac^ αμεινον Ικμαθύν τι Βραστίον, CEdip. Tyr. 1408, ^ i η the position in which we stand, it is better ;' but Matth. Gr. Gr. § 480, 3, in explaining this passage, appears to me not quite correct, who renders by ' because we are in this state,^ regarding ίνα as put for otl Ινταυθα : cedibus ex magnis subito se conderet, unde mundior exiret vix libertinus honeste, Hor. Sat. ii. 7, 11. § 8. — This ellipsis occurs in particular phrases; as, paucis diebus, quibus eo ventum erat, Caes. B. G. iii. 23, for post eos dies quibus, and is preserved in postquam, priusquam, &c. the indeclinable or adverbial forms which have probably arisen out of similar consi- derations. To these may be added particular phrases, as dignus qui, idoneus qui. General Remarks, § 9. — [a) This ellipsis of the pronoun is very common after cognosco, mitto, dico, and some others ; as, quce ignorabant de L. Domitii fuga cognoscunt, Caes. B. C. i. 20, 'they learn what they were previously ignorant of:^ Pausanias dux ^partes, quos Byzantii ceperat, tibi mi sit, Nep. in Faus. 2, 'those whom he had captured:^ qum videbantur dicebat, Nep. in Timol. 4. {b) And likewise after scio, intelligo, expedio, ostendo, demonstro, &c. as, proeliumque commisit, quam milites sui scirent cum quibus arma conferrent, Nep. in Eum. 3 : ostendam lierum quid sit pericU fallere, Ter. Andr. v. 2, 26, ' 1^11 teach you what risk you run in deceiving your master.' Ohs. 1. The instances, however, of this latter kind (ό) are hardly to the point ; for although our English idiom fills up the supposed ellipsis with the demonstrative, it does not therefore follow that it is proper to do so in Latin. Indeed, in these cases, the antecedent generally finds a place in the relative clause, or quis, quid follow, which are strictly equivalent to such a combination of antecedent and relative, marking, as they do, a strictly objective sentence. And the Latins likewise distinguish this usage, by placing the verb in the relative clause in the subjunctive mood, which is another mark of the objectiΛ'e sentence, and proves that it differs from the species of ellipsis which we have before been discussing. This point will be more fully elucidated when we come to speak of the influence of the relative on the mood of the verb. The same remark applies to quid when it follows a verbal noun ; thus, conscientia quid ahesset vii-ium, Liv. iii. 60, ' from a consciousness of the deficiency of strength,' where we do not resort to the supposition of ellipsis. 12 REPETITION OF THE ANTECEDENT. [CH. IT. 2. To the examples last named, § 9. (b) and also Obs. 1, belong a numerous class of instances, in which the neuter pronoun id is commonly regarded as elliptical ; not but that the practice of the Latin writers would lead us to suppose, that it was much more frequently absent than present ; as, qtiid ea postulet pervidendum, Cic. de Finib. v. 16 : deinceps videndum est quce sit liominis natura, ibid. v. 12 : quid sit primum est videndum, ibid. Tusc. Qusest. i. 9. It is here at once obvious that the relative clause may be regarded as the nominative to the impersonal verb, or id may be supposed to be wanting. But, on examination, instances of this sort will be found to be similar to those just named above. The relative clause is, in reality, objective to the impersonal verb, as is proved by the use of the subjunctive mood in cases of this kind, and by the analogy of the phrases, utendum est opere, fruendum est bono. If id is supplied, therefore, it does not stand as an antecedent, but in mere apposition, and is to be regarded as objective to the foregoing verb. In English, when a relative sentence is the subject of an impersonal verb, the 'it' is always expressed, when the verb comes first ; but if the relative sentence is placed first, there is no need of the impersonal symbol of the subject. Thus we say, either, ' it makes no difference whether you stay or go,' or ' whether you go or* stay makes no difference.' § 10. — The subject of ellipsis as containing the explanation of the use of a great number of relative particles and adverbs will be afterwards resumed, when we have first explained some of the peculiarities of the relative pronoun well deserving of minute attention. CHAPTER II. REPETITION OF THE ANTECEDENT. § 11. — Having thus treated, at some length, the subject of ellipsis, I pass on to notice some instances in which the antecedent is repeated in the relative clause, and may be said to be redundant. These instances, though not very common, are easily understood, and require no particular comment. Perhaps C£esar most abounds in this practice; and, indeed, it is easy to see that it suits the accuracy of historical detail, much better than it consists with the concise elegance of mere literary composition. Thus, erant omnino itinera duo quibus itineribus, Caes. B. G. i. 6 : diem dicunt qua die, ibid.: ut semel Gallorum copias proelio vicerit, quod ρ roe Hum, ibid. i. 31 : in castris Helvetiorum tabulce repertce sunt, quibus in tabulis, ibid. i. 29 : insulce de quibus insults, ibid. v. 13 : in ea parte, quam in ρ art em, ibid. iv. 32 : ού yap άνηρ ττρότταν ημαρ ες -ηίλίον κατα^ύντα ακμηνος σίτοίο Βυνησεται αντα μάχεσθαι .... ^ 12.] REPETITION OF THE ANTECEDENT. 13 ος δε k' άνηρ κ.τ.Χ. Horn. II. τ, 162: ουκ οΐσθα, σον πατρός μίν ος πρυνφν ττατηρ, Soph. Ajax^ 1263. And sometimes, in addition to the repeated noun, another adjective is inserted ; as, αποπέμπομαι ίνννχον ο'φιν a ν περί παι^ος Ιμον, του σωζομίνου κατά θρ^κην δι' ονείρων εΤδον φοββράν 6 -φ ι ν εμαθον, ε^άην, Eurip. Hecub. 72. § 12. — This species of repetition ^ is, perhaps, much more usual where a word of equivalent meaning to that of the antecedent is added to the relative clause ; as, locus, iter, res, &c. : as, quern si fata virum servant, Virg. JEn. i. 546: igitur ad Catabathmon; qui locus, &c. Sail. B. J. 22 : deinde Philenon Arce, quern locum, &c. ibid. : helium scripturus sum quod populus Romanus gessit, quce contentio, &c. ibid. 5 : hue accedebat munificentia animi et ingenii solertia, quibus rebus, ibid. 7 : eorumque clientes semel atque iterum armis contendisse, &c. Cses. B. G. i. 31 : qui non longe a Tolosatium finibus absunt, qum civ it as, ibid. i. 10: idus tibi sunt agendce, qui dies, Hor. Od. iv. 11, 15 : quod si hominibus honarum rerum tanta cur a esset, quanto studio aliena petunt, Sail. B. J. 1 : proximus lictor, quern ministru m, ibid. 12 : propter siccitates palu- dum, quo perfugio usi, Caes. B. G. i. 38: igitur Vaccenses quo Metellus, Sail. B. J. 70 : legatos, quod 7iomen apud omnes sanc- tum, in vinculo conjectos, Cses. B. G. iii» 9 : nam cum animus, &c. quae virtus est appellata prudentia, Cic. de Leg. i. 23 : ττροδηλον γαρ, otl εξ Ιούδα άνατίταΧκεν 6 Κύριος ημών, εις η ν φυΧην.,,, κ.τ.Χ. Heb. vii. 14. And sometimes the substantive in the relative clause is one of equivalent signification to the whole proposition that has preceded it ; as, legatos ad eum mittunt, cujus leg ationis, Caes. B. G. i. 7: atque hmc causa conjungit, quam conjunctionem, &c. Cic. Cat. iv. 7: inde repente impulit ; impulsu quo, Virg. Mn. viii. 239. This is done, moreover, where the noun in the relative clause is elliptical ; as, Numidce insidiati sunt ei, quas non solum effugit, Nep. in Hannib. 6, where insidias is mentally supplied. 0^5. 1. In several of these cases the relative is, however, merely put for the demonstrative pronoun and conjunction ; and, as we shall have occasion to see hereafter, may be diiFerently classed. 2. We might have quoted, as coming under the rule, the following : — Nomadumque petam c onnuh ia supplex, quo s ego sum toties jam ^ This is done where a proposition has preceded, and where the relative has reference to a subordinate word in the proposition, though more rarely ; nam C anu si lapidosus : aquce 7ion ditior urna ; qui lo cu s a forti Diomede est conditus oUm, Hor. Sat. i. 5, 91. 14 REPETITION OF THE ANTECEDENT. [CH. II. dedignata mar it ο s, Virg. ΙΕώ. iv. 535; but maritos does more than occupy the place of Nomadum, and explains the pomt of view in which they had been disdained. It would therefore be better to translate, ' whom in the character of suitors for my hand I have so often disdained.' 3. The passage quoted from Sallust above, tanta cura, quanta studio, is also analogous to an extensive class of examples which will come before us presently. In the mean time it may serve to illustrate this paragraph. § 13. — ^Out of the usage described in the last article, § 12, may possibly have originated a class of examples in which the relative pronoun refers to a different antecedent than the word with Λvhich it is made to agree. In other phraseology it is objective to the word with which it agrees as an adjective; as, pei^ idem tempus advorswn Gallos ah ducihus nostris, Q. Ccepione et M. Ma7ilio, male pugnatum. Quo metu Italia omnis contremuerat, Sail. B. J. 114, where it is customary to translate by ' from the fear of which.' Of the same kind are quo conspectu, Eutrop. vi. 20 : quo terrore, Liv. i. 58: tumulus quo summo, Virg. ^n. iii. 22: quo tumultu, Sail. Cat. 43 : quo pavore, Liv. iv. 19. Ohs. Many, or most, of these examples are cases of the initial qui after a pause, and may be treated of under another head. See § 12. Ohs. 1. This construction is not peculiar to the relative, but is a well-known demonstrative usage; as, atque hoc metu latins vagari 'prohihehat, Cses. B. G. V. 19: h a c ira consules in Volscum agrum legiones duxere, Liv. ii. 22, ' with anger on this account :' hoc dolore, Cees. B. G. v. 4. § 14. — When the antecedent refers to a class, rather than to specified individuals, the relative pronoun often takes the word genus in its own clause ; as, additis perfugis quod genus, Sail. B. J. 56: essedariis quo genere, Caes. B. G. iv. 24: quod cap- tioso genere mterrogationis utuntur, quod genus, &c. (Incert.) § 15. — When the verb sum appears by itself in the relative clause, then this second antecedent appears as a predicate; as Carcasone et Narbone, quce sunt civitates Gallice p7Ovinci(e, Caes. B. G. iii. 20 : ea qu(S secuta est hieme, qui fuit annus Cneio Po7n- peio, Marco Crasso, Coss. ibid. iv. 1, ^ which was the year of the consulate of Pompey and Crassus.' The same observation may be extended to passive verbs of calling or naming. Ohs. This last instance is otherwise remarkable, from the substitution of the ablative absolute, in apposition with annus in the nominative ; but is easily explained like other idioms, in which the mode of denoting a thing is put for the thing itself, from the frequency of the expression. A similar instance in Greek is the following : — f ν ημερ^- εττίλαβομένου μου της χειρός αυτών, Heb. viii. 9. The names of the consuls are else- where put for the year, usque ad P. Sulpicium et C. Aurelium consules, Corn. Nep. Hannib. 7. See also Liv. iv. 30, annum C. Servilium Ahalam, L. P. Mugillanum consides. 16.] ELT^IPTICAL RELATIVE. 15 CHAPTER III. THE RELATIVE ITSELF ELLIPTICAL^ AND THE USE OF THE DIRECT FOR THE RELATIVE STRUCTURE. Elliptical Relative, § 16. — Cases of this sort are not very frequent, and occur mostly in familiar idioms ; as, velut ova venarum^ αίμορροί^ας GruBci appel- lant, Celsus ii. : verrucarum qucBclam genera dolentia ακροχορ^όνας GrcBci appellant, ibid. : tres Notus abreptas in saxa latentia torqiiet. Saxa vocant Itali, mediisqiie in fluctibus Aras, Virg. ^n. i. 109 : in Lemelios Ligiirice pars est, Tac. Agric. 7 : qui mercenarius agrum ilium ipsmn mercatus aravit, Hor. Sat. ii. 6, 12, ^ who purchased the very land which as a mercenary he ploughed \' urbs antiqua fuit, Tyrii tenuere coloni, Virg. ^n. i. 12 : est locus Hesperiam Graii cognomine dicunt, ibid. i. 530: terra procul Thraces arant, Virg. ^n. iii. 13. Direct for Relative Structure. This is seen in the following : — Maharbale Himilconis filio, eum prcefecerat Hannibal, if a impigre rem agente, Liv. xxi. 12: is eo cum eocercitu, ibid. iii. \: is enim cum Sp. Postumio consul fait, ibid, iii. 2: id remedium timorifuit, ibid. iii. 3 : adeo civitates, &c. ibid, iii. 4: it a tum, &c. ibid. : f rater idem, ibid. iii. 5. Ohs. 1. In the last examples the demonstrative pronoun or adverb is introduced, which proΛ^es that the structure was not intended to be relative. The editors often include sentences of this sort within brackets ; but this appears to me unnecessary, and even to be productive of an erroneous view of the case. It is certainly better to regard this usage as a particular construction or peculiarity of style, than to resort to the mechanical process of walling up a whole proposition between barriers, as if it were prohibited. The writers who employ it evidently do so for the sake of an agreeable diversification, and to avoid the monotony of the relative construction. 2. In other cases the ellipsis is more than one of the relative. In the following we must understand eo quod, and the abrupt omission is very marked ; an haccis opulentet oUvce, pomisne, an pratis, an ami ct a V itibus ulmo, Hor. Epist. i. 16, 2. 3. Sometimes, in Latin, as in Greek, the participle stands in lieu of the relative construction, and expresses the cause of a particular condition ; as, felix una ante alias jussa mori, Virg. Mn. iii. 321, 'happy in being required to die.' 16 ELLIPTICAL RELATIVE. [CH. III. § 17. — The relative is often elliptical in English. Besides the many cases in which the conjunction ^that/ answering to the Latin relative ut, is left out^ we have such as these : — ^ I have two boys, seek Percy and thyself about the field/ Shakesp. I. Henry IV. act V. sc. 4 : 'in the fair rescue thou hast brought to me/ ibid. : ' than those proud titles thou hast won of me,' ibid. : ' there's some great matter sheM employ me in/ Two Gent. Veron. act iv. sc. 3 : ' she loved me well, delivered it to me/ ibid, act iv. sc. 4. § 18. — In certain customary phrases, in appellations, and in particulars of admeasurements, such as content, height, distance, breadth, &c. the relative is often not employed; as, άφικνουνται ΙπΙ τον Μάσκαν ποταμον το εύρος πΧεθρίάίον, Xenoph. Anab. i. 5, 4: ενταύθα ην ποΧις έρημη, μεγάλη, όνομα δε avTy Κορσωτη, ibid. There is likewise a common species of ellipsis in Greek and Latin, which has become so usual that we hardly notice it, in the phrases, ελεζε τάΒε, ελεζε or ελεγεν ταντα, τοιαύτα έλεγε, τοίάδε είπεν, dixit hmc, dixit talia. These phrases refer, as is well known, to some- thing previously narrated, or subsequently to be narrated, and, consequently, the plena locutio would be, in English, ' such as we have related,^ or ' such as we are about to relate / or of a narrative already stated, ' such was the statement / and of one to be given, 'he said as follows.' In this use the demonstrative pronoun is often repeated in Greek; as, τοίάδ' ηκονεν κακά τοίάδε λέγω, Eurip. Hecub. 571 — 575. § 19. — It may be objected, perhaps, to many examples that we might quote from Greek, that they illustrate equally well ellipsis of some other kind, as, for instance, of the participle. But as the participle stands for the relative and finite verb in many instances in Greek, the cases are at least analogous, and in the selection of examples below I shall regard them as such. Thus, ό ηγησόμενος ουδείς εσταί, Xen. Anab. ii. 4, 5, ' there shall be none who will act as guide. ^ The transition from the participle to the direct con- struction is certainly an equivalent change to that which we are considering. Even if we suppose that the Greeks preferred the direct to the relative construction, it is not out of place to intro- duce the examples here, as there is still the same departure from the analogy of our own language, where the relative is in use ; and to the English student of the classics it is most important to dis- tinguish these differences in the practice of the several languages. (a) Transition from participle to finite verb. This is common in the New Testament ; as, δίά την άληθειαν την μεν ου σαν εν ημΧν, κα\ μεθ' ημών ε στ α ι, 2 Ερ. John, i. 2 : το πνεύμα καταβαΐνον και εμεινεν, John's Gosp. i. 32 : τους Xiy οντάς άποστόΧονς § 19.] CHANGE FROM RELATIVE TO DIRECT CONSTRUCTION. 17 ξίναί, και ουκ ε ι σ ι, Rev. η. 2. 9. iii. 7. 9^ and elsewhere. This practice seems to be very characteristic of St. John's writings, or rather he indulges it freely. It is also met with in St. Paul, where it is probable, from the punctuation, that the editors do not under- stand the construction : ε γ ε / ρ α ς αυτόν εκ νεκρών, και '/c ά θ t σ ε ν, Eph. i. 20, where they put the colon after νεκρών. {b) Transition from the relative to the direct construction. The first I may quote is, r ο μεν ου ^ννατ αΧλος ^Αχαιών πάΧλειν, αλλά μ ί ν οίος επίστατο πηΧαι Άχίλλεΰς, Hom. II. π. 142, and again, II. r . 389, ^ w h i c h no other of the Greeks was able to wield, but Achilles alone knew how to wield it.' I of course choose to regard το as the relative, and remark, in passing, that this example is of a different kind from some I shall have to quote presently, the relative and demonstrative being here each in a distinct mem- ber of the whole period, which sometimes are found in one and the same member. Again, όστις ετ αβλητος κα\ άνοντατος οζεϊ χαΧκω δίνεύοί κατά μεσσον, ayoi δε ε Παλλάς ^Αθηνη χειρός ελοΰσ', αυτάρ βελεων άττερύκοι ερωην, Hom. II. δ'. 540, where we have I instead of ov. To this head belong examples in which the struc- ture commenced is not persevered in, but a fresh one is begun where no relative is met with. Thus, oi γε τον φνσαντ ε με ούτως άτίμως ημνναν, αλλ' ανάστατος αντοΐν εττ ε μφΘην, CEdip. Col. 423, where I would put a comma only at ημνναν : ττείσεται γαρ aWo μεν αστερίες ουδέν, γί^ς• δ' άττ εισ ιν άβΧαβτ^ς, CEdip. Tyr. 223 : βαθι βαθι τανδε γαν σοι ν ι ν iKjovoi κτισαν, Eurip. Phcen. 681 : "Έκτορ, τίτττε μάχης άποπανεαι ; ο ύ δ ε τι σε χ ρ -η, Hom. II. π. 721. These last four cases might be consi- dered as containing nothing more than the direct, in lieu of the relative construction, and so far belong to § 16 ; but it will be seen on examination, that they are essentially different ; the first member of the sentence implying a construction which is sud- denly departed from in the second, and which differs from a mere parenthesis. A curious instance is the following, ου καθώς Κάιν εκ του τΓονηροΰ fjv και έσφαζε τον άδελφον αύτον, 1 John iii. 12. This may be regarded as a mere case of the ellipsis of the relative pronoun 6ς after Καϊν, or we may adopt an arrangement, of which cases are not wanting in EngUsh, ^ not of the evil one as Cain was, and killed his brother.' It is absurd to say, that καί has here the force of qui, though this substitution might make out a construc- tion. In such phrases as the following, Stallbaum has noted, that the relative is usually wanting, or, what is perhaps still more correct, that a direct construction takes its place : λέγω jap δί) το c 18 CHANGE FROM RELATIVE TO DIRECT CONSTRUCTION. [CH. III. εναντίον rj 6 ποιητής εποίησεν 6 ττοιησας, Plat. Euthyphr. 12, A : δι' εΰνοιαν τουναντίον είπες η δύναται, ibid. Phsedr. 275, A, ^for I say the opposite of what the poet feigned/ &c. ; ' the opposite of what is its real value.^ Among others, the following good in- stances may be quoted : νφ' η ς εμασχαλίσθη καΐ εζε μαζεν^ Soph. Elect. 437 : ilium ego jure despiciam, qui scit quanta sublimior Atlas omnibus in Lybia sit montibus ; hie tamen idem ignoret, for sed qui in the latter clause, Juven. Sat. xi. 23 : μανθάνειν παρά τούτων, ο t μήτε προσποιούνται ΒιοάσκαΧοι είναι, μητ εστίν αυτών μαθητής μη^ε).ς τούτου του μαθήματος, Plat. Men. 90, Ε ; where see also Stallbaum^s note to the passage, in which he quotes, Gorg. 452, D. De Hepubl. iii. 395, D. Ibid. vi. 505, Ε : οία λέγεται' τε νπο των ποιητών, και ύπο των αβαθών Ύραφεων τά τε άΧλα ιερά Ί]μίν καταπεποίκιΧται, και δ?) και τοις /ζεγαλοις ΐΐαναθηναίοις 6 πεπΧος μεστός τ ώ ν τοιούτων ποικιλματων ανάβεται εΙς την άκρόποΧιν ; Plat. Euthyph. 6, Β & C, where supply οΊΌις implied from οία, and note the change of construction to τοιούτων, instead of ών or οίων, see Stallbaum : και νυν τι \ρη Βράν ', δ στ ι ς εμφανώς θεοΐς εχΘαίρομαι' μισεί δε μϋ '^ΈΧληνων στρατός, Soph. Aj. 449. This example illustrates also the ellipsis of the demonstrative before δστις, and it was, perhaps, with a view to make the real sense clear, that the writer suddenly reverts to the direct structure, and intro- duces με. See also Matth. Gr. Gr. 472, 3. Again, ουχ οράας, οίος και εγώ, καΧός τε με'γας τε ; πατρός δ' είμ ά-^αθοΧο, θεά δε /χ ε ^είνατο μητηρ, Hom. II. φ', 108. The practice we have been exemplifying here is very frequent with αυτός, νίν, μίν ; as, η φύσει μεν f/v βασί- Χεια Ααομε^οντος, εκκριτον δε ν ι ν δώρημα κείνω 'δώκεν ^ ΑΧκμηνης ΎΟνος, Soph. Aj. 1273: ε^' ον άν ιδτ?ς το πνεύμα καταβαΐνον, και μενον επ' αύτον, John i. 33 : ττερι ώ ν Βιενεχθίντες και ου δυνάμενοι έτΓΪ Ικανην κρίσιν αυτών. Plat. Euthyph. 7, D. And to this class belong two examples quoted above, Hom. II. π\ 142. /. 389. In the Greek of the New Testament this practice is carried still further, and the αυτός actually occurs with the relative in the same clause, of which examples are given elsewhere. In this way is explained the construction in John xv. 6, 'If a man abide not in me, he is cast forth as a branch, and is withered ; and men gather them, and cast them into the fire, and they are burned :' ώς το κΧημα, και εΕ,ηράνθη^ και συνάγουσιν αυτά, και εις πυρ βάΧΧουσι, και καίεται. Here the αυτά is once expressed in the second of four clauses, of w^hich two having the verb in the active, and two in the passive, the αυτά serves for both subject and object, the sense being certainly, that the branches are withered, gathered, cast into § 19.] DIRECT FOR RELATIVE CONSTRUCTION. 19 the fire, and burned. Our translators should have supplied the relative here, instead of copying a construction which is almost unintelHgible in English. (c) Direct for relative construction. In addition to instances given above, under the general rule, there are the following : turn virgam capit ; hac animas ille evocat, ilia fretus agit ventos, &c., Virg. ^n. iv. 242, where note that hie and ille are used of the same subject. In English, we have ^ a prince called Hector, Priam is his father,^ Shak. Troilus and Cress, act i. sc. 3 : ^ when that this body did contain a spirit, a kingdom for it was too small a bound ;^ I. Henry IV. act v. sc. 4. In a somewhat similar manner the English translators of the New Testament use ^ him^ in place of the relative ; as, ^ him would Paul have to go forth with him,' and the same is true of the Greek original, where τούτον is employed. Thus also ίνθα, and corresponding words, have in Greek often a relative force; as, ίνθα φόνιος ην δράκων, Eurip. Phoeniss. 657. In instances of this sort, the editors almost in- variably punctuate more fully, as if there was a break in the con- nexion ; as, Virg. JEu. ix. 340, suadet enim, ^c. ; and again, 354, sensit enim, ^c. ; and thus their force, as relatives, is put out of view, to which I object. Eodem modo primus eques hostium agmi- nis fuit, prceerat Aruns Tarquinius, filius regis, rex ipse cum legio- nibus sequebatur, Liv. ii. 6 : indutus at olim Demoleos agebat, Virg. ^n. V. 264. The repetition of the article in Greek with an adverb, in apposition with some preceding noun, is of the same origin ; as, εϊ γά^ο εν ταίς ζνμφοραΐς ταΐς νυν, CEdip. Tyr. 496. This, in other languages, is effected by means of an adjective or a relative clause, ^ the calamities which at present afflict us.' {d) The relative put a second time in place of the conjunction. Thus ος ουκ οθετ α'/συλα ρίζων, ο ς τοΕ,οισιν ίκη^ε θεούς, οι "Ολυ/χ- πον εχουσιν, Hom. II. ε'. 