An Unrecognized Construction of the Latin Subjunctive: The Second Person Singular in General Statements of Fact

Английский Статья 20 век William Gardner Hale

William Gardner Hale on the second-person singular subjunctive in Latin general statements of fact.

An Unrecognized Construction of the Latin Subjunctive

The Second Person Singular in General Statements of Fact

Author: William Gardner Hale

Source: Classical Philology I, January 1906, pp. 21-42.

My paper deals with such Subjunctives as the independent ones in the familiar passages: A. Nil satis est, inquit, quia tanti quantum habeas sis; Hor. Sat. i. 1. 62. B. Assem habeas, assem valeas; Petron. 77. C. Ubi mortuos sis, ita sis ut nomen cluet; Plaut. Trin. 496. D. Quom inopiast, cupias: quando eius copiast, tum non velis; Plaut. Trin. 671. E. Quod in manu teneas atque oculis videas, id desideres; Plaut. Trin. 914. F. Unum quom noris, omnis noris; Ter. Ph. 265. G. Accipe nunc, victus tenuis quae quantaque secum Adferat. In primis valeas bene; Hor. Sat. ii. 2. 70. H. Dum tibi fit quod placeat, ille ringitur: tu rideas, Prior bibas, prior decumnbas; Ter. Ph. 341. I. Nam cetera maleficia tum persequare, ubi facta sunt, hoc, nisi provi- deris ne accidat, ubi evenlit, frustra iudicia implores; Sall. Cat. 52.4. J. Quem neque gloria neque pericula excitant, nequiquam hortere; timor animi auribus officit; Sall. Cat. 58. 2. K. Nec porro quaecumque aevo macieque senescunt, Nec, mare quae inpendent, vesco sale saxa peresa Quid quoque amittant in tempore cernere possis; Lucr. i. 325. L. Quae si ipse exsequi nequeas, possis tamen Scipioni praecipere et Laelio ("a Scipio and a Laelius "); Cic. Sen. 9. 28.

These look alike. If there were a comfortable category in the grammars into which to put them all, one would not think of separating them. As the title of the paper implies, I am brought to regard such Subjunctives as expressing general statements of fact,-as corre- sponding, in effect, to Indicatives in the same second person singu- lar indefinite. I wish to translate example A by, "because you are worth just what you possess ;" B by, "have a penny, you are worth a penny ;" C by, "when you are dead, dead you are," etc. In accordance with this view, § 542 of the Hale-Buck Grammar reads: "A general statement of fact is sometimes expressed by a Subjunctive of the Second Person Singular Indefinite;" and to our example C, there given, is added the statement, "the second sis has the force of es."

The construction itself will prove to be a simple one to solve, but the exhibition of previous opinion upon the matter must be a somewhat tangled affair, since grammars, grammatical treatises, notes in commentaries written by professional grammarians, and notes by commentators who are not professional grammarians, have all to be taken into the account. It is also obviously impracti- cable to refer to every note ever written upon any of these pas- sages. The selections, however, will fairly cover the ground.

So far as I know, no grammar or grammatical paper had rec- ognized the existence of the construction for which I am contend- ing. The statements regularly made (where any are made) with regard to the range of the Subjunctive of the second person singu- lar indefinite are of a kind to exclude such an interpretation. Thus Madvig, § 370, says: "Diese Form findet sich in bedingter Rede, in hypothetischenAussagen und in Fragen uber das, was geschehen wird und kann, in Nebensatzen mit Conjunctionen, in Relativsat- zen (mit qui oder einem unbestimmten Relativ), und in Vor- schriften und Verboten." To this is added the Anmerkung: "Ein solcher conjunctivischer Bedingungssatz fuhrt nicht den Con- junctiv im Hauptsatze herbei." Quite recently (1905), Nutting, Studies in the Si-Clause, pp. 84, 85, has distinctly shown the conviction that there is no Subjunctive in a main sentence corre- sponding to the common second singular indefinite in subordinate clauses. "One looks in vain," he says, "for a clear case of such leveling in Plautus." And Elmer, in his note on F (as will be seen below), in effect says that a general statement of fact in the second singular indefinite must be expressed by the Indicative.

The grammarians of course all provide for the Subjunctive second person singular indefinite in conditional or corresponding relative clauses which, if in any other person or number, would be in the Indicative, as in memoria minuitur, nisi eam exerceas, Cic. Sen. 7. 21; but they know nothing of the power for which this paper contends. If anywhere they chance to make use of an ex- ample of the type before us, it is placed under some already recog- nized and entirely different category, as we shall see below.