404. {e) The relative elliptical in one part of a sentence, but inferred from a preceding relative. Many examples are met with, where the relative only occurs once in a sentence, and must be supplied again in another case or form by inference. Thus, posteaquam in vulgus militum elatum est, qua arrogantia in coUoquio Ariovistus usus ; omni Gallia Romanis inter dixisset, impetumque /e- cisset, ^c, Caes. B. G. i. 46, where we must supply quomodo, inferred from qua in the former part of the sentence : δ ν ημείς ηθεΧομεν βασιλέα καθιστάναι και εΒώκαμεν και ελάβομεν τηστα, Xenoph. Anab. iii. 2, 5 : ^ whom, and to whom, and from whom :' o) μη ξένων εζεστι μηδ' αστών τίνα όόμοις ^εχεσθαι, μη^ε προσφωνεΐν τίνα, CEdip. Tyr. 790. I had quoted this passage without refer- c 2 20 ONE RELATIVE SERVING FOR TWO CLAUSES. [CH. III. ence to the various readings, before noticing Wander's note in his edition, ' permirum est recentissimos editores revocare potuisse vitiosissimam scripturam a consuetudine loquendi GrcRcorum alienis- simam ω μη, κ.τ.λ. quum rectissime jam Erfurdtius olim, quod solum scribi a poeta pohdt, ov μη restituisset.' Again, Ιξ ων ζζς και ποί^Τς και τΓολίτεύεί, Demosth. de Cor. νη' : but this is hardly, perhaps, to the point ; for the relative having taken the place and construc- tion of the demonstrative, although the two last verbs take the construction εκ τούτων, the a will stand indiiferently for any case ; and thus many instances will be found, in which two verbs, taking a different case, are coupled with one relative, which has assumed the case of the demonstrative, by the ordinary Greek attraction or substitution, and which does not, in reality, stand for two cases itself. The same maybe said of this instance, ει τι τών αΧλων, ων νννΐ ^ίάβαλλε καΧ διεξϊ^ει, Demosth. de Cor. ^', the ών being a case of attraction with αΧΧων ; for, although the verb of accusing takes a genitive in Latin, διαβάλλω is construed with an accusative, see Θ', and preposition, or genitive of the person ; and Εί^ζψί likewise takes an accusative, see η\ So too ^ιεζίρχομαι, i^^ Some of the cases may be classed under this section, or § 19. {b) ; as, Ik τών φοινίκων, ο ι ήσαν εκττετττωκοτες τους δε και Ιζίκοπτον, Xen. Anab. ii. 3, 10, where supply ών, ' some of which they cut down,' or regard the last member as direct. Again, in consequence of a 'zeugma, we require to understand another case of the relative in relative sentences, as in the following, dii cujus jurare timent et fallere numen, Virg. JEn. vi. 324, where supply mentally, per quod with jurare. This, however, is not peculiar to the relative; and we may understand, per cujus numen, thus preserving the case of the relative. The following alFords an example unless other- wise explained ; ο υ ς 17 μίν πόΧις απηΧασε, σοι δ' ήσαν φ ίΧο ι, Dem. de Cor. κε ; but the change here is probably not without rhe- torical point and emphasis, and is introduced with this express view, ^ whom the state banished, but they were your particular friends.' The last instances which here follow are however free from objection : ο ν μητ οκνείτε, μητ αφητ έπος κακόν, CEdip. Colon. 728, ^ whom neither fear nor ο η w h ο m.' Even here, however, the relative has reference only to the possessive pronoun εμης, and had the construction been made by αλλά μη οκνεΐτε εμί, our case would only have exhibited a zeugma of two verbs governing dif- ferent cases of the personal pronoun. Again, ους κεν εν Jvoίηv και τοΰνομα μνθησαίμην, Horn. II. γ'. 235, where understand ών : nam qui cum ingeniis conflictatur ejusmodi neque commovetur animus, Ter. And. i. 1, QQ, where supply cujus: cut comminus ensem § 19.] RELAT. AND DEMONST. CLAUSE AMALGAMATED. 21 condidit adsurgenti, et multa morte recepit, Mn, ix. 347 : τις νυν ίχονται καπαναίρονται δόρυ, (Edip. Colon. 420^ ^respecting this battle which they are now preparing to begin, and in which they are lifting their spears against each other.^ W under com- pares with this, qiiam (cervam) procul incautam nemora inter Cresia fixit pastor agens telis, liquitque volatile ferrum nescius, Virg. ^n. iv. 71. It is convenient to explain all these cases, by supposing the relative assumed mentally, which we do naturally enough in endeavouring to accommodate the construction to an English one. But we might put the rule differently, and I am not sure that it would not be better to do so, by stating, that after a relative pronoun, when two verbs coupled by a con- junction occur, the second is not required to be in construction with the relative. This mode of representing the matter appears to me to agree better with the freedom of construction, which the Greeks particularly observed in many other cases, and which is not to be explained by inferring that they did this, or omitted to do that, which we now do in English. In all reasonings of this sort, it must be recollected, that although we may explain on the analogy of our own tongue, and while this may suffice for furnishing a rule which is intelligible, we must not assert that this furnishes a clue in all cases to the origin of a given practice. (/) The relative and demonstrative clause melted into one ^. This usually takes place after the interrogative relative pronoun, the relative of the explanatory member being absorbed. Thus, quis novus hie nostris successit sedibus hospes? Virg. ^n. iv. 10, for quis est hie novus hospes qui: similarly we have quid hoc veneni savit in prcscordiis F Hor. Epod. iii. 5, for quid est hoc veneni quod : and the practice is very usual in Greek ; as, όστις δδε κρατίει; Hom. II. 7γ'.424, ^who he is that is thus prevailing?^ τις ποθ' 7} δ' ό/χτίγυρίς ; ^sch. Choeph. 10: καΙ νυν τ ί τοντ αυ ψασί παν^ημω ττολεί κη ρυ y μα θύναι τον στρατϊ]^ον άρτίως; Soph. Antig. 7, ^ what is this proclamation which they say ?' ποίου yap ανΒρος r η ν δ ε μηνύει τυχην; Soph. (Edip. Tyr. 102, for ττοίου γαρ ανδρός εστίν τ/δε -η τύχη^ ην μηνύει. See also Eurip. Ph^n. 145, τις δ' ούτος ττερόί ; It is not, however, wholly confined to the interrogative ; as, ut nihil in iis non appareat. Incert. for nihil est quod. This is done with the adjectives primus, princeps, used ^ Even with the demonstrative, where no relative appears, this is done ; as, tanta hcec Icetitia oborta est, Ter. Heaut. iv. 3, 2, ' so great is this joy which has arisen.' 23 COMMON ORIGIN OF DEMONST. AND RELATIVE. [CH. III. to express the first person, or the party principally, who was afFected; as, quos pars cwitatis Helvetm insignem calamitatem populo Romano mtulit, ea princeps posnas persolvit, Cses. B. G. i. 12. Here no relative is expressed, but is implied, as we see in trans- lation, ' that was the party which chiefly suffered retribution/ Ohs. What is called the use of the article for the relative pronoun in Homer, is nothing more than this use of the direct construction. If we take the common example, τον βάλε ^eUov ωμον, and analyse it, we have a verb which would take, in the ordinary way, either an accusative of the person or of the thing separately, used with both coincidently ; and τόν may either be what we call the relative or the personal pronoun. The same passage, however, often occurs with /xtV, as II. ε'. 188. λ'. 583 ; and with κατά, as II. ε . 98. In Homer's time, it is probable that the article, pronoun personal, demonstrative, and relative, were much confused, or were merely variations of a common form or forms, of which the uses were not always very broadly or precisely distinguished. Take as a specimen, άλλα τ a μ ε ν πολιών εΙ,ετΓράθομεν, τα Ζέ^ασται, Hom. II. α. 125, where the first τά has the force of a, and the second of ταντα. Again, we have τάττερ for αττερ, as τ άτ ε στυγέονσι Θεοί περ, II. ν'. 65. This use of article, as relative, is common enough in the Attic tragedians ; though, from the introduction of a point before many of the cases, they are rather regarded by the editors as cases of the direct clause: thus, τ ό v, ώ ττυρφόρων άστραπάν κράτη νέμων ώ Ζεϋ ττάτερ, υπο σω φθίσον κεραννω, Soph. (Edip. Tyr. 195. In the following, the article is put for the relative, Έλένας ίπΐ λέκτρα, τ αν καλλίσταν δ χρυσοφαης " Αλιος ανγάζει, Eurip. Hecub. 629. A further evidence of this confusion in Homer may be inferred from another practice not uncommon with him. When he has commenced with a relative pronoun, separated from its verb by a long parenthesis, the form of the relative pronoun is again given, rather than that of the personal pronoun ; and there are besides cases in which 6ς is better translated by ' and he :' as, δ ς τίκτε Θεανώ καλλιττάρτ^ον, Horn. II. λ'. 224 : ο ς ρα τότ Άτρεί^εω 'Αγαμέμνονος άντίος ήλθεν, ibid. 231. See Trollope, also II. φ'. 198, and Theogn. 205, quoted by him. I have already noticed above, a double use of the relative, where the last has merely the force of the copula, II. ε'. 404. Again, as already shown, the construction begun by the relative, is often terminated by the personal pronoun ; as, δ ς μέγα πάντων Άργείων κρατέει και ο Ι πείθονται 'Αχαιοί, II. α'. 78. Just as ος is followed by ρα, so is τόν : thus, τόν ρά ποτ αυτός, II. φ'. 35, SO also II. β'. 21. 309. 742. 4^'. «402, and other places, which an index will supply. This pa is a common appendage of the relative ος, just as άρ of the interrogative ; as, τις τ άρ σφωε θεών, II. α\ 8. Το the proof of the original oneness of these forms may be added the Attic use of ος, which maintained its place in later times, in the well-known phrases so frequent in Plato, ή I' ος, ή ^' η, and in the expressions, κα\ ος γελάσας εψη, Protag. 310, D : και ος είπεν έρνθριάσας, ibid. 312, Α, and elsewhere. In the later Greek writers, it is curious that the relative forms δς μέν, ος ^έ, usurped the place of the demonstratives ό μέν, 6 ^έ. This is supposed not to have got into use in the time of Demosthenes, though it is found in two psephismata in the Crown Oration, and in some readings in the text, § 71. See Bremi's note § 19.] COMMON ORIGIN OF DEMONST. AND RELATIVE. 23 to his edition. It was of earlier occurrence, however, in Doric writers, and its use in legal formulae is in favour of its antiquity. How this usage originated, it is perhaps impossible to say ; but it may be thought to bear some analogy to the phrase εστίν a and its cases, which are used in a somewhat similar way. Thus, in the following, tri ^έ και εν τοΊς βαρ- βάροις εστίν οΐ ς νυν, Thucyd. i. 6 ; and again, της τε άλλης 'Ελλάδος εστίν α γωρία^ ibid. i. 12. St. Paul uses the ος μεν form in his epistles in three different ways ; first, as answering to ος U, as καΐ ο ς μεν ττεινξ. ο ς ^ε μεθύει, 1 Cor. xi. 21 ; secondly, opposed to άλλος ^ε, as J μεν γαρ, άλλω ^ε, 1 Cor. xii. 8 ; and thirdly, without the second or contrasted member, as καΙ ους μεν εθετο πρώτον αποστόλους δεύτερον κ. τ. λ. ibid. xii. 28. It is at least curious, that the English language should have retained this use of the relative for the partitive adjective, or something which very closely resembles it. Thus, ' And such a flood of greatness fell on you. What with our help, what with the absent king, what with the miseries of a wanton time,' Shakesp. I. Henry IV. act v. sc. 1, where 'what' might be supplied by 'partly.' It is not always easy to decide, however, whether Homer is using what is termed the article, demonstrative, or relative pronoun. Thus, take for example, μνησάμενοί τα εκταστ-ος ενί μεγάροισιν ελειπον, Hom. II. τ. 339. Here, if we allowed Homer the use of the article, and considered the construction as determined by the sense, we might put λειπόμενα in place of the sen- tence following τά : or we may consider the relative ά to be elliptical, and τά as put for ταϋτα : or we may suppose that the demonstrative is wanting, so commonly the case in Greek, and that τά is the relative supplying its place. The article differs from the relative in the later form of declension, chiefly by the former retaining the initial τ through most of the oblique cases, while the latter always commences with the aspirated vowel. In the Doric dialect, however, the initial r was preserved in the nominative plural of the article, proving that, like the relative form, it has experienced a change of its original form, by dropping it in some of the dialects. Thus, τ at Αιωνύσου πόθεν ε^εφανεν συν βοηλάτ^. γάριτες οιθυράμβω ; Find. Olymp. xiii. 18 : α r' ΈλενσΙς τα ί Ο' ι/ττ' Αίτνης υφιλόίρον καλ- λίπλουτοί πόλιες, α τ Εύβοια, Find. Olymp. xiii. 110, where take α for the article, and not the neuter plural of the relative. This passage is quoted by Matthiae in his Grammar, as exemplifying the use of the article with proper names ; also ταί οχΘαι, Find. Fyth. i. 18. And, on the other hand, we have the τος form of the relative pronoun in other positions than the initial one, where we cannot resort to the supposition that the direct structure is put in place of that by the relative ; as, οε^αι ^ε' υΐ στέφανων εγκώμιον τεθμόν, τον άγει πεδίων εκ Πί'σας πεντάθλω άμα σταόίον νικών όρόμον' άντεβόλησεν τών άνηρ θνατός ουπω τις πρότερον, Find. Olymp. xiii. 29, ' and receive from him the customary comus which he conducts from the plains of Fisa, having conquered in the stadium and quinquertium, which never mortal man before obtained.' The τος and ος form occur together in two contrasted members of a sen- tence, in their partitive use, in the following : rot μεν γένει φίλω συν Άτρέος Έλέναν κομίζοντες, ο Ι 3' από πάμπαν είργοντες, Find. Olymp. xiii. 58, 'some with the offspring of Atreus, whose friends they w^ere, seeking to carry back Helen, and others to prevent it.' In addition to using pa in common, the relative and demonstrative both take μεν and ^ε, as τα μεν. Find. Olymp. xi. 7, occurring after a point, for d μεν, agreeing with άφθόνητος αίνος as a noun of multitude, and Fyth. i. 21. They 24 COMMON ORIGIN OF DEMONST. AND RELATIVE. [CH. III. both likewise take ττο-ε, as τόν ττοτε, Find. Pyth. i. 16, here relative ; also ττερ, as τάττερ, άττερ, besides other similar combinations. An examination of the demonstrative and relative forms, both in Latin and Greek, shows plainly that they have all the same origin. Thus, hujus and cujus, hie and qui, are in reality the same words, since the letters h, c, and q are often no more than the symbols of an aspirated vowel, as may be seen in the derivation of the English ' whom,' from the Latin quern. Again, with regard to 6 and oc, the aspirate shows that there was something beyond the mere vowel-somid in the first of these, which may have sometimes preceded and sometimes followed, or both together. Thus, 6 might become 6ς, or even toc, though probably the reverse process represents the truth more nearly, viz. that which supposes that the sibilant at the end was gradually softened till it disappeared, while the initial breathing or τ came to be represented by the aspirate mark. In Latin, by a similar method, w^ might obtain iq (hie) or qi (qui)^ The distinction between the Ρος and τος, or relative and demon- strative form, has been well preserved in English ; as, for instance, in 'whose' and 'those;' and it would be interesting, could we make out a complete history of these modifications. With regard to Home Tooke's derivation of qui from καΙ o'l, adopted from Voss, I hardly know what to say. Undoubtedly, in many instances, the relative pronoun can in signi- fication be resolved into the copulative conjunction and demonstrative ; and in consonance with this is certainly the frequent use of τε in Homer after what is called the relative ; as, κείται dyy)p ΰ ν r' ~iaov ετίημεν'Έκτορί ^ίω, II. ε. 467. See Trollope's note, who quotes Hermann to the effect that ό'ς originally signified hie, and υστε the same as qui, or et hie, a force which the relative often possesses, both in sense and construction, as we shall elsewhere see. Another good instance to the same effect is or ε προ'ώωσιν Ιόντα Κίρκον, ο τ ε ajxiKorjat ή)όΐ'ον φέρει όριήθεσσιν,ΙΙ. ρ'. 756. But, on the other hand, the aspiration of ό'ς in Greek admits the suppo- sition of some consonantal sound before it, having the same force as the digamma, or English w, which is equivalent to the qu of the Latins, and answers to the old orthography of ' whom,' as found in some authors, qliam. Even, however, supposing this to be fanciful, there was still a τος form in Homer, preserved in the oblique cases, -^^hich bears the same relation to 6q as τόσσος to οσσος, as talis or tantus to qualis, quantus in Latin, or as the English 'those' to 'whose.' That the relative is often equivalent to the copula and demonstrative, that the demonstrative and copula often fill its place, particularly in the earlier dialects, and that the Latins often render the latter by qui, are points beyond dispute, and all tending to render the derivation of qui from κα\ ος feasible ; still I do not think the etymology at all demonstrated, though it appears more satis- factory than the majority of such derivations usually are. ^ That the final c in these cases is a mere ending, which has nothing to do with the root, or which is indicative of a peculiar sound sometimes given to the vowel, seems to be established from the forms hie, hi, singular and plural, ne and nee, harum and harmic, the Greek vvv and Latin nunc, the termination hoe in the ablative compared with other ablatives in o, that of hu7ie compared with U7n in other accusatives, the ending ee attached to some words, as his, hisee, also by comparing illue and illud, istuc and istud, quiequam and quidquam, and the German hin with the Latin hinc. § 19.] COMMON ORIGIN OF DEMONST. AND RELATIVE. 25 I have not, however, the slightest doubt, that the distinction between the relative and demonstrative did not at first exist. I do not mean to say that Homer made no distinction, as this would be altogether inaccurate. The Latins used the initial relative after a full point, without limitation, in lieu of the demonstrative aud copula. This does not appear to be the case in Homer, so far as I have noticed ; for in these cases he uses the τος form for oblique cases, reserving the ος form for the more strictly relative clause, occurring in the middle of a period ; though he likewise uses τά freely in this last situation, accompanied commonly by μεν and ^ε, which seems to connect the τά in its origin with the demonstrative. He does not, any more than Pindar, use the τος form for the nominative singular, and the nominative plural is in general o'/, though Pindar, as We have seen, makes use of rot, ταί frequently enough in the plural. On the other hand, the particle pa is very usual with Toy, as with ov the relative, in the beginning of a period, where the Latins use the relative form. The inter- change of the article and relative, or rather the use of the same form for both, is exhibited familiarly enough in the German language, while relative adverbs are used by many writers, where in English we require the demonstrative forms ; and vice versa, the demonstrative forms are often to be translated by relative ones in our language. Thus, dadurch, ' whereby,' woven, ' thereof,' Kant's Criticism of Pure Reason, p. 218. The suppo- sition, that only one form originally existed both for article and pronoun, is confirmed by the practice of the Hebrew tongue, in which the use of the relative is unknown, and whose relative clauses are expressed by means of distinct parenthetic propositions. In Homer, as we have seen, there is some difficulty in distinguishing the two classes of words ; for although separate forms had in his time got into use, yet the necessities of the metre may sometimes account for the use of one in preference to the other. Even in much later periods, the so-called article had not lost its force as a relative, as may be seen from many passages in the Attic poets. Thus, γνναι, νοείς εκείνον, omv άρτίως μολείν εφιέμεσθα, τ ό ν θ' οντος λέγει, CEdip. Tyr. 1025, ' and whom this man speaks of:' μ?) καΙ μάθη, μ' ηκοντα, κακ\έω το ττάν σόφισμα, τω νιν αυτίχ αίρησειν ^οκώ, Soph. Philoct. 13, ' the whole cunning device, by which I expect very shortly to get possession of him.' Having alluded to the use of ος μεν, ος ^έ, for δ μεν, 6 Si, I shall be excused for noticing another class of cases, in which 6 U is used, not by way of apposition or contrast, but where its place might be supplied by the relative pronoun ; as, τίς oh Ιυνασγαλα. κακοΊς τεοΊσι, ^ίχα γε Διός ; ό S' ετΓΐκότως ηεΐ (sc. Ζευς), ^ilsch. Prom. 167 : φοβούμαι, τον μυριωττον εϊσορώσα βούταν, δ ^έ πορεύεται, ibid. 584. See also the same play, vv. 679, 881^. Many other examples of this kind occur in the Tragic poets. See Monk's Hippolytus, 280, who observes, that the prsepositive article with μεν, γάρ, ci, is often put for εκείνος. This is true in these cases, but he has not noticed the whole peculiarity, viz. that in general there is a direct construction, in place of which, in our own language, we should usually employ the relative. The Latin language, on the other hand, very commonly uses a relative at the beginning of a fresh period, where a demonstrative would be expected ; as, cuj us reprehensio me vehementer movet, Cic. pro Muren. 2. 3 : quce si causa nan esset, tamen Blomfield. 26 THE RELATIVE CONSTRUCTION [CH. IV. &c. ibid. 4. 8. The remark of Matthias on the passage, aypiq. ΙυντακεΙς νόσω νοσεί τΧημων ""Ορέστης' δ ^ε ττ ε σων εν ^εμνίοις κείται, Eurip. Orest. 34, I do not understand, where he says, that it is more suspicious by reason of no opposition being implied. It is strictly analogous to the examples quoted immediately above, from the Prometheus of ^schylus. Homer makes use of the same thing, χητε'ί τοιοΰ^' νιος' 6 S ' άλλοΕαπω ενί ^ήμω ε'ίνεκα ριγεΒανής Ελένης Ύρωσιν ποΧεμίζω, II. τ. 324. CHAPTER IV. THE RELATIVE CONSTRUCTION AND THE DIRECT IN THE SAME CLAUSE. § 20. — The Septuagint Greek version of the Old Testament, and also that of the New, contain many instances of this class 5 which are, however, not wholly confined to these writings. But as αυτός is the word which occurs most frequently in these cases, it has been usual to treat this peculiarity as one which arises out of the redundant use of αυτός. This mode of representation does not give a sufficient view of the matter ; the fact being, that either the construction by αυτός, or the relative alone, may be made use of, and, according to the general usage of language, the one should be exclusive of the other, while here, on the contrary, both are, as it were, jumbled into one. This is not confined to αυτός, as we shall presently see ; while, again, there are many other modes of expression, in which this pronoun appears superfluous, or rather seems, as it were, put to indicate the government of the verb that accompanies it, which has been separated from its own natural connexion, by some independent construction, as the genitive abso- lute or so forth, where no relative construction has been employed. The repetition of the antecedent substantive, in the relative clause, has already been abundantly illustrated, but this is quite a differ- ent thing from the repetition of the demonstrative pronoun ; be- cause, although the latter is but a pronominal representative of the antecedent substantive, and therefore may be thought merely to occupy its place, it really contains, in itself, much more ; viz., a peculiar law of construction as well as a correspondent sense. The writer of the Apocalypse indulges this practice more, perhaps, than any other. Thus δέδωκα Ινώπιόν σου Θύραν ανζω^μένην rj ν ουδείς δύναται κΧΰσαί αυτήν, Rev. iii. 8. Some editions read και for rjv, but if this reading has been adopted merely to obviate the difficulty in construction, it is of little use, as the examples of § 20.] AND THE DIRECT IN THE SAME CLAUSE, 27 a similar kind are exceedingly numerous ; άγγίλοίς οίς εδόθτ/ αντ οΐς, ibid. vii. 3 ; δ ν αριθμησαι αυτόν, ibid. νϋ. 9. See also Septuagint, Gen. α'. 11, 12. κΙ\ 3, 37, 42. κη. 13. λα'. 13. λθ'. 20 ; also in the following, άττο του ξύλου ο υ ενετειλάμην σοι τού- του, Gen. y. 3, 11, 17; at Isa. r{. 20, the very prepositions are repeated in both members. Ohs. Nearly allied to this is the subjoined construction, in which it is the relative clause, however, that appears to be redundant : αντ οΰ τα σημεία a iiroUt, John ii. 23. v. 36. vi. 2. 3 Ep. John i. 10 : and again, τα έργα σου a ττοιεΊς, John vii. 3. Gen. ii. 2. But this construction admits of explanation on another principle, and is analogous to another very common practice of the Greeks, viz. the removing the subject out of the subordinate relative clause, and putting it as the direct object to a verb or other word preceding it, whose proper object is, according to Latin and English practice, the whole relative clause. See my Accus. and Infin. chap. iv. p. 38. So Κρέοντος, οίά μα βεβουλευκώς 'έχει, Soph. QEdip. Tyr. 674, in answer to otqv πράγματος μηνιν τοσην^ε έγεις, where Creon is taken from its own clause, and put answering to Ίτρά-γματος^ in lieu of πραγμάτων understood. Thus, in the case above, the natural construction would be τα σημε'ια α αύ^ός εποίει. In lieu of which, αντός is made to depend on σημε'ια, as the latter of two substantives, precisely as in the following English example, formed exactly on the Grsecian model, the verb takes the subject of the relative clause for its object, with the remainder of the clause in apposition, instead of the whole clause as an object; 'then you perceive the body of our kingdom, how foul it is,' Shakesp. II. Henry IV. act iii. sc. 1, for ' then you perceive how foul the body of our kingdom is :' τον ^ ε Αάϊο ν ώυσιν τίν εΤγε φράζε, τίνα δ' άκμην ήβης έχων, QEdip. Tyr. 713, ' tell Laius what nature he had,' for ' tell what sort of person Laius was ;' and note also that έχων follows the construction understood, and not that expressed. Some supposed cases of anacoluthon might be explained in this way. I may remark, in passing, that not merely the subject, but even the object of the dependent relative clause is sometimes removed in this way ; as, ηάν^αρον άντί- θεον Βίζημενος, ε'ί που ε φ ε ν ρ ο ί, Hom. II. ε. 168 : καΐ ε υ χη ν ^έ τίνες αυτοϋ εζέφερον, ώς ευχοιτο, Xen. Anab. i. 9, 11, though ώς ενχοιτο may be regarded as redundant or exegetical. Sometimes the opposite to this takes place, particularly with certain uses of quis and τις, where a dative is transferred from the preceding verb to the relative clause, of which it then becomes the subject ; as, sed nan ante datur telluris operta subire auricomos quam quis decerpserit arb ore foetus, Virg. ^n. vi. 140, for nan datur alicui quam decerpserit, where the dative to datur becomes nominative to decerpserit ; but falling in this position after quam, the unaccented form of the word is used, viz. quis, according to usage after other relative forms. This is, perhaps, properly a Greek structure ; as, *'Έιλληνί καΐ βαρβάρω πορεΰεσθαι, οπη ης ηθελεν, Anab. i. 9, 13, for τινί Ελλήνων. In some other instances it happens, that the subject of the relative clause is made, not the object, but the subject of the preceding ; as, εΐ τα μεν χρήματα λνπεΊ. τινας υμών εΐ Βιαρπασθησεται, Dem. pro Cher. ιγ. We can, however, explain this last example, in accordance with another principle extensively prevalent in Greek, whereby usages 6 28 THE RELATIVE CONSTRUCTION [CH. IV. which would in another language be expressed impersonally, are put personally. Thus, 'if it grieves you that the moneys be plimdered' is rendered ' if the moneys grieve you that they are plundered.' § 31. — This same thing occurs with the relative and demonstra- tive adverb^ which are thus found in one and the same member; as δ 7Γ ο υ εχεί ε κ: ε Τ, Rev. χϋ. 6. where, as above, some editions reject εκεΤ, though they retain it in other cases strictly analogous ; as δ•7Γου τρίφεται ε κ ε Τ, ibid. χϋ. 14. Again, we have, οίος ουκ: lyivETO αφ' ου οι άνθρωποι lyevovro εττι της yvg, τ ηΧικουτο ς σεισμός οντ ω μίγας, ibid. χνί. 18 : δ ττ ο υ ύ) yvvrj κάθηται επ αυτών, ibid. χνϋ. 9. See also Septuagint, Gen. ly. 3. κ. 13. Obs. 1. Hence, as we have seen in the previous chapter, where there are two members in a sentence, the first takes the one, and the second the other construction ; as, μεθ' η ς επόρνενσαν οί βασιλείς της γης, και εμε- θύσθησαν οΙ κατοικοϋντες την γήν εκ τοΰ οίνου της πορνείας αυτής. Rev. χνϋ. 2. So John i. 33. 2. The nearest thing akin to this double use of demonstrative and relative in the same member, occurring in St. John's Gospel, is this : ο υ εγώ ουκ εΙμΙ αί,ίος 'ίνα λύσω αυτοί/ τον Ιμάντα τοΰ υττοίηματος, John i. 27. I mention this, not by way of discrediting St. John's alleged author- ship of the Apocalypse, but as worthy of critical notice. In other respects, there is a correspondency of style, in the use of parts of sentences put absolutely out of their natural syntactical connexion, and in the employ- ment of iVa, which would induce us to argue, that the writer of the Gospel, and the other books ascribed to the same author, is one and the same. 3. In illiterate English composition we meet with the same thing ; as, ' within this stone here one is laid, W h ο m none with vice could her upbraid,' Tombstone in Chippenham Churchyard. 4. There are several examples in Latin, in which the relative pronoun is said to be redundant, in which the demonstrative and relative occur together, but which I explain differently. Thus, the following, quoted by Facciolati, adeon' me fuisse fungum, ut qui illi crederem, Plant. Bacch. ii. 3, 40 : qui te Jupiter diique omnes, percontator perduint, ibid. Men. v. 5, 31 : De Apollonio quod scribis, qui illi, &c. Cic. Att. lib. iv. ep. 7: and again, qui ilium dii omnes perduint, Ter. Phorm. i. 2, 72. This may evidently be explained by the analogy of many passages, in which the relative stands merely for the copulative conjunction and demonstrative pronoun, while it is here put merely in place of the conjunction, through the unnecessary change to the direct structure, which is really implied in the use of the relative alone. With regard to some of these, the indig- nation may explain the confusion of structure. So α προλέγω υμΊν, καθώς και πρυείπον, οτι οί τα τοιαύτα πράσσοντες, κ.τ.λ. Ερ. ad Galat. v. 21. 5. Somewhat different from the examples given above, is that which follows, in which the relative belongs to one verb, and the αυτός to an interposed subordinate clause, and the construction is strictly correct according to ordinary syntax. The only peculiarity in such examples is, that generally the secondary or subordinate clauses of a relative sentence. § 22.] AND THE DIRECT IN THE SAME CLAUSE. 29 when the relative is in an objective case, do not usually repeat the object again in English, but the verbs in such clauses all seem to be referred to the first relative as a common object. Hence the apparent redundancy in the following : τά γαρ έργα α ε^ωκέ μοι δ πατήρ Ίνα τελειώσω αυτά, John V. 36. Regarded in another point of view, however, an instance like this is similar to others we have brought forward above. To the same purpose are certain passages in which the English translators have imitated this exactly. Thus, 'whom when they had washed, they laid her in an upper chamber,' Acts ix. 37, where we should do better to read, 'whom, when they had washed her, they laid.' The obvious confusion which is thus introduced, as to which verb the relative is really governed by, explains sufficiently why the two verbs should be referred to one and the same object, without the introduction of a second. So again, 'whom when Aquila and Priscilla had heard, they took him,' Acts xviii. 26 : 'whose shoes of his feet,' ibid. xiii. 25, for ' the shoes of whose feet:' 'what God hath cleansed, that call not thou common or unclean,' ibid. x. 15, where ' what' should by itself be equivalent to ' that which,' and is almost always so used. The practice of supplying each clause with its own object is not even confined to αυτός, for we have, though rarely, examples like the following: και ους αν λάβωσι των Ελλήνων ^εινά ν(3ρίζειν λέγονται τους "Ελλη ν ας, Xen. Anab. vi. 4, 2. § 22. — Agreeably to what has been shown above^ we may, I think, defend the integrity of such passages as the following, in which the personal pronoun is placed in close contiguity with the relative in the same member. They are, ov, κα\ irpXv Ις φως μητρός εκ ■γονής μοΧειν ayovov ΆτΓολλων Aaito μ Έθίσπισεν, Eurip. Phoen. 1597 : fj ν χρην σ εΧαννειν τ η ν δ' υπέρ Νείλου ροάς, ibid. Androm. 641 : και παΧς ατεκνος, fj ν δδ' ε% ημών γεγως ελα Si οίκων τ η ν S^ εΐΓίσπάσας κόμης, ibid. 700 : οί μεν τοσούτοι τταίδες ουκετ είσί μοι, Eurip. Hec. 804, which the editors seem to think corrupt, but I own I do not. Again, we have αΐτ εττεϊ ουν χειμώνα y ς? Eurip. Hecub. 1172: α τον tovS^ οΰποτ ευφράνει βίον' 6 ς γ' ω κάκιστε, σκήπτρα και θρόνους έχων, α νυν 6 σος ^ύναιμος εν θηβαις έχει, τον αυτός αυτού πάτερα τ ό ν d' άπηΧασας, Soph. CEdip. Col. 1348, where both change of person and non-agreement with the last subject may be noted. But it may be remarked here, that the change of person is really only imaginary, for the first τοί^δε, as well as the last τόνδε, are used ^εικτικώς, a common practice in a personal address, and hence the relative has its person determined by the sense, which in actual dialogue could not give rise to misapprehension. § 60. — The freedom of Greek construction often allows the sense, and not the exact words, to determine the agreement, in other cases. Thus, otcu yap ώφθην ευτυγρυσ , αιδώς μ' έχει, εν τφδε πότμίιύ τ υ γ χ ά ρ ο υ σ ', Ί'ν' ειμί νυν, Eurip. Hecub. 949, where see PflugFs note to his edition, verse 970, and parallel passages there quoted. In this example, τνγχάνουσα is conformed to αΐ^οϊιμαι, implied in αΐΒώς μ έχει, while ellipsis takes place, the sense being ' for I am ashamed to be seen in adversity, in the condition at present allotted me, by those by whom I was seen in prosperity/ And thus in other cases the relative has reference Η 98 ACCUSATIVE AFTER VERBS OF ASKING. [CH. XII, to an implied subject ; as, μη γαρ α/δε Βαίμονες Oeiiv μ αψωνον τησΒε της αράς ίτι, υ ς με, κ. τ. λ. (Edip. Col. 862. See also the examples under the head of Ellipsis. This construction, κατά σννεσιν, is not uncommon in Thucydides ; as, a tl χρη κάντανθα πιστενειν, η ν εικός, κ. τ. λ. Thucyd. i. 10, for και ττερΧ ταύτης της στρατείας \iyovTi, So μέλλοντας below for the genitive. See Arnold. § 61. — Several cases occur in which the relative is found acci- dentally, which do not particularly belong to the relative syntax, or present any peculiarity merely characteristic of it. Thus the subject is sometimes expressed only in the subordinate member in com- parisons; as, Caule suburbano qui siccis crevit in agris dulcior, Hor. Sat. ii. 4, 15, but this is often the case with adjectives; as, ex unis geminas mihi confides nuptias, Ter. And. iv. 1, 51 : Calabris Luc an a mutet pascuis, Hor. Epod. i. 27: Picenis cedunt pom is Tiburtia succo, ibid. Sat. ii. 4, 70 : quippe duos pro uno domino acceptos, Liv. iii. 9. In the annexed example, the causal relative is put after another relative pronoun as the antece- dent : ως our' av ο ς vvv σκητττρα και Θρόνους έχει μείνειεν, ουτ αν ούζεΧηΧνθ ώ ς πάΧιν εΧθοι ποτ ανθις' οί γ ε, κ. τ. λ. Soph. (Edip. Col. 421. Occasionally, a verb which takes ordinarily one of two constructions, takes both together, and the relative may happen to be found in such an arrangement ; as, ο υ ς ού μη ττοτε χώρας φν^όντες τησ^^ επεύΕ,ωνται θεοΐς, ibid. 1021. Here the only peculiarity is, that <^εύγω takes either an accusative of the person, or a genitive of place, but both constructions are found in this place, and the relative is the accusative of person. § 62. — Verbs of asking, in Greek and Latin, take two accusa- tives, one of the person from whom any thing is asked, and the other of the thing sought. This is also in Greek extended to the person or thing respecting whom a question is asked ; as, οπότερον κα\ έρωτας ; Plat. Euthyd. 271, A, ' respecting which do you inquire ?^ ov έρωτας ; ibid. B, ^ respecting whom you ask ?' This is not peculiar to the relative, as another substantive is sometimes put in apposition ; as, ο δε συ έρωτας την σοφίαν αντοΐν, Θαυ- μάσια, ibid. C, ^ concerning what you ask, viz. their wisdom, it is wonderful.^ But still this construction does not appear to have always satisfied the writer, as περί αυτών is subjoined. Thus, a νυν Βη ηρωτώμεν περί αυτών ; Plat. Theaet. 185, C, which in this form strongly reminds us of the construction παν ρΐιμα apyov .... πε^η αυτού, Matt. xii. 36, though different. The last passage from Plato, in fact, contains two constructions. The one περί αυτών in place of an accusative after ηρωτώμεν, in the sense of ^ whom do you § 67.] RELAT. AS SUBJECT AND OBJECT. — INTERROGATIVES. 99 speak of?^ which sense Ιρωτ^ς appears to have in the three pre- vious examples; and the other a, the proper construction after έρωτας, in the sense of ^ asking/ While speaking of these passages, it is as well to notice, that Stallbaum defends the use of οττότερον, Plat. Euthyd. 271, A, the relative adjective, in preference to the mere interrogative form πότερο v, in this and some other instances ; in English, ' I would wish to know which of the two you mean.' So δ Ti ίστί TovTo, Plat. Men. 74, D, in lieu of τί ίστι τοΰτο ; where this last-named editor observes, " Sic οττότερον dictum Lysid. 212, C. Euthydem. 271, A. De Repub. i. 348, B. Gorg. 502, Β : atque oTi prorsus eodem modo positum^ Phileb. 23, C. Euthydem. 287, C. Sophist. 236, Ε :" and again, Euthydem. 271, A, he says, *^ Per- tinet hue etiam locus Lysiae Accusat. Nicom. p. 840, ed. Reisk == 259, ed. Bremi, /cat γαρ tol, ώ άνδρες δίκασταί, ΙττειΒη Ικείνων ^ίκην ου δεδωκεν, οποίαν και νυν την αρχήν κατεστησατο ; ubi Mark- landus et Taylorus errarunt mirificeP § 63. — The relative, in the example which follows, stands both as a subject and object ; as, omnia quae aut amisi aut ex necessariis advorsa facta sunt, Sail. J. B. xiv. 16. This instance is quoted by Stallbaum, in his note to the Charmides of Plato, 156, C, as illus- trating a similar Greek passage in the text, otl τ avr a οντω \iy ου σι τε καΧ ε χει, § 64. — We have seen above how Stallbaum defends the relative in preference to the interrogative form, and we must now advert to others, in which he prefers the interrogative form to the indefinite or ordinary relative form, where another relative has preceded. We have already touched on some cases nearly allied, under the head of double relative uses. Thus, οτφ r ί ποιεΧν rj τταθεΐν ύττό του πεφνκε, Plat. Phaedr. 271, A : φροντιστεον τ ί ερουσιν, Crit. 48, Α, for ο Τί, where see his note to the passage, and parallel instances, Phileb. 17, A. Gorg. 500, A. ibid. 448, E. Phsedr. 271, A. Charmid. 160, D. De Repub. iii. 414, D. ix. 578, E. Legg. i. 632, C. vi. 767, C. vii. 803, A. Alcib. i. Ill, E. 114, A. Demosth. de Coron. 275. § 65. — In English, as in Greek, we sometimes make the article precede the interrogative ; as, ^ The what ?' and sometimes where the relative is not interrogatory; as, ^ There is a history the which observed,' II. Hen. IV. act 3. sc. 1. § 66. — Stallbaum notes a peculiar collocation with the relative, και vvv δ?) a έλεγαν, for a ελεγον νυν δι?, ' what I said but now,' Plat. Euthyd. 288, B, and quotes Garatonius ad Cic. Orat. pro Mil. 16. 42. § 67. — After the verb ^ to be,' in English, when the word ^ same ' Η 2 100 OF CERTAIN GREEK PECULIARITTES. occurs_, it is not followed by the relative, as in Latin. Thus, Livy says, eosdem et Romanos et hostes esse qui per tot scBcula fue- runt, but we say ^ the same as/ When no comparison is implied, then the construction in Latin and English is the same; as, se quoque eundem dictatorem qui fuderit, Liv. iv. 31, in English, ^ the same dictator who/ § 68. — The interrogative relative ^how^ is generally made in Latin by the ablatives quo or qui, and according to the rule that the redditive member is put in the same case as the interrogative, we have the answer in the ablative, which is in Latin the gerund in do. The Greek πώς, on the other hand, is followed in the redditive member by the nominative of the present participle. In like manner, the nominative of the participle has often in Greek an instrumental force, in other cases not interrogative. § 69. — The Greeks are fond of crowding more into a sentence than its strict syntax will permit. See the use of two subjects in relative connexions, where one is put as a direct object to the fore- going verb, as in the case ^ ye see your calling, brethren, how that not many wise men are called ;^ and thus, in the following, we have an additional verb ; as, τοσούτον μόνον σε Ιγιγνωσκον, όσον η κονο ν ^Αθήναιον είναι, Xen. Anab. iii. 1, 45. § 70. — I have elsewhere noticed, § 37 («). Obs., instances of όστις used for ος, also after τις, and also followed by 6ς of the same subject; and in the following it appears to be used of a definite subject; as, την Ύισσαφερνους απιστίαν όστις λέγων, («>ς κ.τ.Χ. Xen. Anab. iii. 2, 4, though the meaning perhaps is, ' in that he or any other person does so f and again, όστις, Q5dip. Col. 959. Aristoph. Vesp. 632. 711. And in this passage, η ^υσπετώς αν τους ε μου ς αθΧους φεροις οτ ^) θανείν μεν εστίν ου ττεπρωμενον, -^schyl. Prom. 777. Blomf , where also, though οτ^) has a definite reference to Ιμονς or its primitive, the sentiment is generalized, ^ to whom as in my case, and wherever the same exemption exists,^ &c. The following is peculiar: εστί δ' όστις καΧ κατελήφθη, Xen. Anab. i. 8, 20, ^ there was one individual who was taken by surprise,' as the use of τούτον immediately after seems to compel us to adopt this definite sense. We have seen the Latin aliquis used in the plural sense of ^ some,' and we may also here imagine that όστις is a singular put for plural, which would doubtless be the fact, were τούτους read in lieu of τούτον» CONCLUSION OF THE FIRST PART. SECOND PART. PRELIMINARY REMARKS. I MUST beg the reader to keep in view^ that, in what follows^ it is the doctrine of the Latin subjunctive, as employed in connexion with the relative and its family of words, that I am chiefly aiming to elucidate ; and that although reference will be made to parallel Greek constructions, this is in accordance with my general design, by which the two languages are made to illustrate each other. To have embraced the whole syntax of the Greek conjunctive and optative was never intended by me, nor should I, in all probability, have been able to have added any thing new to what has been so effectually done in the larger Greek Grammars of Buttmann, Mat- thiae, Rost, &c. now so generally in use. Τ have confined myself, theiefore, throughout these pages, rather to what seemed to me susceptible of further explanation, in a way not as yet exactly attempted, and also as to which I was sanguine enough to hope I might add something of practical value. It has been already noted at the commencement, that the whole is to be regarded as a con- tribution to the department of which it treats, and it may be added, that much is professedly left untouched, which would require to be comprised in a work aiming at systematic completeness, and to exliaust the subject which it embraces. CHAPTER XIII. OF THE RELATIVE AND VERB. § 71. — Having thus extensively exemplified the various pecu- liarities of the relative, in reference to its own character and uses, I come to consider what influence is exerted by it on the mood of the verb, with which it is connected in its own clause ; and whether there may not be discovered some general and extensive 104 COMMON PRINCIPLE INVOLVED IN THE USE [CH. XIII. law on this point. The use of the conjunctive or subjunctive mood with the relative, is a point on which a good deal has been written, and much advanced to no purpose ; since, after all, there are some cases, where the practice seems to have been to a certain extent arbitrary, and others, where a potential force is intended by the writer, though not easy to be recognized. In English, the potential, by a mere idiom, is put for the indicative, as in the expression, ^ I should think so,^ which, taken strictly, is non- sensical; because a person either thinks in one way or another, and when he states that he thinks, he by no means asserts that his thinking one way in preference to another may not, in point of fact, be groundless. The use of the potential in this case marks only a modesty or hesitancy on the part of the speaker, impljring that, so far as he has considered the matter, he is of this or that opinion. Thus, too, in all languages, there will be found two or more modes of expression, often, for the same precise idea ; as, ibr instance, the use of possim, nolim, malim, velim, as softened irdi- catives, answering to the English ^ I could wish,' ^ I should prefer,' &c. And again, many present tenses subjunctive are rendered in English by the imperfect tense ; as, monet imitetur, ' he advises that he should imitate,' mandat adeat, ' he orders that he should go to,' where we can easily discover the idiom by choosing the English infinitive, ^ to imitate,' ' to go,' wherewith to render the original. § 72. — In addition, however, to what is here advanced, I may observe further, that there are many cases in Latin, in which, when translating into English, we use the potential, where precisely the same principles are applicable, in accounting for the mood eir- ployed, whether we call the Latin mood subjunctive or potentid. The Latin conditional mood, as it might better be termed, is but one and the same, whether we translate it by the English indicatiie^ or with the addition of potential or optative signs. There is nat, then, a different principle involved in the three cases, but it is with reference to the first, viz. the Latin subjunctive or the conditional mood translated by our indicative, that we find it more particularly necessary to furnish rules, as the English language has in some cases preserved the evidence of this dependency, by the introduction of auxiliary marks, and has dropped it in others, or exhibited it in a different way. In what follows, therefore, I have not thought it necessary always to discriminate the so-called subjunctive and potential, as a common principle attaches to both, and it is not always possible to do so, while both may be conceived occasionally to CO- exist. On the other hand, there is no doubt that many § 73.] OF THE SUBJ. AND POTENTIAL. — GREEK PRACTICE. 105 writers have erred, in ranking under one special rule cases more properly belonging to another special rule^ where, however, both rules are but particular exemplifications of one general law. § 73. — It must, nevertheless, be admitted, that the Latins were much more rigid in the observance of their syntactical rules than the Greeks, with whom they shared a kindred language ; in so much so, that of the deviations from strict rules of which they were sometimes guilty, the greater part were direct imitations of Greek idiom; and the style of many Latin authors determines pretty nearly the extent of their fondness for, and acquaintance with, their great literary precursors among that favoured race. The employment of the subjunctive after the relative, of which I am about to speak, was not a Greek practice ^ ; for although the optative does occur after relative particles, in certain connexions, where, and for the same reasons that the subjunctive occurs in Latin, this is only in particular cases, and to a limited extent. Of this practice I shall say something further below. However closely the two languages may in many points correspond, it does appear that the Latins had greatly refined upon the practice of the Greeks in this respect, and that they adhered with great uniformity to their own rule. There is no question, too, that this rule enabled them to convey, with great accuracy, some nice distinctions, most important to the full and perfect understanding of what they wrote, and more especially valuable to those who have now to * This must be taken with due quahfication. In one point of view, the Greeks may be said to have carried the practice further than the Latins, in as much as they had two moods, optative and conjunctive, and these modified by the introduction or omission of aV, or κε, κέν in Epic, after relative words. Thus, after ος, oVric, οίος, &c. where the relative does not mark a definite person or thing, but merely a probable, or supposable, or supposed case, the optative is used with or without aV, according as the relative sentence is to be represented as dependent on some condition, or is merely a supposition arising in the mind of the writer or speaker. Thus, after universal negative propositions, the optative with av in the relative sentence is common, in the same way as the subjunctive in Latin ; as, ο ν /c ε στ Li' ο τι αν τις μείζον τούτου κακόν πάθοι, Plat. Phied. 