Whatever the grammarians may do, the commentators, it would seem, would frequently have occasion to discuss such examples. The most striking feature of the case is, however, that they gen- erally pass them without comment. Sometimes, indeed, they actually give an explanation of the dependent verb in such com- binations, and pass in silence the far more difficult independent one. Thus Greenough, in his edition of the De senectute (1873 and in various reprints), says of our example L, "si . . . . nequteas, gen. cond. (§ 309 a)," but says nothing of possis. Occasionally the dependent verb and the main verb are dealt with together and without distinction, with a reference to some rule, such as the grammars afford. This reference, as is natural under the circum- stances, never fits. Thus Gudeman, dealing with example I in his edition of Sallust, says, " persequare . . . . provideris inplores: Subjunctive of indefinite second person singular," and refers to the Allen and Greenough Grammar, 266a. This section, when one turns to it, proves to deal with independent "EIortatory" Subjunctives, and to read (266), " The Hortatory Subjunctive is used to express an exhortation, a command, a con- cession, or a condition;" and (a), "The Second Person is used only of an indefinite subject, except in prohibition, in early Latin, and in poetry." Not only the general headings "independent" and "'h ortatory," but the examples given (e. g., isto bono utare dum adsit, Cic. Sen. 10. 33), show the type meant. But nisi provideris is evidently not independent; and it is also difficult to believe that Sallust means to direct people to implore in vain, unless, indeed, the editor conceives the idiom to be one of ironical command (not provided for in the reference), in which case the explanation should certainly be given. More frequently, if any- thing at all is said about the main Subjunctive, it is simply classed with the dependent one, and the Subjunctive idiom in the second singular indefinite in " conditional " clauses is referred to or assumed. Thus Munro, in his comment upon our example K, puts possis with si non relinquas, Lucr. i. 515, quibus careas, ii. 4, etc. Thus, again, Greenough, in the note in his edition of the Satires and Epistles of Horace, 1888, says of our example A, "tanti sis (you are r ated att), etc. The subjunctive is the regular one of the second person with indefinite subject." But in his Gram- mar Greenough (the same is true of his revisers) makes no pro- vision for an independent Subjunctive of the second person singular indefinite, except as hortatory or potential; while the treatment of the dependent member of the example in question comes in § 309 a (518 of the new edition) under the head of "General Conditions," to which it is obvious that sis does not belong. It is clear, also, that general conditions are really meant; for the rule reads: "The Subjunctive is sometimes used in the second person singular to denote the act of an indefinite Subject (you =any one). Here the Indicative of a general truth may stand in the apodosis."'

Yet the tendency naturally produced by the sight of the com- bination of dependent and independent Subjunctives in the same indefinite second person is a good one, and has in a number of instances led, even in the absence of an admitted grammatical category, to the solid interpretation. Thus Wagner, in the note in his Trinummus (1875) upon our example E, says, " id desid- eres, 'that one is apt to forget': the second person subjunctive expresses generality." Greenough (1888), as we have seen, translates tanti sis of A by " you are rated." Friedlander, in his Petronii Centa Trimalchionis (1891) translates our example B by 4"G laubt mir, habe einen As, so giltst du einen As." Morris, in his edition of the Captivi and Trinummus of Plautus (1898), paraphrasing example C, says, " when you are dead, you are dead in the full sense of the term." He makes no grammatical com- ment. Rolfe, in his edition of the Satires and Epistles of Horace

[1] The word " sometimes " is not happy. The Subjunctive is regularly used, there being but a very small number of exceptions in the whole range of the literature. Hardly any usage in Latin is more fixed. Or, if the intended meaning is " one some- times finds examples of . " then, by the same token, the word " sometimes" ought to be added to every rule in the Grammar. Moreover, the phrase " may stand" in the last line should read " generally stands."

(1902) puts the main and subordinate clauses in example A to- gether, rightly translating, " because you are rated by the amount of your possessions, " and adding, "habeas and sis are subjunctive because of the indefinite second person singular." No grammati- cal reference is given. The procedure of Wagner, Greenough, Friedlander, Morris, and Rolfe seems to imply a subconsciousness of the force naturally to be attributed to the independent con- struction; but these excellent opportunities which Greenough and Morris, as grammarians, had for calling attention to a much mis- understood, and very variously interpreted, construction, and of enunciating its true character, pass unused. Note also, under 4 below, p. 27, a discordant explanation of the exactly similar per- sequtare of example I (namely as Potential) in the Greenough and Daniell Sallust.