89, Τ), So, too, in interrogative sentences; as, και τις ποτ εστίν, ο ν γ' εγώ φέζαιμί τί, Soph. CEdip. Col. 1167. And thus, too, when the pronoun is used indefinitely without a direct reference. Again, when a relative clause follows a verb in the optative, expressing a wish ; as, ό ττ ττ ό r ε ρ ο t ττρότεροί υπέρ ορκια π η μ η ν ε ι α ν, ώΗ σφ' εγκέφαλος •χαμά^ις ρ ε ο ι, Hom. II. γ'. 299. In the oblique oration, particularly after λέγω, ερωτάω, είπον, when it is intended to mark that the opinion is expressed by the speaker, not the historian, as in Latin. See Matthias, Griech. Grammatik, 3rd edit. Leipzig, 1835. § 527, a. 527, b. β. § 528, 529. 106 OBJECTIVE RELAT. SENTENCE WITH SUBJUNCTIVE, [CH. XIII. translate their works. I have elsewhere been led to notice the great frequency with which the Greeks place, as an accusative to the governing verb^, what in Latin is made the subject of the governed relative clause, and have shown by numerous examples, that the relative not only takes the place of the elliptical demon- strative, as it does in Latin, but undergoes very frequently its government, so that the ordinary distinction between the syntax of the demonstrative and relative is here confounded. Here, then, was a mark of dependency, which the Latins signified, in a totally different way, by the use of the subjunctive mood, leaving the relative clause in possession of its subject, and the relative of its own peculiar construction, but marking clearly the objective nature of the clause. I shall make no apology for introducing the use of the subjunctive here ; for although its employment after the relative is connected by a tie of close affinity with many other cases where the relative does not occur, and may thus seem to belong more properly to the general doctrine of the subjunctive ; yet it is equally true, on the other hand, that the relative in this use may be regarded as the type or pattern to which the whole class of these usages has been conformed, and which affords by far the greatest number of exemplifications of the practice. In an earlier part of this work, I have shown how extensively words of relative origin serve the purpose of conjunctions, not only in the adverbial state, but where the forms are declinable, so that the relative in the construction of the genitive absolute may even be made to serve this purpose. It would hence appear, that the sub- junctive does properly belong to the doctrine of the relative pro- noun, and though this work has been undertaken with a different view than that of adding one to the many systems of rules on the use of the subjunctive, it cannot be here wholly omitted. I shall, however, confine myself rather to the illustration of general prin- ciples, than the mere accumulation of specific rules, which are to be met with in all the grammatical writers, and which may be studied more advantageously in connexion with examples for practice in composition. Objective relative sentence, § 74. — The most general and extensive law which can be laid down on this subject is, that where the relative clause, or that 2 Or cause it to depend on a preceding word sometimes, as the latter of two substantives; as, ύ} αγγελία τών πόλεων 'ότι ά^εσΓασί, Thucyd. i. 6L § 74.] INFINITIVE, UT AND THE SUBJ., ACC. AND INFIN. 107 which suppUes it, constitutes the whole object of a verb, or any governing word, the verb in that clause is put in the subjunctive mood ^. Thus, qum in eo reprehendat ostendit ; qu(2 ipse intellig at, qu(B civitas queratur proponit ; monet ut in reliquum tempus omnes suspiciones vitet, Caes. B. G. i. 16. Ohs. 1. From this rule would be excepted postquam, priusquam, <^c., in which the preposition may be regarded as governing a following clause. We must here therefore suppose a mere ellipsis, of which the relative clause is explanatory, according to the explanation given below. So ut in reference to time 'when' takes the indicative. In the oratio ohliqua, postquam is followed by a subjunctive for other reasons ; as, nunc post- quam cernant.. . . hah e ant quarum alterum latus Italice s i t, Liv. xxi. 30. 2. To enforce and explain what was said above on the common prin- ciple involved in the use of potential and subjunctive, it may here be noticed, that monet ut vitet is as strictly a subjunctive, as reprehendat or intelligat. Some verbs point out a circumstance, or inquire into one, and are followed by a declinable relative, coupled with a subject or not, as the case may be, and the verb in the subjunctive ; or they require to be fol- lowed by something expressing an act, which in Latin is either the infini- tive, or ut and the subjunctive, and in English the infinitive. The use of the infinitive, or ut and the subjunctive, is determined mainly by whether the dependent verb (viz. the verb so placed in the infinitive or subjunctive) is or is not the principal object of the governing verb. If the dependent verb is the sole object of the verb preceding, then the infinitive is usually employed, or otherwise ut and the subjunctive. Now, when an act is spoken of, a personal agent is implied, and if the governing verb has any object besides the governed verb, that object will be a person, which at once explains the principle of the rule on this subject, given in my tractate on the Accusative and Infinitive, page 38, § 4, " All or most of the verbs which would, after the analogy of the English, appear to be followed by a dative or accusative of the person, or by an ablative of the person, with the preposition * from' before it, and a simple infinitive, allow or require this infinitive to be changed into ut and the subjunctive." In many cases, verbs will take either the infinitive, or ut and the subjunctive, according as the personal object is intended to be prominently marked or not. The accusative and infinitive, on the contrary, is made use of, when the word in the accusative is not itself the object of the preceding verb, but consti- tutes with the infinitive that object; and hence the accusative is more generally that of the thing than the person, the same verbs taking ut and ^ I do not know how long this principle has been developed in the shape of a rule. It is laid down in a short system of rules, prefixed to the later editions of Ellis's Exercises, by the author of the articles ' Prosody ' and * Versification ' in Dr. Rees's Encyclopaedia, and now again omitted in the latest editions of that work, most probably, as being but badly adapted for elementary instruction. Doubtless the principle has been often noticed before, as it has been by myself, before seeing the digest to which I have referred ; but I do not at present recollect to have seen it any where so distinctly put forward as the foundation-stone of the whole doctrine. 108 THE INDICATIVE SHOWS THAT THE CLAUSE [CH, XIII. the subjunctive, or the accusative and infinitive, according as the accusative of the person or thing is employed. However, in English, we should also translate monet ut vitet by ' he advises that he should shun,' and hence vitet may be regarded as a potential. Ut in the sense of ' in order that,' is also usually rendered by potential signs in English. But in many of its other uses, after adeo, ita, tantus, <^c. it is otherwise. Relative clause not objective, § 75. — The rule contained in the last article will require exami- natiouj before it can be correctly understood, or otherwise it will seem to contradict fact ; for wherever the pronoun is objective is elliptical before the relative clause, the latter may be said to fill its place, and does therefore in a sense become objective ; yet scores of instances of this sort may be found where no subjunctive follows. Thus, an per liter as agere quce cogitas, Nep. in Con. 3 : quod facere voluit effecisset, Cic. pro Sext. 38. 81 : egone, qui, quod did potuit, nan dixerim, Cic. pro Plancio, 34. 85 : postquam quce voluerat dixerat, Nep. in Lys. 4 : ut quod vel corporis vel loci vel studii ratio detrahit cur a restituat, Cels. i. 2. And thus perpetually in all writers. See also instances quoted by me, under the head of Ellipsis, and elsewhere in the body of this work. On the other hand, where the assumption that is is elliptical does not explain the whole proposition, the subjunctive does then very commonly follow the relative, as in the example given in illus- tration of § 74. If the relative clause, though occurring after a verb, merely expresses the place or manner, — in other words, if it occupies the situation rather of a qualifying adverb than of an object, it is then not regarded as objective ; as, sumpsere unde cuique prowimumfuit, Liv. iii. 27: quoquo modo potuimus sustinuimus, Cic. Amic. 12. 41. It is not uncommon to meet with the explana- tory and objective clause both following the same verb, as in the fol- lowing instance, where an examination of the passage will at once point out the distinction : sed, quod magis ad nos pert i net et nescire malum est, agitamus : utrumne divitiis homines an sint virtute beati, Η or. Sat. ii. 6, 72. (a) In point of fact, it matters not at all to the use of the sub- junctive, whether the demonstrative pronouns, hie, ille, is, are omitted or inserted, the pronoun being often put in by way of emphasis, as we have elsewhere seen. Thus, neque enim est id celare quidquid reticeas, sed cum quod tu scias id ignorare emolumenti tui causa velis eos quorum intersit id scire, Cic. de Oflf. iii. 13 : sed ego hoc meis ponderibus examinabo, non solum quid cuique debeam, sed etiam quid cujusque intersit, § 75.] IS NOT OBJECTIVE.— DEMONSTRATIVE WANTING. 109 et quid a me cujusque tempus po scat, Cic. pro Plane. 32. 79 : haec cum viderem quid ag erem judices, ibid, pro Sext. 19. 43 : hoc enim ipsum utile put are, quod turpe sit calamitosum est, ibid, de Off. iii. 13. These instances are mostly cases of the potential ; but this is of no consequence. Likewise when an adjective stands in lieu of is or hie ; as, equidem audita dico quce frequentia fuerit, Cic. pro Sext. 33. 72. Or when the relative clause is made by ut and the subjunctive; as, illud ocius extorquebis, ut haec oculo contenta sit uno, Juv. Sat. vi. 53. (b) In some particular phrases, the indicative occurs, where, . according to the rule, w^e should expect a subjunctive ; as, sed hoc, nescio quomodo, frequenter in me congessisti, Cic. pro Plane. 34. 83 ; but it is to be observed here, that nescio quis, and its cases, nescio qui, quomodo, ^c. must be regarded as parenthetic phrases, on which the verb does not depend, in English, ^ but thou hast heaped on me, I know not how,^ &c. {c) It is hence evident, that if an ellipsis of the demonstrative takes place, there will still be instances in which the subjunctive is proper, if it be proper to use it where that demonstrative is expressed ; and this brings us to the development of the funda- mental principle involved therein, viz. that the mere ellipsis of the demonstrative or antecedent member does not constitute the relative clause an objective one ; for although it seems to fill the place of the object, in these cases, it is often only an explanatory or apposition clause, not absolutely necessary to the completeness of the proposition, as w^e shall presently see. Thus it happens that the very same verbs will take a relative clause after them, either with their verb in the indicative or subjunctive, as the case may be; as, qualis ess et natura montis et qualis in circuitu ad- scensus qui cognoscerent misit, Cses. B. G. i. 21, where the object to cognoscerent is qualis esset natura montis. On the other hand, quae ignorabant de L. Domitii fuga cognoscunt, ibid. B. C. i. 20, where the meaning is not ^ they learn w^hat they were ignorant of with respect to the flight of Domitius,' but, ^ they learn the fact of Domitius's flight, a circumstance of which they were ignorant,^ to which mode of interpretation quce in the plural offers no real impediment. In the following, the relative clause is merely expla- natory : h(BC enim una nos cum cmteras res tum quod est diffi- cillimum^ Cic. de Legg. i. 22. And in this both indicative and subjunctive occur after the same verb, coupled with a conjunction : et quid te ego velim, et quod tu queer is, scies, Ter. Andr. iii. 3, 4, ' you shall know what I want with you, (which as yet you know nothing about,) and a fact which you seek to know (inde- 110 DIFFERENCES OF MEANING WITH INDICATIVE [CH. XIII. pendently of my information)/ where the difference between the objective and explanatory clause is clear enough. In the following example, the principle is still more clearly marked, for we have an indicative clause in apposition with the subjunctive : habes igitur, quod ex me qucesisti, qui es s ent optimates, Cic. pro Sext. 63. 132. Here qui essent is strictly objective, and the subjunctive marks the dependence on qucesisti; since the phrase qucBvere qui erant would mean, ^ to seek those known persons who were noble,^ while qucerere qui essent means, ^to seek to know who were and who were not noble,' a sufficient difference, the first construction making qui an explanatory clause of some definite, though unex- pressed, party ; and the second making the whole clause an object of inquiry. In conformity with this we have quod quaesisti after habes, expressing a known and determinate condition. Notice here, however, that both clauses may be conceived of as depending on habet, the one in the indicative, the other in the subjunctive ; or the subjunctive clause may be regarded as filling the place of quod, and being governed by qucesisti. Again, sed recordare qui turn fuerint consulum nomine hostes ; qui non siverun t — sed a d- emerunt, Cic. pro Plancio, 35. 87. The first subjunctive marks a dependence on recordare ; the indicative, in the following clause, indicates that the second qui is merely explanatory of hostes. Let us take now the following : mirari se non sacrilegorum numero haberi qui supplicibus eorum nocuissent, aut non gravioribus pcenis affici qui religionem minuer ent, quam qui fana spoliarent, Nep. in Ages. 4. Here the qui supplicibus eorum nocuissent, by reason of the ellipsis of eos, becomes the accusative of the subject to haberi ; but that is not the reason why nocuissent is in the sub- junctive, this being due to another principle to be hereafter named, viz. the use of the subjunctive in the relative clauses, dependent ' on a preceding accusative and infinitive in the oratio obliqua. Or we may say that qui does not here relate to any specific individuals, but designates a class, and that the clause is hypothetical, requiring in English a potential sign, ^ those who should hurt.' The inde- finite relative in Greek is also followed by the optative and con- junctive. {d) In order to free this subject still further from misconception, and to show the universality of the rule, I shall now allude to another class of instances, in which the student may find some difficulty in the application of the principle ; such as the phrases, sunt qui dicant, reperti sunt qui insilirent, quod vix invenitur qui desideret gloriam, Cic. de Off. i. 19, &c. of which mention will be made presently. Here it may be said, that these verbs take no § 75.] AND SUBJUNCT. TEST OF THE OBJECTIVE CLAUSE. Ill object, but only a predicate after them ; to which I reply, that if the term objective is regarded as inapplicable, our rule must be enlarged. But the same analogy holds as in the case of objective clauses properly so called. In these last, if the relative clause is only explanatory of the elliptical object, we use an indicative ; if it constitutes the object itself, then the subjunctive; and so with predicate clauses, if the relative clause only explains an elliptical predicate, we employ the indicative, if it is itself the whole pre- dicate, then the subjunctive. Even many apparently subjective relative clauses have the verb in the subjunctive, when the whole clause constitutes the subject, and is not a mere explanation of an omitted subject. Of this description are the following: quid ea postulet pervidendum, Cic. de Fin. v. 16: deinceps videndum est quce sit hominis natura, ibid. v. 12: quid sit primum est videndum, Tusc. Quaest. i. 9. Now in all these, the relative clause appears to take the place of the subject, and yet the subjunctive is found in them, as uniformly as if the clause occu- pied the objective position. There are two methods of reconciling this apparent difficulty with the general principle. The first, by showing that the Latins, in other impersonal usages, did hold the impersonal to be the governing word, although the word or sen- tence so governed might be made the subject of the impersonal; and the second, by showing that the subjunctive, in these cases, has originated under circumstances in which the clause was objec- tive, if the first argument be disputed. Thus we say in Latin, fruendum est bono, utendum est opere, rather than bonum est fruen- dum, opus est utendum ; and the Greek verbal impersonal governs its noun likewise ; as, ττολλτ^ν δη την μετάστασιν και μεΎολτιν δείκ- τέον, Demosth. Olynth. Α. ^' : η μη^άνα άλλον αΐτιαιτίον, ibid. Philipp. Δ. ιη, which appears conclusive as to the question ; while on the other hand, if this be regarded as insufficient evidence, we have only to show that it would be correct to say, quid ea postulet pervidere, videre quae sit hominis natura, in the active con- struction, analogy teaching us, that in a change from the active to the passive structure, the whole object of the active becomes the subject of the passive; and if that object is a sentence, its gender will be regarded as neuter, or, in other words, the changed con- struction will be impersonal. The subjunctive, for the same reason, is used in relative clauses, after any other impersonal, whose personal use implies an object; as, non minori curce est qualis respublica futura sit, Cic. de Amic. 12. 43. But, in order to see the necessity of this use of the subjunctive still further, I may observe, that it is manifestly necessary to the sense. Thus, sunt -*^ 112 PARENTHETIC FOBMUL^E WITH THE INDICATIVE. [CH. XIII. qui dicunt would mean, ^ they are the persons who are known to say ;' sunt qui dicant, ^ there are persons (of that class) who say f vix invenitur qui desiderat, ^ that particular person who is known to desire is scarcely to be found ;' vix invenitur qui desideret, ^hardly a person is found to desire,' In the next example, qui being only explanatory of the subject, the indicative follows : fortes igitur et magnanimi sunt habendi non qui faciunt sed qui propulsant injuriam, Cic. de Off. i. 19. In this example, equally with the former, qui designates a class, not any definitely-known individual, but here the verb habendi has a predicate, /orie^ et magnanimi, and the relative clause belongs to the subject ; on the contrary, sunt and invenitur must be regarded as impersonal usages, and the relative clauses as predicates, not as forming any portion of the subject, any more than in the more strictly impersonal uses of the participle in dus, so clearly marked by the neuter termination in m, (e) The rule with reference to the subjunctive in objective clauses is sometimes Λdolated in dialogue or a personal address, where, in fact, the clause in question is direct, though seemingly dependent on another verb, which is in reality put in paren- thetically. But generally these are not relative clauses. I have elsewhere shown, see my Accusative and Infinitive, page 34, § 3, that this is the case after οΐμαι, εγώ μϊν οίμαι, ψο\ δοκεΐ, οίεσθαί γε χρη, βούλομαι, obsecro, oro, dico, credo, video, inquam, aio, ne time, opinor. The various combinations of nescio quis, nescio quod, nescio quomodo, ^x. are used also parenthetically, and the verb following is in the indicative. Even in other arrangements, nescio itself stands in the indicative, where other words are put in the subjunctive ; as, me miseram qucs nunc nescio, Ter. Hec. ii. 1, 8. Vide § 83. Certain adverbial combinations, as quidvis, quidlibet, for quidvelis, quidlibeat, may also be added. Nescio, however, when followed by an, takes the subjunctive. Marks of the objective relative clause, § 7 β — An objective relative clause may very often be known by the use of quid in it instead of quod^, and by the presence of some * In a former part of this work, it was noticed, that though, as a general rule, quod is used adjectively, and quid substantively, a genitive is fre- quently met with after quod. See page 50, note 5. The prevalence of the indicative, however, in examples of this sort, marks a distinction between quod and quid thus used substantively, quid being the proper § 76.] HOW THE LATINS MARKED AN OBJECTIVE CLAUSE. 113 substantive^ in agreement with the relative as an adjective, in its own clause, of which we have brought forward so many instances in the previous part of this work. Under this head are included adverbial combinations, as quojmcto, quomodo, quemadmodum, quam- obrem, qiiare, ^c. The following are examples : quid sentirent aperuerunt, Nep. in Eum. 13 : ostendam herum quid sit pericli fal- lere, Ter. Andr. v. 2, 26 : natura quid efficere possit videtur experta, Nep. in Alcib. 1 : docet quae res sit, Cic. de petit. Consul. 11 : si copies suce cognoscerent adversus quos ducerentur, Nep. in Eum. 3: videamus quatenus amor debeat, Cic. de Amic. 11. 35 : videtis quanta sit facta labes, ibid. 12. 41 : discent quemadmodum Jiaec fiaiit, ibid. : turn intelligitur quam inopes fuerint, ibid. 15. 53 : saiu sciant quomodo his utantur, Cels. ii. 18 : qum sonitum det causa videmus, Virg. ^n. iii. 584 : qu(Rrit quibus locis sit Aspis, Nep. in Dat. 4. Sometimes, however, after verbs of asking, the ipsissima verba of the speaker are given ; as, si quis nunc qucerat quo res hcec pertinet? Hor. I. Sat. ii. 23. Hence all indirect questions require the verb in the subjunctive, i. e. whenever the question, instead of being directly asked, is dependent on a foregoing verb. Con- sequently, in an indirect question, quis, quid, quam, quantus, quot, qualis, cur, quare, ubi, unde, uter, quo, utrum, an, ne, num, take the subjunctive. {a) By a reference to the former part of this work it λυΙΙΙ be seen, that I have treated this use of a substantive in the relative clause as a transference out of the demonstrative clause ; and such was, no doubt, the origin, or, at all events, we can explain the matter syntactically in this way. I do not thereby mean to have it under- stood, that the idiom of the language wOuld allow the replacement of this transferred noun, or that it can with any propriety be generally allowed. This is, on the contrary, one principal method by which the Latins distinguish an objective clause from a merely explanatory one, and the few deviations that occur are accordingly marked as exceptions ; such as, damnatum po^nam sequi oportebat ut igni cremaretur, Caes. B. G. i. 4 : ilium ut vivat optayit, Ter. Adelph. v. 4, 20 : seine me in quibus gaudiis sim, Ter. Eunuch, v. 8, 5 : audine tu ilium quid ait, ibid. v. 8, 7 : fac me ut sciam, ibid. Heaut. i. 1, 32 : qui ejus formam cognoscere studebant qualis esset. word to be used in objective sentences, or with the subjunctive. Thus, atque etiam illud ipsum, quod ac erbitatis hah et ohjurgatio, sicjnifi- candum est, Cic. Off. i. 38, where the indicative is used, because the clause is explanatory of the subject ; but ostendam quid sit pericli, see above, where the quid sit is objective. 114 OBJECTIVE RELATIVE CLAUSE^ HOW KNOWN. [CH. XIII. Nep. in Eum. 11 : exque propinquo copias, quantce et cujus generis essent speculandas, Liv. xxi. 46^ instances containing an imitation of a well-known Greek construction. In all these^ the subjunctive follows^ just as much as if the object of the first verb were made the subject of the dependent verb. On the other hand^ this trans- ference does sometimes take place, both in Greek and Latin, where the whole relative clause is not objective, and where we may with propriety restore the lost object, by taking it back from the relative clause; as, quern dederat cursum fortuna peregi, Virg. yEn. iv. 653. {b) There are two classes of instances in Latin, where the ante- cedent substantive finds its place in the relative clause, in the one of which it may be removed, in the other not ; but this generally depends on the nature of the verb preceding. Thus, dicite quce regio Anchisen qnis habet locus, Virg. ^n. vi. 669. Here the indi- cative follows, and not the subjunctive, and regio and locus may both be made objective to dicite. But in this, causasque requirit inscius jEneas, quce sint ea flumina porro qidve viri complerint, ibid, vi. 711, we could not make either flumina or viri objective to requi- rit. There is a further peculiarity in this passage, viz. that causas is put as a direct object to requirit, while the relative clauses, con- taining each a different subject, are properly such, and causas must be regarded as in apposition with them, or else the relative clauses must be regarded as exegetical of causas. In Greek, not only is the subject of the objective relative clause transferred as object to the preceding verb, but the structure of the relative clause altered, as Ιπιστάμ^θα γαρ Μυσονς ot oIkovglv, Xen. Anab, iii. 2, 23, for otl ol Μυσοι οίκουσιν. That this is the right con- struction is clear from the sense of Ιπιστάμεθα, and the use of on in the same connexion lower down. Again, Μενελεων δε ττεύ- Θομαι, εΐ νόστιμος γε και σεσωσμίνος πάΧιν, ^sch. Agam. 575. At other times, in addition to this peculiarity, another clause is made to intervene between the accusative and the clause to which it would in Latin be the subject; as, ορώντας τους στρατη- γ ο ύ ς, ot δίά πιστεο)ς αντοΧς εαυτούς ενεχείρισαν, οία πεττον- Θ ασιν, Xen. Anab. iii. 2, 8. Or this subject of the relative clause is even put in the accusative, before the clause by which it is governed ; as, τον μεν οντω φανερώς α^ικοΰντα, και πόλεις κατά- \αμβάνοντα ουδείς- πώττοτε τούτων ειττεν ως αΒ ι κ εΐ και 7ΓόλεjU0i» 7Γ ο ι ε Τ, Demosth. Philipp. Δ. ιε'. Other General Principles. § 77. — The same rule which applies to the objective clauses of verbs, applies with equal force to those of any other class of words. § 78.] OPTATIVE AFTER ώς, OTt, TN GREEK. 115 as adjectives or verbal substantives ; as, ignari quid gravitas, quid integritas, quid magnitudo animi, quid denique virtus valeret, Cic. pro Sext. 28. 60. § 78. — The objective clause in Greek does not take its verb in the subjunctive, at least not for the same reason as the Latin, even when the antecedent is in the relative clause, nor after τις. The use of oTi, likewise, after verbs of knowledge, &c. as είδέω, γιγνώσκω, with the indicative, is famiUar to every body. Thus, τί δ', ύ βον- Χηθεΐμεν £ t δ έ ν α t μη μόνον ΤΓ ο ΐ ο ί άνθρωποι ε ι σ ί ν rj ττοίοι ίΤΓΤΓΟί, άλλα και τίνες αυτών, κ. τ. λ. Plat. Alcib. i. Ill, Ε: οράς όπως ρα^ιως αμα καΐ ποίητίκώς Ι^ζίρΎασάμεΘα, Lucian, Char, et Merc. (β) The cases in which the optative is used after ώς, οτι, belong to the same class as those in which the subjunctive is used in Latin, in the oblique oration, and thus we may account for the use of 6τί and the optative, particularly after verbs of ' telling,' because in these instances the statement is that of some other person than the narrator. This practice is exemplified in the following : άρω- τίοντος δε του ΑαρεΙου 6π ο^απη ειΐϊ...όδ' αμείβετο, τίνες δε οι Ώαίονες άνθρωποι είσ ι, και τί κείνοι εθελοντές εΧΘ ο ιεν ες Σάρδίς . . . εφραζον ώς εΧΘ ο ιε ν * . . εϊη δε 17 ΐίαιονίη . . . είη σ αν δε ... ό δε είρώτα ει κα\ πασαι είησαν, κ. τ. Χ. Herod. Terpsich. 13 : μνήμην παΧαιών σπερμάτων εχουσ, υφ' ών Θάνοι μεν αυτός, την δί τίκτουσαν Χίποι, Soph. CEdip. Tyr. 1222. Notice, however, the mixture of construction at είσι in the first passage, where either the change is from indirect to direct, and the Greeks allow them- selves many liberties of this sort, or the indicative is purposely put, to distinguish the positive information here required, from the mere reason they might assign why they came, at εΧθοιεν, But the Greeks do not adopt this practice in the oratio recta, as the Latins do, and from the great frequency with which we find the direct object in Greek, where the Latins usually make this object the subject of a whole objective clause, there is reason to believe that they did not, to the same extent, regard the relative clause as objective, but rather as exegetical or explanatory. At all events, the difference between Greek and Latin practice in arrangements of this kind, seems to have some connexion with the difference of the syntax employed. The use of the subjunctive in the oblique oration after the relative, when the clause is explanatory, is to be accounted for by another rule. {b) The use, however, of the optative and conjunctive, in the oblique oration in Greek, is very much modified by the great freedom with which the writers in this language pass from the i2 116 USE OF δτί BEFORE A QUOTATION. [CH. XIII. oblique to the direct structure without notice. Matthiae, in his Grammar, has given many examples of this, see § 529, Obs, 5, of his third edition, and explains the well-known practice of introducing the on before the ipsissima verba of the speaker, in this way. Thus the indicative is found very frequently in the oratio obliqua, where the optative would be expected, and often after otl the indicative is interchanged with the optative in the same passages. Thus, ίλεγε δτ), ως ήλθε αρχήν 6 Σόλων — /cat θεησάμενος πάντα τον ίωντον οΧβον άποφΧαυρίσείε, Herod, 1. 86 : λέγων ώς εττ ι ρ ρ io υ σ ι οι "Ελληνες και α ττ ο\ α μφ ο ιτ ο συχνούς^ ibid. ix. 38. In a similar way, after on, in the sense of ^ because ;' αρρώ^εον, otl αυτοί μεν εν Έ.α\αμ7νί κατημενοι νπερ γ^ς της ^Αθηναίων νανμαχεειν μεΧΧο ιεν, νικηθεντες δε εν νήσφ άπο- Χαμφθεντες π οΧ ιο ρ κη σ ο ντ α ι, ibid. viii. 70, where Matthiae supposes that the indicative marks a determinate and real fact that would arise, the optative one that would probably arise ; though this is not necessar}^, if the freedom above spoken of is conceded. § 79. — An interrogative sentence is not regarded as objective when the direct interrogation takes place. Though, from the preΛdous analysis there seems reason to suppose, that the relative pronoun has gained its interrogative force from its use after such verbs as posco, rogo, peto, qucBVo, in a dependent clause. The syntax, however, does not bear testimony in favour of such a view. Thus, quce vero promulgata illo anno fuerunt ? quce promissa multis ? ^c. Cic. pro Sext. 30. Q>Q> : quid dicam ? qui locus orbis terrae jam non erat alicui destinatus, ibid. We have seen, moreover, already, that the interrogative quis, and the various interrogative adverbial forms, are in the oblique oration construed with an accusative and infinitive, particularly where rhetorical indignation is conveyed, so that the interrogative appears to follow its own laws. A good illustration is contained in the following : Sy. Quid est ? Ctesipho. Quid sit? Ter. Adelph. ii. 3, 8, where the real direct question has the indicative, the repeated question, depending on rogo understood, has the subjunctive expressed. Of interrogative uses in the oratio obliqua with the infinitive, I may quote quid mirum aut inexsuperabile esse ? Liv. xxi. 30 : quid periculi, quid laboris exhaustum esse, ibid. : quicquam videri, ibid. : quid agitur ? quid agatur? Terent. Adelph. iii. 3, 19. General OhservatiGns. (a) It is somewhat curious, that where, to a considerable extent, die whole sense of a passage depends upon whether we use the subjunctive or indicative, that the verb which indicates this should be elliptical, as in the § 80.] REL. CLAUSE EXPRESSING ANOTHEr's SENTIMENTS. 117 following : magnaque et comitum cemulatio, quihus primus apud principem suum locus, et principum, cut plurimi et acerrimi comites, Tacit. Germ. 1 3. Here, if we supply est, the meaning is, that those who hold the chief place, and those who have the most comrades, display great emulation ; whereas the meaning most probably is, that the emulation is called into exercise, in order to obtain these relative degrees of eminence. (δ) That the infinitive mood, as it is called, always an objective one, except when used as an abstract verbal noun, is of the same origin as the subjunctive, little doubt can exist, when the strict correspondence of its two principal forms with those of the subjunctive is considered. The t of the 3rd person imperfect and pluperfect being dropped, is the only mark of difference, and the person does not require to be distinguished, as it is either expressed before it in the objective case, or assumed from that of the governing verb. Thus there were two or more methods of expressing an objective clause ; one by using a personal clause, with a nominative subject usually introduced by a relative, or by a conjunction of the relative class ; the other by an accusative and the impersonal form of the sub- junctive (or the infinitive in other words). Thus, when two verbs are used to express one general notion, i. e. where the first expresses a modal condition, and the last the action, an objective form is used ; as, consuescit facere. CHAPTER XIV. SPECIAL RULES FOR THE USE OF THE SUBJUNCTIVE AFTER WORDS OF THE RELATIVE CLASS. § 80. — It is always necessary in using qui with its verb, to dis- criminate whether the relative clause merely states a fact, or con- tains an explanation given by the writer ; or whether it expresses an allegation of some other party ; in other words, whether or no the writer is suggesting a fact or reason of his own, or of some one else. Thus, id esse consilium Caesaris iit quos in conspectu Gallide interficere verer etur hos omnes in Britanniam transductos neca- ret, Caes. B. G. v. 6 : qui in condone palam dixerint linguam se evellisse M. Catoni, quce semper contra extraor dinar ias potestates libera fu is set, Cic. pro Sext. 28. 60 : bello superatos esse A7^vernos et Rutenos ah Q. Fabio Maximo, quihus Populus Romanus ignovisset, neque in provinciam redegisset, neque stipendium imposuisset, Caes. B. G. i. 45. Ohs. 1. As I have before stated, the same case may belong to two or three special rules. In the first example, vereretur is in the subjunctive, because it contains the allegation, not of the writer, but of other parties ; and if it did not, it might still be the subjunctive, because it may constitute an integral portion of the object of a preceding verb in the subjunctive. See § 93, Ohs. 3, below, ' to kill all whom he feared to kill in the sight of 118 SUBJUNCTIVE IN THE OBLIQUE ORATION. [CH. XTV. Gaul.' In the next example, /wmei marks the reason assigned by Cato's enemies, not by Cicero. And in the last example, though the writer here speaks himself, yet he is held to be a distinct party from the historian, it being a mere accident that the historian and speaker are the same party. 2. There are some apparent violations of this rule, {a) Where we must suppose that the writer is throwing in an explanation of his own, similar to what would be assigned by the other party, and therefore marked as such by the indicative, (b) In certain mere circumlocutions, as ea quce gesserat, usque qui prcesunt, de quo agitur, and a number of others, where the whole might be expressed by one word, or at least where such a word might be coined so as to contain the whoTe notion, were it necessary, (c) Where the allusion is to a recorded fact, even though it is introduced not as a statement of the writer, but as containing a part of the thought or sentiment of another person. Zumpt is of opinion, however, that the occurrence of the indicative in these instances is not in keeping with strict Latinity, but that they must be regarded as inadvertencies of the writers, which are not to be imitated. In the two cases quoted by him, se hand raulto post, quam jpestilentia in urbe fuerat in juventutem grassantem in Suhura incidisse, Liv. iii. 13 : legatos nuntiare jussit, Q. Fabium consulem JEquis helium afferre eadem dextra armata, quam pacatam illis antea dederat, ibid. iii. 2, it appears to me that the 'year of the pestilence,' and ' the peace made by Fabius,' forming a part of the Roman annals, might with strict propriety be referred to in the indicative if the writer so preferred. § 81. — It is in this waj that the subjunctive is so common with the relative after an accusative and infinitive ; because the relative clause in this oblique oration then so often contains the sentiments of the person who is the subject of the infinitive^ who is of course not the writer. The tw^o last examples in the preceding § 80 are in point, to which I may add, classem Lacedcemomorum quoe suhducta esset, Cic. Ofi". iii. 11: perutile esse consilium quod Themistocles ajferret, ibid. : ibi fama est in quiete visum ah eo juvenem divina specie, qui se ab Jove diceret missum, Liv. xxi. 22, ^ so the report said :' quas duo diversa maria a mp lectantur, ibid. xxi. 30 : eos ipsos quos c em ant, ibid. : non se existimare Romanos sine ope divina helium gerere ; qui tantm altitudinis machinationes tanta celeintate promovere possent, Caes. B. G. ii. 31 ; but this may also be explained by the next article, as the qui has here also some- thing approaching to the causal force. It is not, however, merely because the relative clause contains the thought or sentiments of the speaker, after an accusative and infinitive, that the subjunctive is used, but because also the case put by the speaker is a wholly general one, and does not refer definitely to any party. Thus, ad hcec Ariovistus respondit ; jus esse belli ut qui vicissent Us quos vicissent, quemadmodum vellent, imperarent, Caes. B. G. i. 36, ^ that any persons who conquered should have the rule of § 82.] REL. CLAUSE CONTAINING AN EMPHATIC GROUND. 119 those whomsoever they conquered.' This passage^, too, would be diiFerently classed by diiferent parties. Some would say that it depends on the rule given below, that with a verb in the subjunc- tive imperarent, the relative clauses essential to the sense of the whole proposition have likewise the verb in the subjunctive; others, that the relative clause following the accusative and infi- nitive, and containing the statement of another, requires this; or, without regarding either of these rules, some would say, that the condition expressed by the relative clause is not definite, but hypothetical. Ohs. When the relative clause, following an accusative and infinitive, does not contain the sentiments or reasoning of the party who is the subject of the verb that renders the construction oblique, then the indicative is of course usual. Thus, scire igitur oportet omnia legumina, qiK^que ex fru- mentis panifieia sunt generis valetitissimi esse, Cels. ii. 18 : quo minus mii^um est opus pistorium valentissimum esse quod ex frumento, adipe, melle, caseo constat, ibid. § 82. — When the relative with its clause is introduced, not as a mere extension of the principal subject, or as merely explanatory of the subject, but contains an emphatic ground or reason implied why something was done, or a defining circumstance of great importance to the sense of the passage, then the subjunctive is found in it. In the examples which follow, each of the relative clauses might be rendered by cum and the subjunctive ; they con- tain a reason why such a man would not sufier the republic to be overthrown, or why the inference of the writer had been formed ; and they give great emphasis to the statement. Thus, non est passus ille vir, qui sceleratissimos cives, qui acerrimos hostes, qui maximas nationes, qui reges, qui gentes feras atque inauditas, qui prcedonum infinitum manum, qui etiam servitia virtute victoriaque domuisset; qui omnibus bellis terra marique compressis, impe- rium Populi Romani or bis t err arum terminis definisset, rem- publicam everti scelere paucorum ; quam ipse non solum consiliis, sed etiam, ^c. Cic. pro Sext. 31. 67 : non illi ornandum M. Catonem, sed relegandum : nee illi committendum illud negotium, sed imponendum putaverunt : qui in condone palam dixerint linguam se evellisse M. Catoni, ibid. 28. 60 : quod Morini Menapiique supererent qui in armis essent, neque ad eum unquam legatos de pace mi sis- sent, Caes. B. G. iii. 28 : quod Harudes qui nuper transportati essent in Galliam fines eorum popular entur, ibid. i. 37. Ohs. In all these examples, the difference of meaning between qui with the indicative and subjunctive is plain enough. In the first, qui and its clause is not a further designation of the particular vir, but is an essential 120 SUBJUNCTIVE AFTER QUI CAUSAL•. [CH. XIV. part of the subject οι passus est, without which the proposition would have no meaning, in English, ' a man who had so distinguished himself was not the man to suffer the republic to be overthrown.' In the third example, qui is not a mere extension or further explanation of Menapii, but states a fact which explains why Caesar led his army there, in English, not because 'certain Menapii who were in arms remained,' but because 'the Menapii were in arms, and he must subjugate them.' And in the fourth, the relative clause also contains a reason why the Harudes could so easily effect their purpose, although the subjunctive might have been used on independent grounds. The meaning here is not that ' certain Harudes, viz. those who had lately crossed into Gaul, were laying waste their territories,' but that ' the Harudes, from having lately crossed over, were embracing the oppor- tunity thus afforded them of doing so.' Instances, however, do occur, where the indicative is used ; as, neque vendundam censeo quce libera 'st, Ter. Adelph. ii. 1, 39, ' nor do I think that one who is free is to be sold.' But here the individual spoken of is present, or has been so but a moment before, and the meaning may be ' nor do I think that she (pointing to her) is to be sold, a person who is free.' The indicative or subjunctive occur in examples of this kind according as the observation is general or par- ticular. § 83. — The use of qui causal, as it is termed, approaches nearly to the one just mentioned, and the subjunctive follows it in this signification ; as, accuses eum qui se pr(Ssidio munierit, non lit te oppugnaret, sed ut vitam suam posset defendere, Cic. pro Sext. 36. 78 : qui gnatum haber em tali ingenio praditum, Ter. Andr. i. 1, 71: maluimus iter facere pedibus qui incommodissime navi- ga s semus, Incert. : Laocoonta ferunt sacrum qui cuspide robur Ice s erit, Virg. ^n. ii. 231 : qui vim tantam in me et potestatem habeam tantce astutia, Ter. Heaut. iv. 3, 32 : qui non a b s ti- neas manum, ibid. iii. 3, 4 : qui me hine ecetrud as, ibid. iii. 3, 28 : saii7i' sanus es qui me id r ogites, ibid. Andr. iv. 4, 10 : quem ego igitur rogem ? qui hie neminem alium vide am, ibid. Ohs, In many cases of this kind, qui may be rendered in English by ' seeing that.' The indicative is found in some editions ; thus, infelix f qui non sponsce prcecepta furentis audierat, Virg. Mn. ii. 345, which is changed to audierit in other copies. § 84. — The tv^^o preceding rules might have been included, with some others which follow, under the more general one now to be given, viz. ^That whenever the relative clause is absolutely necessary to the sense of the whole proposition^, or consti- ^ In a certain sense, the relative clause is essential to the meaning in propositions like the following : jam faciam quod vultis, Hor. Sat. i. 1, 16 : trahit quodcunque potest, ibid. 34 ; but faciam and trahit would both take a simple word for their object, of which the quod may be regarded as explanatory. But quivis, quilihet, quod indtis, quodcunque potest^ are mere periphrases or general formula. See above. § 86.] SUBJUNCTIVE IN INDEFINITE PHRASES. 121 tutes an integral part of it, then the subjunctive is used after it.' Thus, § 85. — After an absolute negative proposition, the relative takes its verb in the subjunctive mood. In this case, the whole truth of the proposition depends on the limitation which the relative clause conveys. Thus, civis est nemo in tanto populo, ^c, qui rem ullam de meis bonis attigerit,— qui defenderit, Cic. pro domo sua, 41. 108 : inveniebatur nemo qui se suffragium de m,e tulisse confiteretur, Cic. pro Sext. 31. 68 : fuit non nemo qui agnosceret Thyum regique nunciaret, Nep. in Dat. 3. When the proposition is almost an universal negation, the subjunctive follows in like manner, and for the same reason ; as, unus est solus inventus qui dissideret, Cic. pro Sext. 62. 130. Ohs. This is also a Greek construction ; as, ov κ εστίν, οτω μείζονα μοίραν j^Eijuaiju' ^ σο£, ^sch. Prom. 291. § 86. — For the same reason, the subjunctive is used in the relative clause after the indefinite phrases, sunt^ reperti sunt, inventi sunt, answering to the Enghsh ' there are,' and some others of the same class; Sis,fuere qui cognoscerent, Sail. Cat. 61, in finem: sunt qui dicant [passim) ; reperti sunt complures nostri milites qui insi- lirent, Caes. B. G. i. 52. Ohs. This rule applies as well to relative particles and adverbs, such as ubi, unde. In the same way, non est, in the general sense ' there is not,' influences the mood of the verb in the relative clause. The following observations are copied from the English translation of Zumpt's Latin Grammar. " As the use of the subjunctive in these expressions depends on the relative's characterizing the class which is indefinitely referred to, the indicative is used if there be any thing which fixes the verb to a definite person or persons. Thus, quidam, denoting, as distinguished from aliquis, a person known, but not specified, takes most commonly an indicative. So sunt nonnuUi, sunt multi, are often used with an indicative, and must be so, where definite persons are meant. But with sunt qui, the indicative is very rare in prose authors, and its use is probably to be referred to the influence of a Greek idiom." I add an instance in which the indicative is used after sunt multi, though the parties are perhaps not definite ; as, sunt autem multi, et quidem cujpidi sjplendoris et glorice qui eri- piunt, Cic. de Off. i. 14, and also of the subjunctive after unus, where the party is indefinite ; as, quidam ex Gallis unum e Romanis qui e s s et optimiis provocavit, Eutrop. ii. 6. Valpy, in his edition of Cicero's Offices, has the following note on the passage, sunt enim qui, non audent dicere, Cic. Off. i. 24 : "In the present instance, and in others similar, where sunt qui is followed by a verb in the indicative, Ernesti has altered it into the subjunctive. I. F. Heusinger asserts that this has been done against the authority of MSS. : that where the matter predicated in the sentence is simple and absolute, without doubt or reserve, the indicative is properly employed. In the present book, 30, occurs the following 122 SUBJUNCTIVE AFTER QUIS EST QUI ? [CH. XIV. passage : sunt his alii multum dispares, simplices et aperti qui nihil ex occulto, nihil ex insidiis agendum put ant ; veritatis cultores, fraudis inimici ; itemque alii, qui quidvis ρ er ρ eti antur, cuivis deser- V i ant, dum quod velint consequuntur. Tn the first limb of this sentence, pufe?it would be as improper as in the second, perpetiuntur and deserviunt. Alterations of this nature, in conformity to the grammatical notions of copiers and editors, are more difficult to effect in poetry. Thus, Lucre- tius iii. 115, est aliud quod agitatur et accipit: ibid. vi. 703, sunt aliquot qua rum dicere non satis est: Horat. Od. i. 7, 5, sunt quihus opus est.