The only comment I have noticed which makes a sound ex- plicit statement about the independent construction is that of Sch-itz (1881), who, in his note upon our passage A, says "sis, nicht es, weil allgemein, nicht an H. direct gerichtet; 'man gilt."' To this he adds, "Im Indicativ wulrde man eher die 1. Person verlaingen, weil der Geizige doch sich selbst damit entschuldigen will." He is right in his general feeling, but not right in imply- ing that in the second singular indefinite the Indicative could not be used. See below, at the end of this paper.

The explanations which clearly attempt to bring one or arnother of the independent Subjunctives in sentences A-L under one or another of the categories provided for in the grammars are of six kinds. I shall first merely state them, without argument.

The Subjunctive is that of the Oratio Obliqua. This is the favorite older explanation, and is still frequently found. Thus Wusteman, revision of Heindorf (1843), says of our example A, "man bemerke den im Latein. gew6hnlichen Uber- gang von der orat. recta in die obliqua;" H. Duntzer (1869) says "sis, Ubergang in die abhangige Rede;" Kirchner (1885) says "im Lateinischen geht die Rede aus der directen Construc- tion (nil satis est) in die abhangige (quia tanti . . . . sis) uiber, was auch in Prosa nicht selten vorkommt;" and G. T. A. Kruger in various editions (as in 1860 and 1866), by a still bolder device, says "sis, statt es, als ware vorhergegangen: nil satis esse ait. Also auf einer Anakoluthie beruhend." G. Krufger says the same, in re-editing his father's edition in 1876. Among the later editors, Kiessling (1886) says the same in the sentence, "der Konjunktiv sis = esse putaris." Palmer (my edition is of 1888) seems to mean the same by his phrase "general opinion" in saying, "the subjunctive sis is used because it is the general opinion, indefinite." And Harper's Lexicon (1879) means the same, in putting example A under the Subjunctive illustration of the statement, "construed with indicative in asserting a fact; with subjunctive in stating an assumed reason, or one entertained by another mind."

The Subjunctive is that of Command or Exhortation ("Imperative"). Thus Blase, in his recently published "Tem- pora u. Modi" in the Historische Grammatik d. lat. Sprache, p. 136, puts our D as his first example under the head of the Jussive, Hortative, or (true) Optative, and accounts for non as modifying, not the sentence as a whole, but the single word velis. The same explanation seems to be meant, p. 123, for C, which is translated by "so bald man tot ist, soll man es ganz sein." And the same is explicitly suggested by Antoine and Lallier as possible for I. See below, at the end of 4.

The Subjunctive is Permissive. Thus Elmer, on example H, says, "'rideas, bibas, etc.: subj. of permission, involving the yielding of somebody's will; to be carefully distinguished from the potential."

The Subjunctive is Potential. It is by the Potential formula ("you can laugh, drink your wine before him," etc.) that Morris H. Morgan, in his translation of the Phormio (1904), renders these same Subjunctives which Elmer tells us are "to be carefully distinguished from the potential." The express word "Poten- tial" is used of one or another of our examples by many commen- tators. Thus Dillenburger (my edition is of 1867) both affirms this explanation and denies 1 above, in the note: "sis non per anacoluthiam positum est pro es, quasi praecederet nil satis esse ait, qutia, sed est potentialis modus, ut apud Lucilium." Likewise G. Krttger, who in the edition of 1876 explained sis in A as due to Oratio Obliqua, explains it in the edition of 1889 as Potential. This same explanation is given by Lucian Miller (1891) for A and B, by Blase op. cit., p. 142, for B, by SchtVtz op. cit. for G, and by Schmalz Bell. Cat. (my edition is of 1885) for J.

Unfortunately this word "Potential," as used in our gram- mars, is ambiguous. It includes meanings so far apart, at their extremes, as "may possibly" or "can," and "would certainly." The former idea alone should bear the name Potential, while the other (seen most easily in the common Subjunctive conclusion) should bear some name like that of Ideal Certainty, which I have given it in my own grammatical writing. [1] It is the true Potential use, apparently, that Kruger and Miller have in mind, and likewise Freeman and Sloman upon our example D. This, at any rate, is what is meant by Greenough and Daniell in their Sallust, in the note upon persequare in our example I. The translation given is "one may prosecute," and the reference (Allen and Greenough Grammnar 311a) is to the statement, "The Potential Subjunctive is used to denote an action not as actually performed, but as possible." Similarly Antoine and Lallier, Salluste, Co7ijuration de Catilina (1888), write in their commentary: "persequare pourrait etre un subjonctif imp6ratif, puisque le sujet est indetermin6; mais en réalité, comme le note M. P. Thomas, c'est le subjonctif potentiel: 'on peut ne les pour- suivre que lorsqu'ils sont accomplis.'"