^' Thus far the editor referred to above. I may remark, however, that in three of these examples alius or aliquot is inter- posed, which, taken in their connexion, as in the case of nonnulli and multi quoted above, often imply a more specific reference than the simple phrase sunt qui. The use of the subjunctive perpetianiur above, is, by Valpy's own showing, dependent on other considerations than those involved in the mere use of sunt qui, and it will not therefore be thought that I am con- tradicted by his example, where after alii the subjunctive does occur. The first alii may have had a definite reference, and the second have been wholly general. I only remark, that when alius is used, this reference is very often definite, and the indicative is used accordingly. § 87. — The relative is always put with the subjunctive after the expressions quis est ? quid est ? used interrogatively ; as, g μ i 5 est^ qui se, cum contra me ferebatur, inisse suffragium confiteatur? quis est, qui non ρ r of ι teatur se affuisse, ^c. ? Cic. pro Sext. 51. 109: quis fait qui non arbitraretur, ibid. 50, 108: quis esset tantus fructus qui g auderet, ibid, de Am. 6. 22: quce enim domus tam stabilis, quce tam firma civitas est, quae non possit everti, ^c. ? ibid. 7. 23: quis est in quo non fixum sit? ibid, pro Sext. 60. 128. Ohs. And thus continually in the same author. This rule, of course, applies to all the inflexions of the phrase quis est qui ? In Greek we use the indicative ; as, τις J3' άτρωτος όστις ουκ άποίσεται, Eurip. Phoen. 594; but also sometimes the optative, with or without αν, as Plat. Euthyd. 292, E. 293, A. To prevent mistake, it may be observed, that the rule holds good only when the phrase quis est qui is wholly in the interrogative member ; as a question is sometimes asked by quis est, and answered by qui in the redditive member, which is of course a different case. Thus, vir bonus est quis ? qui consulta patrum, qui leges, juraque servat, Hor. Ep. i. 16, 40. § 88. — The relative is put with the subjunctive in cases where it is englished by ^ to/ ^ as to/ in the phrases dignus qui, idoneus qui, ^c. Instances of this sort are easily understood; for, in a general w^ay, if we render by the word ^ that/ we require to use the potential in English. Thus, indignus qui possideam, Hor. Sat. ii. 3, 236 : qui ilium decrerunt dignum suos cui liber os committer ent, Ter. Hecyr. ii. 1,15: nihil est Thaide liac tua frater dig nius quod ametur, ibid. Eun. v. 8, 21 : dig nam me putes quam illudas, ibid. Heaut. iv. 4, 19 : idoneus tibi § 91.] SUBJ. AFTER COMPARATIVES WITH QUAM QUI. 123 videor esse quern tam aperte fallere incipias dolis, ibid. Andr. iii. 2^ 13 : adeoTi' videamus vobis esse idonei in qui bus sic illudatis, ibid. Andr. iv. 4, 19 : Pompeius i done us non qui impet7'et, Cic. pro leg. Manil. 19. 57. § 89. — So also after is put for talis, or after tam ; as neque enim tu is es, qu i, qui sis η e s c i a s, Cic. ad Divers, v. 12 : se enim eum esse dixit qui cog ere t non qui expugnaret, Nep. Ages. 5, In Greek, the indicative is used in the same circumstances ; as, ουκ ίστιν ούτω μώρος, ος θανείν Ιρα, Soph. Antig. 220, ^ no one is so foolish as to desire to die :' τις ουν όντως ΙστΙ δυστυχής όστις βονΧησεταί ; Dem. de Class, θ' : but with the optative in the fol- lowing, ουδείς όντως Ύ]\ίθι6ς Ιστιν όστις ονχ Ικανον Soirj, ibid, η, and in examples quoted above. Obs. This rule is of some extent, and comprehends a variety of cases, where the force of the English ' that,' or ' such that,' may be detected. On the other hand, is, put for talis, is sometimes followed by the relative and indicative, where the force of the English ' as ' is intended to be con- veyed ; as, itaque ego i s in ilium sum quern tu me esse ν i s, Cic. Attic, vii. ep. 8 ; and so in the following, where it has a causal force, heu me miserum qui tuum animum ex animo spectavi meo, Ter. Andr. iv. 1, 22. § 90. — The subjunctive is used after comparatives with quam qui, in all its cases or inflexions ; as, major sum quam cui possit fortuna nocere, Ovid, Met. vi. 195. Obs. This, like its predecessors, comes under the more general rule of the relative clause being essential to the sense of the proposition. To say ' I am too great ' would be unmeaning ; but the proposition becomes complete when it is added, ' to be hurt by fortune.' § 91. — Qui after quippe is followed usually by the subjunctive, though we then render the passage by ut, in the sense of ' as that :' as, nihil attinet earn ex lege consider are, quippe quae in lege scripta non sit, Cic. de Invent, ii. 45. Obs. The expressions quippe cum, quippe uhi, quippe ut, are employed in the same way. Plautus, however, uses an indicative in similar cases, and Sallust occasionally does so too ; as, quippe cui omnes copice in usu quotidiano, et cultu corporis er ant. Cat. 48. Again, with ut qui we have the indicative or subjunctive, according to the sense ; thus, u t quibus esset persuasum, Cses. B. G. v. 31: ut apud quos plurimum hiems occupat, Tac. Germ. 22, where the first contains a plea why, the second a fact that. In the obhque oration, quippe, like some other words named above, is put with the accusative and infinitive ; as, quippe duos pro uno domino acceptos, Liv. iii. 9. In Greek, we render the quippe qui of the Latins by the dative of the participle; as, καλώς ελεξεν ευλαβον- μενω •πεσεΊ.ν, Soph. GEdip. Tyr. 597. 124 SUBJ. AFTER QUI FOR UT AND QUOD QUIN. [CH. XIV. § 92. — The relative and the subjunctive are often used in place of the English ' that ' with the potential, or of the simple English infinitive present; as, qui cognoscerent misit, Caes. B. G. i. 21 : hinc demus quifruatur, Ter. Adelph. v. 8, 27. Ohs. But, even in cases of this kind, an indicative occurs, when the writer chooses to look rather to the fact than the intention ; as, certos mittit homines ad infimos monies qui ohvii erant itineri adversariorum^^ Nep. in Epam. 9: eliguntur principes qui jura reddunt^ Tacit. Germ. 12, where, however, the meaning is perhaps merely, ' that the princes are elective, and that the dispensing justice is entrusted to them.' § 93. — In the same way, quin and quod^ though not strictly resolvable by ut, both contain in themselves the force of the English ^that/ and are followed by a subjunctive. Quod, in the sense of ^because/ is not necessarily followed by a subjunctive; but in many cases this force seems to obtain where it is so used, because a causal may still be necessary to the sense of the whole proposition. Thus, after verbs of praising, blaming, or accusing, the relative and subjunctive follow more particularly, inasmuch as they contain in themselves the terms of the commendation or charge, not the reason assigned by the writer. Quin is evidently a composition of qui non, and quod, signifying ' that ^ or ^ because,^ is, as elsewhere seen, of the same origin as the neuter of the relative pronoun. In fact, the demonstrative pronoun is often found pre- ceding it, as in the examples subjoined : in eo reprehensus, quod ex prceda tripodem aureum Delphis posuisset, Nep. in Pans. 1: in eo reprehendit, quod prceter modum dicat esse, Cic. pro Plancio, 33. 82 ; for although eo is usually regarded as redundant, and standing in apposition with the whole relative clause, there is no doubt that in the original mode of viewing the construction it was demonstrative to quod as the relative pronoun, but afterwards omitted or retained at pleasure, when quod came to be regarded more as a conjunction. Quin, with its verb in the subjunctive, is found after propositions containing or implying a negation, or expressing an obstacle or doubt. Hence it is common after dubito, nullus est, via? reperitur, quis est ? quid causae est ? eo inficias, and where intervention is implied. The following are examples : non dubitabat quin ei ere- deremus, Cic. Attic, vi. ep. 2 : non cunctandum eosistimavit, quin ^ This passage is perhaps, however, to be understood differently, by making ohvii agree with monies, ' which lay in the route of the enemy,' though ohvius esi has the meaning of ' meet' in other passages, as quicmiqiie obviiis est, Hor. Sat. ii. 6, 51 : ohvius ei fuit Emphyletus, Nepos, Phoc. 4. § 93.] QUOD WITH SUBJ. AFTER A SUBJUNCTIVE. 125 pugna decertaret, Hor. Sat. i. 1;, 20 : nullum intercedebat tempus quin extremi cum equitibus pr cellar entur, Caes. B. G. i. 70 : nee suf- fragatio horum valuit quin prceferrent, Liv. iv. 44 : numquid vis quin abeam ? Ter. Adelph. ii. 2, 39. The Greek μγ] ov has often the force of quin, with this difference^ that it may be followed by a participle, while quin requires the subjunctive. Hence it follows any word of negation or deficiency ; as, ηκεις γαρ ου κεν // γε, τουτ εγώ σαφώς ίζοί^α, μη ο if χ\ δεΐ /i' Ιμοί φέρουσα τι, Soph. CEdip. Col. 355, ^ it cannot be but that you are come bringing some terror to me.' The following will illustrate the use of quod the conjunction: hie quum ab Agnonide aeeusatus esset, quod Pirceum Nicanori pro didi sset, Nep. in Phoc. 3: et ante omnes Q. Servilio Priseo, quod non degener asset ab stirpe Claudia, collaudante juvenem, Liv. iv. 48 : erhninabatur quod Titum filium ab hominibus rele g asset et ruri habitare jussisset, Cic. de Off. iii. 31: questum quod popular entur, Caes. B. G. i. 37. Thus, too, after an equivalent to a verb of accusation; as, qui M, Postumio et T. Quinetio diem dixerant quod pugnatum esset, Liv. iv. 40. And likewise after a substantive implying a sentiment ; as, tamen illo dolore, quod Romce non sis animum tuum libera, Cic. Ep. ad Divers, vi. 1. With the expressions ago gr alias, habeo gratias, we very commonly meet with an indicative after quod, the proposition in these cases being complete and true in itself without the causal. Ohs. 1. The subjunctive or indicative is used, according as the relative clause is essential to the complete sense or not, or according as the writer intends the relative clause to contain a mere explanation of the cause or the terms of the praise, blame, accusation, or complaint. The examples quoted with the subjunctive will fall under the general rule of objective relative clauses, or the more special rules, § 80, § 81, and § 84. 2. Quod, in the oratio obliqua, is sometimes followed by an accusative and infinitive ; as, quo d si quid ei a Ccesare gravius accidisset, η e mi- ne m existimaturum, Cses. B. G. i. 20. This I have attempted to explain above, § 38. In the sense of ' as to that,' it is not put with the subjunctive, unless one of the special rules applies to the case ; as, quod me ad consuetudinem revocas, fuit meum quidem jampridem remjoublicam lugere, Cic. Ep. ad Attic, xii. 28. 3. There is another class of instances in which quod, ' that,' takes the subjunctive, which belong to the special rule immediately following, but must be noticed here for the sake of systematic arrangement. This takes place when quod and its clause follows a verb in the subjunctive mood, as one of the particulars included in the subject of that verb, or rather, perhaps, where the relative clause is itself the subject of the verb, though an apposition word, implying the subsequent clause, is expressed with the previous verb. Thus, quum ad has suspiciones certissimce res accederent ; qu ο d per fines Sequanorum Helvetios transduxisset; 126 SUBJUNCTIVE BY ATTRACTION. [CH. XIV. quo d obsides inter eos dandos cur as set ; quod ea omnia non modo injussu suo, et civitatis, sed etiam inscientibus ipsis, fecisset; quod a magistratu JEduorum accusar etur^ Sfc. Cses. B. G. i. 19. § 94. — When the relative clause is dependent on a verb in the subjunctive mood^ and is essential to the sense of the main pro- position^ then a subjunctive is found in it ; as, delude qumn, quid- nam id esset, quod i^espicere ν e tit us esset, agitaret animo, Liv. xxi. 22. See also the instance immediately above. Obs. 1. The English language does not require this species of attraction, and hence its adoption in Latin constitutes a ground of difficulty to the young student. As in a previous case, however, there are some exceptions, the indicative being used in certain customary circumlocutions, where a single word of equivalent meaning might be capable of standing instead. See § 80. Obs. 2. (b). But if the verb in the subjunctive, on which the relative clause depends, is the subjunctive after ut, preceded by tarn, ita, adeo, then the indicative occurs in the relative clause ; as, jam vero ita faciles aditus ad eum privatorum, ita liberce querimonice de aliorum injuriis esse dicuntur, u t is, qui dignitate principibus excellit, faci- litate par infimis videatur, Cic. pro leg. Manil. 14. 41 : Asia vero tarn opima est et fertilis, ut et ubertate agrorum, S^c., et multitudine earum rerum, qu ce exportantur, facile omnibus terris antecellat, ibid. 6. 14 : atque ita Quirites, u t hoc vos intelligatis a nullo istorum q u i huic obtrectant legi, ibid. 8. 21. 2. A similar kind of attraction exists in Greek; as, γόνιμα ποιητην l•" αν ουν evpoLQ 'in ζητών άν, όστις βήμα γενναΐον Χ ά κ: ο ι, Aris- toph. Ran. 96. See Matthise, Gr. Gr. § 527, β. § 95. — In nearly all the cases here enumerated, it will be found on examination, that the relative clause is indispensable to the sense of the v^hole proposition to which it belongs, when the sub- junctive is used in it, and in other cases not specifically enu- merated here, the rule may be applied with safety. There are a few cases of a different kind yet to be mentioned. § 96. — The verbs dico, puto, arbitror, are often put by Cicero, and occasionally by other writers, in the subjunctive mood, by a kind of attraction, where this mood properly belongs to the verb, which expresses what was thought or said; as, sed ubi Consulem ad tantum facinus impellere nequeunt, ipsi singillatim circumeundo atque ementiundo qu(2 se ex Volturcio aut Allobrogibus audisse dicer ent, magnam illi invidiam conflaverant, Sail. Cat. 49. Had the writer not used dico, he would have said properly quce audisset, Λvhich, by reason of the subjunctive, would have meant, ' what he said he had heard ;' but he actually introduces dico, and puts this in the same mood, while, by reason of this change, the word that would have been in the subjunctive is put into the structure of the accusative and infinitive, and turns over its mood to dico. See Zumpt's Lateinische Grammatik, 8th Ausgabe, § 551. § 100.] SUBJUNCTIVE USED OF AN ACT REPEATED. 127 § 97. — As the verbs volo, nolo, malo, are often put in the present subjunctive, by a mere idiom, as well as some others in the present and perfect subjunctive, merely as a modest qualification, the use of the relative in such connexions is not peculiar. The use of the perfect, to which I refer, is one of extremely common occurrence in our own language, in the phrase, ^ nor would I assert.' Thus, too, in Latin ; nee tamen affi rmaverim nullam Germanics venam argentum aurumne gignere, Tacit. Germ, b : et ut corpora lente augescunt, cito extinguuntur, sic ingenia studiaque oppresseris facilius, quem revocaveris, ibid. Agric. 3, where the first is sub- junctive for indicative, the second perfect potential for present potential. § 98. — The subjunctive is used, too, in relative clauses, where the writer intends to convey the idea of an action frequently repeated. This rule is, however, in all probability, identical with one of the preceding, because the relative does not in these cases refer definitely to any one case, but to all which might have taken place, which not being always known, renders the relative clause in a measure hypothetical. Thus, eo postquam Caesar pervenit, obsides, arma, servos, qui ad eos ρ e rfugi ss ent, poposcit, Cass. B. G. i. 27. It is here implied that Caesar was aware that this desertion had been going on, but he was probably not aware of the extent of it, nor did he know definitely when it had been practised, § 99. — In the direct rhetorical question, which, from the nature of the case, frequently occurs in the oratio obliqua, and where no answer is expected, the subjunctive is used; as, quidnam id rei esset, Liv. iv. 44. CHAPTER XV. CONCLUDING REMARKS. § 100. — I have noticed elsewhere, that the relative with the sub- junctive may be regarded as the type or pattern by which the use of the subjunctive is rendered intelligible in other cases. A large number of the conjunctions are of relative origin, or are mere relicts of older forms of expression, in which ellipsis has taken place, or declinable forms have become fixed and adverbial. Hence depen- dent sentences, with an, ne^ num, ut, cur, quare, quomodo, ubi, unde, ^c. all take the subjunctive, while in direct sentences they take the indicative. Among direct propositions, we must reckon ut when 128 CONCLUDING GENERAL REMARKS. it is a particle of ^ similitude^ or ^ time ;^ but when it signifies ^ that/ ^ in order that/ it marks an objective proposition. The adverb quum, or cum, very often relates to time^ and takes the sub- junctive too, but the use of the subjunctive in such cases marks that it is necessary to the sense of the proposition ; as, quum Jkbc accepta clades esset jam C. Horatius et T. Menenius consules erant, Liv. ii. 51. Here the object of the writer is not to specify the time when these parties were consuls^ but, on the contraiy, to note the time when the overthrow took place, and the historical connexion. § 101. — In Greek, the peculiar power and use of the participle enables us to dispense with the use of the relative in many cases, where in Latin we have no other alternative. § 102. — In conclusion, it may be observed, that there w^ill of course be found occasional deviations from fixed rule, where no very satisfactory reason can be assigned for the use of one mood in preference to another. Sometimes the one, and sometimes the other will be proper, according as the writer wishes more to insist on the fact itself, than another person^s adducing that fact ; that is to say, the relative clause may contain the sentiments of another, and be at the same time a fact of the writer's knowledge, and it may be presented either as the one or the other ; as in the fol- lowing, nam si cum his copiis, quas secum transportaverat, interiisset Darius, non solum Europam fore tutam, sed etiam eos qui Asiam^ incolerent, Nep. in Milt. 3 : or occasional inaccuracies will occur QYen in Latin writers themselves, at variance Avith the usual practice ; or, lastly, a potential force may be intended by the writer, where our own language more commonly uses an indicative. There is considerable difficulty, moreover, in classifying examples, while it often happens that the same instance may be ranged under two or three special rules, all containing particular cases of a general principle, but which it is nevertheless important to pre- serve and set forth, from fear of falling into too great generality, in which there is some danger of the young student^s not being able to follow out the rule prescribed to him. APPENDIX. (α) The grave accent has been purposely left over the last syllable of the last word of several of the Greek examples quoted in this work, to indicate that the quotation is incomplete. (/3) page 4. § 1, Obs. 1. It might be argued, that, ϊΐ currebat were the verb belonging to the relative clause, it would be in the subjunctive like ferret, and that, therefore, it is the verb in the relative clause that is elliptical, curreret. To this it may be objected, that it causes great con- fusion in the disposition of the two clauses, and that it is preferable to regard qui as the subject of currebat. In effect, the two clauses may be viewed as melted into one, and as the fact that he did run is intended to be conveyed, the indicative is properly used, while at ferret the ordinary construction appears. That this is correct, will be seen from the great awkwardness of supposing the construction to be scEfe currebat velut qui fug i ens (for fugeret) hostem ; perscepe velut qui Junonis sacra ferret, where the participle does not correspond with the finite mood. (γ) page 9. § 4, Obs. 2. The following Greek example may be added : ^ΑτρειΕης ^' άν ομιΧον εψοίτα, ε'ί που εσαθρησειεν ^ΑλέΙαν^ρορ θεοειΗα, Horn. II. γ. 449. (^) P^g^ 16• § 19• Similar to the Greek example here given, in which the subject and predicate appear to change places, is the subjoined Latin one: huic autem qui studeant sunt nulli, Cic. Off. i. 37. So ulla esse potuisset, ibid. ii. 3 : homines certe fuerunt, ibid. ii. 4 : nulla esse potuisset, ibid. (ε) page 18. § 19, (b). There is another remarkable use of the direct for the relative structure in St. John, though different in several respects from the one occurring c. xv. v. 6. I allude to the one c. v. vv. 37, 38 : " And the Father himself which hath sent me, hath borne witness of me. Ye have neither heard his voice at any time, nor seen his shape, and ye have not his word abiding in you, for whom he hath sent, him ye believe not." The meaning, obviously, appears to be this, ' And the Father himself which hath sent me, hath borne witness of me, wh ο s e voice though ye have neither heard, nor seen him in a bodily shape, yet ye have not his word abiding in you, because ye believe not him whom he hath sent ;' i. e. their determined unbelief prevented their receiving the fact of his divine mission, of which the evidence was abundant. This must be the meaning and connexion, though it is possible that our Lord had in view at the time the passage, Deut. iv. 12, " Ye heard the voice of the words, but saw no similitude, only a voice." Compare also Heb. i. 2. A somewhat different case, though still referable to the same principle, is Κ 130 APPENDIX. met with Rev. i. 16, η 'όφις αντοΰ ώς δ ήλιος φαίνει εν τγ δυνάμει αντοΰ, where we might say that there is an elhpsis of the relative adverb and pronoun, ' his countenance was as the sun when he shineth in his strength,' but where our translators have faithfully imitated the original, as usual, even to the form of construction. This is not, however, a case of ellipsis, since it is manifest, by comparison with other passages, that the (^αίνεί is put in lieu of the participle ^atVwr, just as in the example quoted page 16. § 19, {a) of this work. It is therefore properly a transition from the participle to the finite verb. At Rev. xiv. 13, there is another example of the direct construction, where resort must be had to some similar explanation. (0 P^g^ 1^• § 1^' ip)• ^^• P^ul uses the article for the relative pro- noun, in this verse, τω τον φόρον, τον φόρον' τ ω το τέλος, το τέλος' τω τον φόβον, τον φϋβον' τω την τίμιων, την τιμήν, ' tribute to whom tri- bute,' 'honour to whom honour,' Rom. xiii. 7. There is here, more- over, an attraction in the first accusative of each member, though it is possible to explain the passage differently, and I am inclined to think that it would be better to regard τω as the article preceding a participle sup- pressed, whose meaning is that of ' having a claim to,' which governs the first accusative, the second accusative being in apposition with όφειλάς, or governed by απόίοτε in the former part of the verse. Could we justify the use of οφείλω as an active verb, in the sense not of 'owing,' but of 'having something owed,' this explanation would be quite satisfactory. (η) page 23. § 19, Obs. It should have been said in four different ways, for we have δς μέν followed by 6 ^έ, Rom. xiv. 2. The reverse of this sometimes takes place. In further corroboration of the views con- tained in this Obs., it might also have been noticed, that Homer uses 6 for 6ς, where the absence of the conjunction ^ε or τε would seem to intimate that the construction is really relative, and not demonstrative, and where the terminal ς is left out for the sake of the metre. See Hom. II. γ. 351, 354, also οτις for όστις, τ. 265, and elsewhere οτε for οστε, οττερ for οσπερ. (θ) page 24. § 19. That the spiritus asper was often equivalent to what is termed the digamma, may be shown by the Latin orthography of frater, frango, of cognate origin with some Greek words commencing with p, from the fact that it often renders hiatus admissible in Homer in a similar way ; and from its being, in some Latin relative forms, represented by g- or qu, the modern w, as may be seen in 'whom,' ' when,' &c., in correspondency with which are Gualterus, Gulielmus, Walter, William, and the pronunciation of gu in Spanish. Often, however, the spiritus asper had more of the sibilant sound, as may be seen in t, se, εζ, sex, ετΓτά, septem, υπό, sub, νττέρ, super, νλη, sylva, 8^c. This again was softened into a mere aspiration, which is made to represent it in modern derivatives. It is, however, probable that it was also sometimes repre- sented by a sound intermediate between the harsh sibilant s, and the softer breathing h, such as t. This is seen in τοί for σοι, θάλαττα for θάλασσα. In correspondency with this we have ης, Latin quis, όστις, quisquis, τε, que. This will suffice to show, that the attempt to connect 6, ορ, τός, hie, qui, even etymologically, is not far-fetched. Thus, then, it has been established, that ο is used for article, pronoun, and relative in Greek ; that ό'ς is used for relative and demonstrative, and coupled with μέν, 3έ, for δ μέν, δ ^έ : and that h, q, c, are frequently only marks of a breathing, either preceding some unsupported vowel-sound, or following it as some- thing upon which to rest. It might not be difficult to connect is, αυτός, APPENDIX. 131 οντος, with the same family, though this would be beyond the scope of the present treatise. (i) page 25. § 19, Obs. at the top. Homer uses the relative form at the beginning of a period, where the demonstrative which should precede it is postponed, as in the following passage, though the use of γάρ shows that it should be regarded as continuous with what goes before, at γαρ νπ ήελίω τε καΐ ηυρανώ άστερόεντι ναιετάονσι ττόληες εττιγβονίων άνθρωττων^ τ άω ν μοι ττερΧ κήρι τιέσκετο'ϊλως Ιρη, II. ^'. 44, which example also illus- trates the subject treated of page 39, § 25, (b), and is of like kind with II. σ. 429, exhibiting transposition from the case of the partitive genitive τάων^ to that of the relative, with which the noun so transferred is made to agree. A curiously abrupt case with όστις inceptive occurs, II. ^'. 240. Homer also uses τοί, ταί for the plural of the pronoun, αλλ' ήτοι ταΐ, νόσφι καθήμεναι, II. ^'. 9, In a great many instances, τοί, ταί plurals nominative are followed by ^ε, γάρ, γε, showing that the structure is demonstrative or direct, ' they,' ' these,' but in others again they appear to be used exactly as relatives. Thus τοί is followed by ^ε, II. α. 447. β'. 52. 149. 151. ν 883. β. 361. 375. σ. 546. 571. ψ. 202. 212. 369. 449. 847. In χ, 254, it is followed by γάρ. In φ'. 608. ψ'. 217, it is followed by γε : and in β'. 52 it is preceded by ol μεν, in a partitive sense, similarly to the example noted Pind. Olymp. xiii. 58, only the order is reversed. In v. 87• p'. 145. v. 299. (p'. 267. χ, 53, it is used for the simple relative, and in β', 346. V. 308, it is followed, like the relative pronoun, by κεν, and in & . 225 by pa. In these cases it will be seen on examination, that το\ ^έ usually occurs in the initial position, and may be regarded as a demonstrative pronoun. The two examples of τοί γε occur where otherwise hiatus would take place before the r. The employment of τοί by itself for the relative pronoun occurs in every instance where hiatus would otherwise take place, or a short syllable requires to be lengthened by position, which is also the case with τοΙ κέν, II. /3'. 346. In the remaining instances, V. 308 with Key, & . 225 with pa, it occurs after a short csesural syllable. These facts seem to indicate clearly, that the τοί form was not unfrequently preferred for the sake of the metre. Another circumstance deserving of notice is, that not only is the τός form used where we translate by the relative, but there is reason to believe that the so-called relative was itself a demonstrative pronoun. Hence we occasionally meet with it in what may be termed its transition state : Ζενο, ο σ τ- ' ανθρώπων ταμίης πολέμοιο τέτνκται, Horn. II. l•' . 84, and again η . 208. This οστε is of kindred nature with ο Si, which is often found where the relative might be put, and which I have elsewhere spoken of, page 25. § 19, Obs. This use of ό ^ε' also might be given as exemplifying transition from the relative to the direct or demonstrative construction, which is well seen in the following : 6v pa ποτ αυτός, νπο στερνοιο τνχτισας, βεβληκει προς στήθος, ο ^' ύπτιος εμπεσε πέτρτ], II. ^'. 106, where, too, notice the repetition after parenthesis, not uncommon in Homer. As it will be more satis- factory to present at one view the way in which these forms of article, demonstrative, relative, and partitive are used, I have supplied the fol- lowing table, which contains the sum of what has been embraced in the Obs. in the body of the work, and in this Appendix : — 132 APPENDIX. (1) ό, 'ό (2) οδε, 6 f ί (3) ος (4) υσ-ις, οης used as article in the more perfect state of the language, but as pro- noun demonstrative* and relative in Homer. as a demonstrative pronoun, and the latter as equivalent sometimes to the relative. as a relative pronoun, hut also as a demonstrative, in which capacity it occurs in particular phrases in Attic. used for one another. (5) οστε, οτε used for one another, and as simple relative, tending to prove that the og was originally a demonstrative. (6) οσττερ, οττιρ (7) ov, τοϊι, and so through all cases. (8) τοί, ταί, nom. ρ1. (9) 6 μεν, . . .6 δε (10) 'όςμ'εν.. ^οςδε (Π) δ μεν . . . .ος δk (12) δς μεν .. .6 δε (13) οι μεν. . . .τοι δε used in like manner for each other. for each other, sometimes even in Attic. used as demonstrative and relatives in Epic and Doric. (14) rot μεν . .οι δε In English, ' the one,' ' the other,' ^ this,' ' that,' 'he,' and also the δς ^έ and 6 δε sometimes nearly the same as 'who.' They are gene- rally used both together in two contrasted members of a sentence, though sometimes one or the other is wanting, as the case may be. No. (9) is most usual in the best writers, and was succeeded by No. (10). The others (11), (12), (13) and (14) are only met with occa- sionally, and (13) and (14) only in the earlier writers, as Homer or Pindar. The latter as mas- culine, either wholly > or chiefly in the Ho- meric and earlier dialects. ' There is Httle doubt J that the use of ή μεν, 1 ή δε, as simple con- i jimctions, has ori- ginated in the same way, though no long- er decKnable. See II. δ'. 258. * That the ό, η, το, of Homer, is a demonstrative pronoim, is, beyond all doubt, in a passage like the following : δια ττρό ^έ ύσατο και της, II. δ'. 138. (fc) page 39. § 25, (c). A very singular example, and different, I believe, from any yet quoted in the course of this work, is the annexed : ne qucere doceri, quam pee η am, aut quce forma viros fortunave mersit, Virg. ^n. vi. 614. Heyne here explains by supplying expectan- dam haheant, and no doubt this may represent the construction, though I much prefer not having recourse to this mode of explanation, which detaches quam, pcenam altogether from the verb mersit belonging to the relative clause. The fact is, simply, that doceri would be followed by poenam, with which quce, that would have followed as the subject of mersit, has become incorporated, and made to agree, so as to form with it the object of doceri. The ordinary and correct Syntax is employed in the second member beginning with aut, and this, in some measure, covers APPENDIX. 133 up and softens the faulty Syntax of quam poenam mersit, which is here, notwithstanding, actually employed. The case will, therefore, fall under the head of 'Attraction of the Relative into the case of the Antecedent.' (λ) page 55. § 33. Siiuilar to the Latin example, ^n. i. 498. vi. 205, is the following in Greek, though the τός, and not the τοΊ,ος form is used ; oio ν ^' αστέρα ήκε Κρόνου παις αγκυλομητέω, .... τ ο ΰ ^ ε τε ΤΓολλοί αττό σπινθήρες 'ίενται' τ ω εΙκυΤ ηϊ^,εν, κ. τ. Χ. II. ό'. 75. (μ) page 60. § 34, (α). Several good instances in proof of this ten- dency in οίος forms, to agree with a noun in their clause, are here exhi- bited with τοΊος, though this practice is less frequent with the demonstra- tive than with the relative form ; ώς Β' οτε τις (five lines inter- vening) τοΊοί τοι, Μενέλαε, μιάνθην α'ίματι μηροί ευφυίες, Horn. II. ^'. 141 to 146. And again, Β'. 275. This might be given as exhibiting the pronominal for the adverbial use, or as a particular case of Attraction, but, I conceive, that the most comprehensive mode of regarding it, is that which refers it to a principle such as I have attempted to establish. Again, τεττί- γεσσιν εοικότες . . . . τοΊ ο ι ηγήτορες ηντο, II. γ'. 151. τ ο ί η οΊ επίρροθοί ήεν Ά θ η ν η, Β'. 390. ο ϊ η ^' εκ νεφεων ερεβεννη φαίνεται άηρ.,.,τοίοι, γάλκεος "Α ρ η ς, ε'. 864. οι ό ς τε πελώριος ερ-χ^εται "Αρης 6 στ' εισι . . . . τ οίος αρ' Α 'ία ς ώρτο, η. 208, where notice also οστε, in the sense of * who ;' τ ο Ίο ς γαρ γαιηογος Έ ν ν ο σ / γ α ι ο ς ωτρυν' Άργείους, ν'. 677. οί ^ε, λύκοι ώς τ υ~ι ο ι η γ η τ ο ρ ε ς, π'. 156. τοΊορ γαρ πόθεον μηστωρα φόβοιο, φ'. 16, in all of which instances it has the force of ούτως. These examples are of similar kind with the qiialis and οίος con- structions, given in the body of the work, and differ from those in which talis occurs after qualis, in Latin. In English we always make this con- struction by the adverb, and not by the relative or correlative adjective, and in Greek and Latin it was principally retained with oloc and qualis. Thus in the Odyssey, where the language has undergone some perceptible change from that of the Iliad, I do not think that any strictly analogous examples are to be met with, at least, they must be very rare, from the examination I have made, without, as yet, meeting with any. This agrees well enough with what might be expected from the facts already ascer- tained. The further we proceed back in the history of the language, the more reason we have for believing that the relative and demonstrative did not exist separately ; and thus, while we find in Homer the 6, ος, and τός form interchanged, so οίος and -ο7ος are more nearly allied, than in an after period, when a wider distinction took place between them, the latter bearing a construction which was, in later times, more characteristic of the former. (v) page 61. § 34, (6). This use of ος after τυ~ιος is Homeric, and is met with II. ω. 182. Odyss. δ'. 776. V. 134. ψ'. 282. (ζ) page 64. § 36, Ohs. 4. It has been suggested already, page 58. § 33 (k), that in the use of qui after tarn, the tarn with its adjective, may be merely an apposition clause belonging to the principal subject, with which the qui has nothing whatever to do. On reconsideration, I have no doubt that this is the proper explanation, but it may be put a little more clearly than in the passage alluded to. Thus, in the instances, ut nemo tarn humilis esset c u i non aditus pateret, Nep. in Milt. 7 : neminem un- quam tarn impudentem fuisse q u i, 8^c., Cic. pro lege Manil. 16. 48 : nemo erit tarn injustus qui, ibid, pro Marcell. 5, 15, the construction is ut nemo esset cui non aditus pateret — neminem fuisse qui — 7iemo erit qui — the tarn humilis, tarn impudentem, tarn injustus, being mere qualifications of the 134? APPENDIX. subjects nemo and neminem. This mode of viewing the nriatter, at once identifies the construction with that contained in the two subsequent exam- ples, num qui ante te tarn nefarii qui id facerent ? — an vero ullam usquam esse or am tarn desertam putatis quo non fama pervaseritl Cic, where the interrogative arrangement takes the place of the negative. Thus, then, there is nothing more in these examples than the ordinary Syntax, contained in the well-known expressions, ne7no est qui, qiiis est qui ? The employment of talis, tantus, tam, merely as intensives, without any qualis, quantus, or quam, occurring after them, has been sufficiently exhibited in chap. vi. of this work. Thus has tam prosperas tamque in- opinatas res consecuta est subita commutatio, Nep. in Dion. 6, where no quam or adverbial relative clause is expected. (o) page 66. § 36, Obs. 6 to 10. That ac, atque, are of relative origin, is not only rendered probable by their being used after compara- tives and other words, where quam is also usual, but from the additional forms, atqui, atquin. The very common occurrence of que appended to relative forms, and the use of at in the same sense as atqui, are further corroborations. As respects inde, there is no more difficulty in supposing that in stood for eo, than that un stood for quo in unde. Hinc, too, is^ evidently closely connected from its use in dehinc (deinde), where it occu- pies the place of an ablative, and from the fact that c was dropped in words of this kind, and formed no part of the root ; as in illi for illic, and illim for illinc. The frequency with which the Greek ο is changed into w, in Latin derivatives, is familiar, and explains the connexion of Greek and Latin relative forms ; and we can also trace a close relation- ship between undique, and de quoque, uspiam and quopiam, usquam, un- quam and quoquam, usque and quoque. The numerous binary and ternary compounds, which words of the alius, quis, and u family form, are like- wise well known, and a strong intimacy is thus marked. It is common to derive ubi from the Greek οττου; but from its relationship to ibi, as compared with that of unde to inde, one is strongly led to imagine that the i, in both words, is of the same origin, as well as the u. We cannot, however, always lay much stress on the preservation of case in these adverbs, nor is it necessary to show that hinc was an ablative strictly in form. In adeo, whose etymology is explained by the use of ut after it, the ablative form is used after ad, which governs an accusative, there being little doubt that the meaning of adeo is that of ad id. (tt) page 68. § 36, Obs. 9. Quin is also used for ' why not,' where its etymology is quo non, or cur non, and also followed by an imperative, where we can no longer substitute these interrogative adverbs, but where it has clearly the same sense, though no longer suiting the construction. Here is the first transition step to its thoroughly adverbial use, just as below we see ne retained in passages v^^here the interrogative sense is con- tained, though the structure no longer properly admits it. In the mean- ing of 'but,' used at the commencement of a period and elsewhere, it has deviated still more widely from its original signification, but the process may be distinctly traced. (p) P^g6 69. § 36, Obs. 9. We find ne after adeo followed by an accusative and infinitive, as in the following, adeon' homines immutarier ? Ter. Eun. ii. 1, 19 : adeon' rem rediisse ? ibid. Heaut. v. 2, 27, where although the 7ie is in progress of extinction, its employment answers to that of other interrogatives with the accusative and infinitive in the oblique oration, or the indignant question. In cases like egone mea bona ut dem APPENDIX. 135 Bacchidi dono sciens ? Ter. Heaut. v. 5, 6, the ne clearly imports the interrogative character of the sentence, while the ut dem depends on some- thing implied ; as, Ί be such a fool as to give,' the construction partaking of the interrogative and interjectional. Two passages have been quoted in note 8 to this reference, containing utne, as supposed to be used for utrumne. These admit, however, of being classed with those immediately preceding, partaking of the interrogative and interjectional character as they do. The ne may here indicate the interrogative manner of the speaker ; the ut, the actual construction. (σ) page 73. § 37, (e). I have neglected to notice some examples of double interrogatives in Latin of a different kind from any mentioned in the place here referred to. Quid, interrogative, is not unfrequently found preceding another interrogative sentence, as in the following ; quid enim censemus, swperiorem ilium Dionysium, quo crucial u, S^c. ? quid? Alexandrum Fherceum, qu ο animo vixisse arbitramur ? quid? Mace- dones nonne Demetrium reliquerunt ? quid ? Lacedcemonios nonne, <^c. ? Cic. Off. ii. 7, where notice also that Dionysium, Alexandrum, Macedones, and Lacedcemonios, are each taken out of their clause, and made to precede it. Similarly, quid? qu ce imhelles dant pr celia cervi ? Virg. Georg. iii. 265. (r) page 80. § 39, (b). The clause to which the relative, as demon- strative and conjunction, belongs, may be a part of another relative clause, belonging to the subject of another objective relative clause, as in the fol- lowing : cujus studium qui vitwperat, haud sane intelligo, quidnam sit, quod laudandum putet, Cic. de Off. ii. 2, where the real construction is, et haud intelligo quid id sit quod is, qui vituperet ej u s (or id) studium, putet lau- dandum. This renders the construction more complicated, and causes it to assume the appearance of anacoluthon, though it may be easily analysed and found not to be so. I have, perhaps incautiously, put the rule as to the initial qui, some- what too generally. It often happens that the ordinary and strictly rela- tive clause is taken out of its natural position and put before that to which it belongs, either as an explanatory or objective clause, and thus qui will stand at the beginning of a period, where it is not referred to any subject immediately preceding it in the former period. Thus, qui recte vivendi prorogat horam rusticus expectat, Hor. Ep. i. 2, 41 : qui cupit aut metuit juvat ilium, ibid. 51 : qui non moderabitur irce, infectum volet esse, ibid. 59, where the qui in each case occurs after a full point. (v) page 90. § 49, Obs. Si is used in other cases, where it is no mark of the hypothetical sentence, but has something of its primitive force, sine or sit, ' granted that,' whichever explanation is preferred ; as, de quibus dicere aggrediar, si pauca prius de instituto ac de judicio meo dixero, Cic. Off. ii. 1, where Valpy quotes Virg. Mn. v. 64, 65. See also Cic. Off. ii. 6, si prius dixerimus, near the end. (0) page 92. § 52, Obs. 3. In the annexed w^e have si qui with the subjunctive, and si quis with the indicative, but in neither case is the clause an objective one, the first verb being in the potential, more pro- perly speaking ; ut s i qu i in foro cantet, aut si qua est alia magna perversitas, Cic Off i, 40. INDEX OF AUTHORS, The editions of ^schylus used in the following Table are Klaxisen's Agamemnon and Choephoroe, and Blomfield's Prometheus. 1 have also employed Matthise's Euri- pides, Wunder's Sophocles, Mitchell's Aristophanes, Dissen's Pindar, Stallbaum s Plato, Spitzner's Homer, and Ernesti's Cicero. The quotations from Cicero usually contain two references ; the first to the book, where the subject is treated of m more than one, and the second to the chapter ; or where there is but one principal division, to the chapter, and to the smaller sectional number printed in the edition referred to, which is wanting in some few instances. It is unnecessary to particularize the edi- tions used for the other authors, as there is a sufficiently near agreement in the various editions to render this superfluous. ^SCHYLUS. SECT. Agamem 249. 36 not. 1 575. 76 b Choeph. 10. 19/ 204. 2 518. 23 Prom. 148. 58 167. 19 ob. 209. 23 291. 85 ob. 584. 19 ob. 679. ib. ob. 777. 37 ob. ib. 70 881. 19 ob. 956. 34 6 Suppl. 455. 23 ARISTOPHANES. Ran» 96. 94 ob. 2. Equit. 781. 22 ob. 1272. 2 ob. 5. Lysist. 615. 2 Vesp. 503. 3 614. 25 e 632. 70 711. ib. A.RRIAN. De Exp. Alex. vii. 30, 2 32 ob. : 3. BEAUMONT AND C^SAR. SECT. FLETCHER. B. G. i. 12. 25 a SECT. 1 14. 38 Maid's Tra gedy, act ii. sc. 1. 27 rem. 36 ob.lO. ib. 16. 18. 19. 20. 48 74 25 b 93 ob.3. 38 C^SAR. ib. 93 ob.2. B. C. i. 2. 48 21. 75 c 12. 52 ob. 2. ib. 92 20. 9 a 26. 38 ib. 75 27. 98 33. 36 not. 2. 29. 11 34. 28 not. 5. ib. 25 b 54. 25 a ib. 38 73. 33 h 31. 11 76. 25 a ib. 12 77. 25 6 33. 33 i ii. 5. 44 ib. 36 note. 18. 25 b 36. 81 32. 33 b 37. 82 iii. 12. 56 ib. 93 15. 25 a 38. 12 20. 4 44. 39 a ib. 48 45. 28 28. 41 bis. ib. 80 V. 144. 48 46. 19 e B. G. i. 4. 76 a ib. 51 6. 11 6is. 47. 57 7. 12 48. ib. 8. 25 6 ib. 52 ib. 49 ob. 52. 86 10. 12 70. 93 12. 19/ ii. 31. 81 138 INDEX OF AUTHORS. C^SAR. SECT. Cicero. SECT. Cicero. SECT. B. G. iii. 3. 46 in Cat. iv. 7. 12 ad Att. vi. 2. 93 4. 32 9. 1 ob. 2. vii. 8. 89 ob. 9. 12 ib. 37 c xii. 28. 93 ob. 2. 20. 15 10. 33 h De Fin. V. 10. 48 23. 8 proMuren. 1.2. 39 a 12. 9 ob. 2. 26. 46 2.3. 19 ob. ib. 75 d 28. 82 4.8. ib. ob. 16. 9 ob. 2. iv. 1. 15 10.23. 25 b ib. 75 d ib. 46 14.31. 25 a Tusc. Qusest. 13. 48 15. 32. 33/ i. 9. 9 ob. 2. 24. 14 16. 34. ib. k ib. 75 d 27. 48 17. 35. ib. k iii. 23. 33 b 31. 25 b 20. 42. 1 iv. 19. 48 32. 11 21. 44. 32 ob. 2. V. 27. ib. V. 2. 25 a 22. 46. 26 De Nat. Deor. ib. ib. c ib.ib. 28 not. 5. i. 14. 36 ob. 9. 4. 13 ob. 25.51. 1 ob. 4. 44. 38 4. 39 d 27. 57. 39 δ De Legg. i. 15. 25 a 6. 80 pro Sulla, 2. 4. 25 a 22. 75 c 13. 11 pro Arch. 3. 5. 48 23. 12 19. 13 ob. 3. 6. ib. DeOffic . i. 10. 48 ib. 39 d pro domo 14. 86 ob. 22. 25 sua, 41.108. 85 19. 75 d 25. 40 pro ib. ib. d 30. 48 Plane. 32. 79. 75 a 24. 86 ob. 31. 91 ob. 33. 82. 93 26. 33 k vii. 35. 36 ob. 9. 34. 83. 75 6 27. ib. not. 8. ib. 37 e ib. 85. 35. 87. ib. ib. c 30. 31. 48 not. 6. 25 a CELSUS. pro Sext. 19. 43. 75 a 32. 25 b Prsefat. 39 j bis. 28. 60. 77 ib. 39 job. 2 i. 2. 75 ib. ib. 80 37. ap.(i) 18. 38 ib. ib. 82 38. 76 note ii. 1. 16 30. 66. 79 39. 39 joh.2. ib. ib. 31. 67. 82 40. ap.(i&) 8. 39 ib. 68. 85 41. 52 ob. 3. ib. 48 33. 72. 75 a ii. 1. ap.(^) 18. 76 36. 78. 83 2. ap.(r) ib. 81 ob. 38. 81. 50. 108. 75 87 3. 4. ap.(^) 25 b CLAUDIAN. 51. 109. ib. ib. 28 De Fl. Mall 60. 128. ib. ib. {β.''' Theod. Cons. 62. 130. 85 102. 36 ob. 4. 63. 132. 75 c 6. ap. {v) pro Mi . 14.38. 37 7. ap.(σ) CICERO. 16. 42. 66 iii. 1. 49 Delnven. h .39. 37 b pro Marc. 2. 5. 33 k 11. 81 bis. 45. 91 5.15. 36 ob. 4. 12. 75 a 58. 37 b ib. ib. 39 e 13. ib. a 59. 33/ ib. ib. ap.U) 31. 93 Or. pro 7.21. 41 DeAmic. 2.6. 33 c Quintio, 13 42. 33 ί ib. 23. 39 job. 1 ib. 8. 44 pro leg. Phil. ii. 44. 38 3. 10. 48 Manil. 3 8. 6 xiii. 6. 4 6. 22. 87 5 11. 39 aob.l Ep. ad 7.23. ib. 6 14. 94 ob. 1. Divers, ad 9.29. 48 8 21. ib. ob. 1. Lent. page 21. 33 c 10. 34. ib. 14 41. ib. ob. 1. Ep. ad 11. 35. 76 bis. 35 44. 36 ob. 4. Divers, v. 12. 89 12. 40. 48 16 48. ib. ob. 4. 19. 1 ob. 2. ib. 41. ib. ib ib. ap.(l) vi. 1. 93 ib. ib. 75 19 .57. 88 ib. ib. 76 bis ill Cat. i. 3. 58 X. 8. 39 aob.l ib. 42. 33 k iii. 3. 50 c xi. 13. 33 c ib. 43. 48 6. 39 a ad Att iv. 7. 21 ob. 4. ib. ib. 75 d INDEX OF AUTHORS. 139 Cicero. SECT. Euripides. SECT. DeAmic. 15. 53. 76 Hecuba, 395. 37 d 17. 61. 48 398. ib. d ib. 64. 1 ob.3 571. 18 22. 82. 33 h 629. 19 ob. ib. 83. 48 691. 39 b ob. Paradox, i. 1. 49 697. 36 ob. 5. 2. 28 58 De pet. 754. 25 b bis. Consul. 2. 31 804. 22 3. 33 k 949. 60 6. ib. a bis. 965. 25 6 ib. 40 977. 55 9. 33 a 981. ib. ib. 40 1172. 59 11. 76 1184. 22 ob. Orestes, 34. 19 ob. DEMOSTHENES. 686. 37 d 33 g Olyuth. A. t'. 75 d 697. B. a\ 50 a oh. Phoeniss. 43. 37 e «'. 39/ 145. 19/ ζ'. 34 290. 23 ib. 34 a 390. 55 ib. ib. d 410. 37 e Philipp. A. ιγ'. 36 ob. 3. 445. 34 b 37 e 519. 2 ιδ\ 36 ob. 5. 571. 39 g Δ. ιη. 75 d 594. 87 ob. De Halones. ι. 56 ob. 5. 657. 19 c DeCherson.i)6'. 25 b ib. 58 ly. 20 ob. 662. ib. 36 ob. 9. 670. ib. Philipp. Γ. r. 22 ob. 681. 19 b Θ'. 25 e 878. 37 e bis Δ. δ'. 26 887. 55 33 e 892. 37 e ι. 32 ob. 2. 940. 25 b ιδ\ 36 ob. 9. 1045. 1 u. 76 b 1235. 35 De Classib. η. 89 1288. 37 e θ'. ib. 1517. 58 De Megal. β\ 33 if 1532. ib. De Coron. τ'. 19 e 1597. 22 26 Medea, 473. 56 V' ib. e Hippolyt. 22. 23 θ\ ib. e 101. 25 b is'. 37 ^ 280. 19 ob. t oo κ t; case of the relative . . . . . . . J "^^ ^*^^• ^* This is done with other words besides αυτός . . 21 obs. 5. in the same member with the relative . . . . 1 9 6 & 20. in the same member with a participle in place of the relative 23 sometimes said to be redundant, but this representation insufficient 20 in the same or in a diffOrent case from the article and participle preceding it, which latter are put absolutely in the nominative, genitive, dative, or accusative . 23 a, 6, c The nominative in this use takes the construction of the relative just as though the finite verb had been put in place of the participle ib. Same construction with other words . . . . ib. c, obs. This construction with another noun or pronoun pre- ceding it in an absolute construction, met with in the best writers ........ 24 C, C identical with q, and probably nearly related to the Greek aspirate in some words . . . . . . . 19 obs. C final in many words a mere breathing or support for the voice, sometimes present, and sometimes absent before u . ib. obs. & note 3. CuUf cunde, cuter, the forms which uhi, unde, and uter assume in compounds 36 obs. 9. and ubi, like the Greek που, ττοί, probably old genitives or datives ib. obs. 9. Cum for quod ib. obs. 6. Cunque appended to ut ib. obs. 6. sometimes pnt alone ior quandocunque . . . . ib. obs. 8. This termination is frequently equivalent to the dupH- cation of the relative form which precedes it, thus ut ut, uhi uhi, quis quis, qualis qualis, &c. are put respectively for utcunque, uhicunque, quicunqtie, qtialis- cunque, ^c 36 obs. 8. & 37 α D. Dative elliptical 3 Declinable form of the relative often used as an adverb . . 36 This is done occasionally in the poets where the relative is in agreement with a noun, in- stead of the adverb being employed . . ib. obs. 4. The same done sometimes with the demonstrative ib. obs. 4. Demonstrative or direct construction used in place of that by J in ^ ^^p (^') the relative (25 i'obs. & note 9. used immediately after the relative structure in the prior member of a sentence, instead of the relative 19 6 together with the relative structure used in the same member, where one appears wholly redundant, frequently found in Greek . . 20 Similar English example . . , . . 21 obs. 3. The same thing in Latin with qui illi, where the common view is, that qui is redundant in INDEX OF MATTERS. 149 SECT, a wrong case, though this may be explained by supposing that the qui has only the force of the copula ...... 21 obs. 4. Demonstrative melted into one clause with that of the relativ^e, particularly after interrogatives τίς, quis . 19 / The same with the demonstrative . . . ib. note. pronoun and the noun with Avhich it agrees often separated in Greek, so that the demon- strative comes last ..... 47 pronoun and its noun separated by the inter- vention of the relative clause, so that the noun comes last, and this may explain some cases of what has been termed transposition 46 adverb coupled with the relative adverb in the same member ...... 21 in agreement with a noun instead of depending on it as the latter of two substantives . . 13 obs. form in lieu of the relatiT^e, in Homer, at the beginning of a period, rendering it probable that the construction is really what it appears to be 19 obs. and copula often supplied by qui, qncp, quod . 19 obs. 27 obs. & 38. adverb put in Heu of relative adverb in German ib. obs. repeated after parenthesis .... 22 obs. pronoun often inserted in Greek interrogative sentences . . . . . . . 25 δ plural with singular interrogative ... 55 form found in both clauses where one should be relative, which is found else'-'liere re- versed 19 obs. 37 6, obs. 2. Absence of, no proof that the relative clause is objective, any more than its presence is that the relative clause is merely explanatory . 75 a, h, c Dependent clauses in English do not admit the relative always as in Latin .......... 38 note I. Difference of Gi^eek and Latin construction in the case of relative clauses 27 ί 2 obs 2 Different modes of explaining the same construction may be J s 1 ' 1 '> resorted to I ^ u * i ' o* i, 4 obss. 1, 3. Difficulty of determining always whether the construction is relative or direct, when the demonstrative form is em- ployed 19 obs. Digamma, expressed in Latin by /or r», appears in some cases to have answered to the sp'iritus asper of the initial ρ and v, and probably to q or qu in Latin App. (Θ') Dorians preserved τ in the nominative plural of the article and relative ib. obs. & App. (i) Double comparatives 33/7 I'elative arrangements in similes 33 p', 37 c? relative compounds answer to indefinites in cunque . Ί ^7 " relatives in lieu of relative and demonstratives in admirative or interjection al passages . . . ib. c interrogatives in the same member . . . . ib. e The same in Latin where 5ΜΪ(ί? precedes . . . App. (σ) Ε. η δ' ος, η δ' η, Attic use of 19 ιί, origin of 36 obs. 9. in particular phrases 8 for ort 36 obs. 9. eWe answering to si, ο si 36 obs. 9. ίίτΓΟΤί for ϊνα or ώστε ........ 49 obs. ίδΟ INDEX OF MATTERS. SECT. tl TIC used like si quis in Latin 4 obs. 2, common with suppressed sense before it . . . ib. used of a definite party ....... 49 supplied by ος ib. Ellipsis so called, not always such, but merely a convenient mode of representation 2}a and App. (p) Negative propositions in Greek have the relative with av and the optative, as in Latin the subjunctive .... 73 note. Nemo followed by homo 32 obs. 3. followed by unus ib. obs. 3. viv, μ'ιν often occur after the relative in a corresponding mem- ber of a sentence coupled only by και. (Direct for relative structure.) .......... 19 6 Nominative elliptical 1 Nominativus pendens . . . . . . . . 23 Nostra after interest not ablative feminine .... 28 note. Noun repeated in the same member ..... 22 note. O. ο for ος , App. (t), {η) for OTi as conjunction . . . . . . . . 36 translated by ' iu that ' ....... ib. obs. 4. Object of a relative clause sometimes removed and put as object to a preceding verb ....... 20 obs. Objective relative clause how distinguished .... 76, &c. ό ^έ not by way of contrast, but for ος 19 obs. App. (») followed by δς μίν ........ App. (ί), (?;) οίος vises of .. , ...... . 3 obs. 1 & 34. with force of ος . . 34 δ & App. {ν) 152 INDEX OF MATTERS. SECT. οίος intensive . . . 34 & 34 c attraction prevalent with it ib. a, App. {μ) frequently doubled 34 c, 37 c redundant 34 d οίο J' difficulty of determining when used adverbially . . ib. οία both used as adverbs ...... 35 with the force of quum ....... 34 for ως, ώστε 4 ........ ib. oTTtp for οσττερ App. \^i), (//) οποίος used like olog and qualis agreeing with a noun in its clause .*......... 34 α οπότερος for πότερος 62 Optative in the oblique oration in Greek after λέγω, ερωτάω, είπον ......... 73 note. after ως, οτι in oblique narration . . . . ^8 a much modified by the freedom usual among Gi-eek writers ........ ib. ό ος signifying 'and he' Avith copula App. (t) interchanged with οίος ....... 34 ό used after τοΧος in Homer and elsoAvhere .... App. {v) for ε'ί τις . . 49 όστις, οίος Avith άν or κεν, κε and the optative ... 73 note. δς μεν . . . δς δε used for 6 μεν . . .'6 δε 19 & App. (t) followed by ό ίέ , App (t), {η) with the second omitted 19 obs. usud partitively as in English the phrase ' what with ' . ib. obs. οσα for ως .......... 36 obs. 4. οσα μη for οτι μη . , ....... 35 'όσος after τόσος Λvith substantive in its ΟΛνη member . . ib. without Γο'σος, in reference to a preceding subject . . ib. with antecedent elliptical ...... ib. after adjectives of quantity and superlatiΛ^es . . . ib. for δς, though then generally after a word of quantity . ib. after a comparative adjective . . . . . 33 A; οστε in the sense of ' who,' from δς ' he' and τε the conjunction 19 obs. App. (t) δσης uses of 2 obs. 5. origin of ......... 25 note. used for ος definite 37 a, obs. & 70 followed by δς in relation to the same subject . . 37 «, obs. „ ,, . Γ 25 α note 7• followmg τις .^ ^3^ ^^ ^^^^ ov for ώστε μη ........ 36 obs. 4. ,t > answering to the Latin qui alius ... 26 or£ for 'at one time' opposed to άλλοΓί, accented οΓε . . 36 obs. 5. for οστε App. (t) for a(f ου 36 obs. 5. oTi signifying 'that' and 'because' of one origin, explained by ellipsis .......... 36 for ri . 62 supplied by f'i rt . . . . . . . . 49 with the accusative and infinitive 39 obs. 4. οΓίς for δσης . . . App. (ι), (j;) ουδείς like quisque put in apposition with a plural ... 32 obs. 2. with τις in the same member . . . . . ib. obs. 3. ώς for δς ......... ^ 36 obs. 5. for οία .......... ib. obs. 5. for όσος, δς . . . ib. obs. 5. ωσττερ for οσττερ . . . . . . . . . ib. obs. 5. Participle in lieu of the relative construction . . . .16 obs. 3, & 19 followed by the finite verb in a second member, coupled only by και 19 α INDEX OF MATTERS. 153 Participle expressed by the finite verb .... This construction common in St. John . Person, change of, from third to second or first with relative Personal pronoun repeated in the same member and demonstrative often put before the rela tive clause to which they are the subject Primus, princeps, absorption of the relative with them Pronoun relative and αυτός in the same member . supplied by the participle in the same con struction . . - . Proper name repeated in the relative clause, or placed there by transposition often put before the relative clause to which it is the subject ..... ΤΓοίος with the demonstrative in its member in the same case its peculiar construction and tendency to attraction Possessive pronoun followed by the personal relative in agree ment with the primitive .... The same thing when the relative is suppUed instead by the participle ..... Postquam, priusquam with indicative, or in the obli(j[ue oi-ation with the subjunctive SECT. App. (i) 19 a 59 22 obs. 39 a, obs. 1 19/ 20 23 25 b 39 a, obs. 1 34 a J25g 1.34 a 58 ib. 74 obs. 1. which with Q represents probably something equivalent to the breathing before the initial ν of the Greeks, or was used for the commencement of some vowel sounds in the same manner as c final for the closing them . Sometimes it answered to the Greek r Qualis in its ordinary construction . to express degree or intensiveness supplied by qui common with a substantive in its clause, like οίος in Greek with a noun in its clause, followed by talis . in an oblique case, with the main proposition to it refers postponed ..... attracted ....... Quam in similes without tarn, in the same way as qualis out talis ....... preferable to quo after post .... used of ' place,' ' time,' ' quantity ' . preceding the comparative on which it depends Quamobrem for propter quern ..... Quantus, a, um in its ordinary constructipn in connexion with one subject of comparison, after a comparative adjective . without a fresh subject intensive ...... qualis after tantus, talis both used interrogatively . for quam tantus .... Qui derived from και oi . used partitively ..... shortened form of quis .... admirative with noun in its clause . with the force of qualis, quantus, particularly when with one or other of these . followed by si, nisi, cum, ^o. ... at the beginning of a period stands commonly for the demonstrative and copula when merely transposed not so used 19 obs. App. (t) 33 b ib. c ib. cSif ib. e. See note also. 33 e ib. e ib. Jc, obs. ib. e 36 obs. 4. ib. obs. 4 44 note 4. 36 obs. 5. 33 a ib. g ib. g ib. d ib. k ib. i 37 c 19 obs. 28 note 5 & 56. ib. 31 33/ 39 a ib. a, obs. 2. App. (r) 154 INDEX OF MATTERS. SECT. Qui in the construction of the ablative absolute . . . 39 c in agreement with a noun on which it should depend in the genitive , ib. cZ for ut and quam 36 obs. 4. after tarn differently explained App. (ξ) initial followed by si, nisi, not a case of anacoluthon . . 39 Qui probabl}^ old ablative of quis ...... 36 obs. 8. Qain supplied by qui non, both in use 36 obs. 9, note 5. Other uses of quin ........ App. (ττ) with subjunctive . . . . . . . . 93 Quippe qui with subjunctive ....... 91 Quisque generally a substantive 32 obs. 1. Quis, quid etymology of ........ 28 derived from τις . . . . . . . . ib. used substantively. . {-^enot.*• more common in objective clauses ..... ib. note. one mark of the objective clause ..... ib. used for cdiquis, after si, nisi, ^c. ..... 48 aliquis used as adjectives ...... 28 note 5. used partitively ........ ib. note 5. The neuter form for cur 36 obs. 6. and qui by duplication with another relative form give rise to quidam, quique, quicunque, aliquis, quisque, quis- quis, quisnam, quispiam, quisquam, most of them answer- ing to certain uses of τις, τις, όστις .... 32 This duplication seems to render them indefinite . . ib. Quisnam, etymology of ........ ib. Quisquam followed by uUus in same member . . . . ib. obs. 3. quidquam, by nihil in same way . . . . ib. obs. . Quisque like the Greek έκαστος is used in apposition with a plural subject as a partitive, and also in apposition with a noun in a different case ib. obs. 2. Quisquis, quicunque not alone the subject of a proposition . ib. obs. 1. Quivis as an indeclinable adjective 36 obs. 8, note 3. Q^^oafZ, etymology of {35 ^bs! 4, note. Quod used adjectively 28 also substantively . . , . . . . • "! 76 i t * in the sense of quod attinet ad id, exi)lained by ellipsis . 36 followed by the accusative and infinitive . . .38, 93 obs. 2. with the subjunctive in an optative sense . . . 39 ^ with subjunctive 93 Quoniam, etymology of 32 Quot more common than quotus 33 R. Relative pronoun originally a demonstrative in Greek construction not used in Hebrew .... often takes the place and construction of the sup- posed elliptical antecedent, in Greek similarly employed in Latin ..... in an oblique case, with the elliptical antecedent the object of the foregoing verb clause objective after the impersonal participle in dum, and the proof of this in agreement with a word not the antecedent, on which it should depend in the genitive . construction supplied by the participle repeated in place of the conjunction .... sometimes appears to stand for the conjunction, where it is said to be redmidant .... 19 obs. App. (i) ib. obs. 2 obs. 1. 4 obs. 2. 3 obs. 1. 1 obs. 4. 9 obs. 2. 13 16 obs. 3 & 19. 19 d 21 obs. 4. INDEX OF MATTERS. 155 SECT. Relative and demonstrative clause melted into one after inter rogatives ....... The same with demonstrative pronoun repeated after parenthesis .... in the same member Avith the personal pronoun put in the demonstrative construction, for which and the copula it often stands .... Its construction with medius, reliquus, totus,par,^c. in construction with two nouns, of which one is dependent on the other in the genitive, and the relative agrees with the former (rare) . twice repeated in the same reference form in both clauses, where one should be demon strative agreeing with an implied antecedent with change of person in attraction frequently in Greek not used of the last-named sub ject, where its place might be filled by και αυτός preceding the antecedent or the word to which the relatiA^e clause is objective .... clause belonging to a subject within another relative clause sometimes prseposed .... in the genitive plural after partitives usually avoids this last construction with singidus, reliquus, plerusque, 2}lurwius, multus, ^c, and agrees with them as an adjective .... Its inflexions used as adverbs in ατε, καθάττερ, ου, ol y,&-c clause when it expresses the sentiments not of the writer, has its verb in the subjunctive . Exceptions in certain circumlocutions clause Avhen emphatic, or containing something essen tial to the sense of a passage, takes the sub junctive ....... when causal takes the subjunctive . clause when constituting an integral part of a propo sition requires the subjunctive after an absolute negative proposition, or an almost vmiversal negation, takes the subjunctive in certain indefinite phrases with subjunctive . after the interrogative phrase quis est qui? and its cases with the subjunctive .... after dignus, idoneits, with the subjunctive after is put for talis, or after tarn, with subjunctive after comparatives followed by qui with subjunctive clause dependent on a preceding subjunctive put with the subjunctive ..... with subjunctive by attraction .... 19/ ib. /, note. 22 obs. 23 obs. 25 a, & 38 26 27 37 δ, obs. 1. ib. b, obs. 2. 58 59 f 25 c,d,e,f Hi L App. (.) 39 b, obs. 40—45 47 and App. (r) 57 ib. 36 obss. 1, 4. 80 ib. obs. 2. 82 83 84 85 87 88 89 90 94 96 S. Si, origin of 36 obs. 9, App. (υ) used for quod or ut ib. obs. 9. omitted ib. obs 9. does not always render a sentence hypothetical . . . App. (υ) Si quando, si qui, origin of, for quandocunque, quicunque , . 36, obs. 9. & 51. Si qui, ce,od, ana si quis, a, id, \)oi\v in n^e .... 51 more common than si quid with the indicative, Avhile the latter is more usual with subjunctive, though not always, unless in objective clauses .... 52 obs. 2. 156 INDEX OF MATTERS. Si quls, origin and employment of for quicunque does not render the clause hypothetical or doubtful ........ used precisely as a relative pronoun with ellipsis, trans- position into its clause, as a partitive, as the subject of the ablative absolute, and with its case altered by attraction ........ Spiritus asper often pronounced s in Latin derivatives, and probably with both the sounds of fr-and ί . . . . Subject of the relative clause removed from its own member, and made the object of a preceding verb, or adjec- tive, or substantive, a well-known Greek construc- tion, sometimes imitated in Latin, and also very rarely in English of the verb in the objective case owing to the trans- ference of that subject into the oblique case of the relative, the relative clause being then taken as a whole for the subject of the verb . noun in the second member in comparisons and object expressed by one word and predicate changing places Subjunctive and potential, their difference Often a verb is both used in objective relative clauses, after the phrases sunt qui, reperti mnt, S[C in the dependent clauses of the oblique narration Sunt multi how used with indicative or subjunctive . SECT. 36 obs. 9. 71- 50 a, b, c, df App. (Θ) 2 obs. 4. 20 obs. 21 obs. 3. 73 note. 25 b 76 6 25 h 61 63 App. {δ) -73 74 obs. 2. 75 d 81 86 obs. Tarn followed by qui Explanation of this construction omitted befoi'e quam often without quam Tantus, talis followed by ut and qui τις use of interrogative with the demonstrative in its member . roi, ταί nominatives plural both as demonstrative and relative, and followed by δε, γαρ, ye, κεν, κε, ρά in situations where hiatus is prevented or a short syllable lengthened τοίος used with similar attraction to οίος, where we should rather expect όντως, common in the Iliad . This confirms the idea of the original oneness of demon- strative and relative forms ...... τόν the oblique form of the pronoun demonstrative followed like the relative by pa, πότε, κεν . . . . . τόσος followed by άλλα του, τφ, τόν and roi, ταί, τά plural used as relatives 33 k App. (ξ) ib. k 2 obs. 5. 20 obs. 28 note 6. App. (t) App. {μ) App. {μ) 19 obs. 33 k 19 obs. u. ubi for ut Un a fixed form equivalent to quo . Unde for de quo used of person, place, cause Unquam redundant after si quando unde, ubi, uter, ut, uti of relative orig Unus redundant after nemo Ut derived from on its real meaning and power not always in sense of ' in order that 36 obs. 6. ib. obs. 9. ib. obs. 7• 52 obs. 1. rin 36 obs. 7. 52 obs. 1. 36 ib. note. ib. note. INDEX OF MATTERS. 157 SECT. Ut its meanings explained by ellipsis in the following, ' in order that,' * as,' ' h(jw,' ' wlien,' ' because,' ' considering that' 36 obs. 2. used for quum, quantum, quam intensive, quamvis, quomodoj and followed hy pote (Gr. ttOts) and cunque . . . ib, obs. 6. after ne in expressions of surprise ..... App. (p) with ne appended to it, supposed to be for utrumne . . ib. obs. 9, note 8. This last usage possibly analogous to the first named, where indignation or surprise is expressed, and not for utrumne App. (p) and the subjunctive how it differs from the simple infinitive and the accusative with infinitive 74 obs. 2. in what cases we may always use it ib. obs. 2. Verbs which take only a relative clause as their object . . 25 note. commonly used parenthetically which have either be- come or are on the road to becoming adverbs, fac, licet, vis, libet, oro, precor, ^c 36 obs. 9. which are introduced parenthetically do not put the verbs of the clauses, which might be made to depend on them, in the subjunctive mood . . . . 75 e which take either of two constructions sometimes put with both 61 i^ BY THE SAME AUTHOR. The USE of the ACCUSATIVE and INFINITIVE in LATIN, with especial reference to those constructions which sometimes occupy its phice, and to tlie pi'actice of the Greek Writers. II. Lately Published, A NEW EXPOSITION of the SYSTEM of EUCLID'S ELEMENTS: being an attempt to establish his Work on a different basis, by a new derivation of the Doctrine of Proportion, and an analytical examination of the nature of a Converse proposition and the doctrine of Identity. Containing, besides other things, a Demon- stration of the Theory of Parallels, as enunciated in the celebrated 12th Axiom, and a Demonstration of the 47th of the 1st, without the aid of that Theory. " We believe that the proof is by far the simplest yet offered to the Avorld, and w^e heartily wish that this qucestio texata may have been effectually put to rest." — Bristol Journal, Saturday, April 6, 1839. " Books of this kind are a proof of the good that is done by stirring up, and keeping going the question of the soundness of the boasted logic of the Greometricians. — ' It is, however, by no means impossible, that at some time or other it may be found neces- sary to place the. doctrine of Proportion at the outset of Geometry, as has occurred to the author. If the grand opprobrium of geometry, the theory of parallels, is ever vanquished, it is very likely to be found dependent on rigid examinations connected with the doctrine of Proportion.'" — Spectator. London, June 15, 1839. " This little tract deserves the attention of mathematicians." — Literary Gazette. London, June 29, 1839. " This is an ingenious book, which will be acceptable to the lovers of geometrical reasoning, a class unfortunately becoming daily more limited." — Athenceum. London, August 10, 1839. " And above all, one of the best improvements of this kind seems to me, your plan of introducing Proportion earlier in the course than usual, and making it the foundation of parallel lines. Upon the Avhole your publication appears to me liliely to be very 'useful : wdiat chance there may be of getting teachers to leave the beaten track is" another question." — Extract of a Letter from the Rev. Baden Powell, M.A., F.P.S., Savilian Professor of Geometry in the University of Oxford, Oct. 16, 1839. III. A TREATISE on PROPORTION ; with a View to a fundamental alteration in the method of teaching the Elements of Geometry, and the simplification of that part of it known by the title of the Theory of Parallels, in which is shown how Euchd's Fifth Book may he treated graphically, and the demonstrations at once con- veyed to the eye of the learner without the use of Algebra. " Dr. Day, of Bristol, has published two small Treatises, the subjects of which are closely connected together, and with that of our present discussion ; they are, 1st, A New Exposition of EucKd, 1839 ; 2nd, A Treatise on Proportion, 1840. " In the first, the Doctrine of Parallels naturally occupies a conspicuous place, while the writer's speculations on it ai^e much illustrated in the second. He treats the whole question in a highly original and compreliensive manner, as connected with philosophical A'iews of the system of Geometry ; following in the arrangement of his deductions a A^ery different order from that of Euclid, and most other Avriters on the subject. His entire method is founded on first treating the subject of triangles, and thence deducing that of Parallels ; and this by means of an intx'oduction of the Doc- trine of Proportion, at a much earlier stage in the course than is usual. A process to which, abstractedly, no objection can, I think, be found, and which is attended with many manifest advantages." — On the Theory of Parallel Lines, by the Rev. Baden Powell, M.A., F.R.S., F.G.S., F.R.A.S., of Oriel College, Savilian Professor of Geometry, Oxford, m\ ΨΛη .Mi^wwi LIBRARY OF CONGRESS 003 064 540 6 J^W Wi'^i w ^r,:^^ ^S»M^ •^li^m \^iMw '-Λ^) J ^ ^ =*■ iSf^l^ I^Xii