[1] 'Delbruiek Verg. Syntax II, p. 371, and Brugmann Kurze vergl. Gramm. II, p. 584, while accepting my term, have misunderstood my meaning.

The Subjunctive is that of Ideal Certainty. This is doubt- less what Roby means in placing our example J under the "hypo- thetical Subjunctive," § 1544 (cf. § 1528 and 3). The corre- sponding English for our examples would then be "one would." This is also substantially what Kiesslin0 means in saying, upon example G, "valeas, Konj. der gemilderten Behauptung, ὑγιαίνοις ἄν," L. Muller in saying, upon the same example, "valeas; wie oft in Satiren und Episteln, urbane Milderung der Behauptung," Heinze, new edition (1906) of Kiessling's Scttires of Horace, in saying, "valeas scil. si parvo vivas," Rolfe op. cit., in slaying, "v aleas: potential subjunctive with an implied protasis, si tenui vietu utaris," Palmer op. cit., in saying, "valeas, a potential or modified future," Wickham Satires, Epistles, and De arte poetica of Horace, 1891, in saying, "valeas, potential," and Bennett Cato itior, 1897, in his note nipon L, "if you should be unable to practice this." This must likewise be what Wickham had in mind in saying of A, "the mood is probably the same as in the original, sis5-' aVlv S;7" though I fancy that this is a restored original, not a known one. And this is also probably what Heinze means, op. cit., in saying of the same example, "der potentiale Konjunktiv sis wie bei Petron. 77 asserm hctbeas assem valeas." (It will be noticed that Heiutze has abandoned Kiessling's explanation.)

The independent verb of example F, unum quom noris, omnis nor is, is explained by at least two of the commentators as not a Subjunctive at all, but an Indicative of the future perfect. Thus Dziatzko says, "noris ist beidemal Fut. exact;" and Elmer gives the same view, with the reasoning which brings him to it, as follows; "noris: in both cases fut. perf. ind. equivalent to fut. If the first noris were the perf. subj. of an indefinite second person, we should have omnis novisti, instead of omnis noris."

The error of procedure in all these explanations is of a kind that shows itself again and againi in the history of grammatical study. Those who have dealt with one or another example of a type have, in repeated instances, explained the example which they had before them, without stopping to make a collectiont of examples sufficient in number to enable a scientific inference of any kind to be drawn. The perfect method is, first, to make a complete collection, and theni to study every example, in the light of all. Human life being, however, what it is, one must, until our work is better organized, be content with a fairly large col- lection. In making this, one must throw aside grammatical pre- possessions, and admit all examples that would naturally seem to belong together. Further, it would be worth while to look, at least, at any sentences that might be found, in which ideas of the same general kind seemed to be expressed, but by a different mechanism. Further, and without any doubt, as much of the context of the Stibjunctive examples as could possibly be illumi- nating should be added for each. Here again our prevailing method is sadly defective. In a number of examples of the type which we are studying, the context immediately and irresistibly overthrows the explanations given. The material being thus gath- ered, every explanation that might account for one or more of the mass of examples, but would not account for all, or that in any example would fail to fit the context, should be rejected. The result of this process in the present instance will be the overthrow of all six of the explanations given for the examples cited above.

Assuming this for the moment to be the case, the next step should be to seek for an explanation that will explain all the examples, and that will be in harmony with the context in all. If one is thereby brought to believe in the existence of a con- struction not recognized by the grammars, one need not be alarmed, if one has confidence in the dispassionateness of the pro- cedure that has brought him to his conclusion.

Finally, one may properly seek for a probable origin of the usage thus reached, out of some construction or constructions known actually to exist. The two steps, however,-the establish- ment of the existence of a construction, and the search for its origin,-may be entirely disconnected. Failure to take the second step successfully will not of itself invalidate the first. A number of constructions are surely known to exist of which the origins are not surely known. In the present case, however, both steps can, I think, be securely taken.

In the collection which follows, the examples are arranged by groups of ideas, or for convenience in reference:

At bona pars hominum,d ecepta cupidine falso 'Nil satis est,' inquit, 'quia tanti quantum habeas sis.' Quid facias illi? Iubeas miserum esse libenter Quatenus id facit. Hor. Sat. i. 1. 61.

Aurum adque ambitio specimen virtutis viriquest. Quantum habeas, tantum ipse sies tantique habearis; Lucil. Schol. ad Iuv. iii. 143.

Mirum quin tu illo tecum divitias feras. Ubi mortuos sis, ita sis ut nomen cluet; Plaut. Trin. 495.

Tarda sunt quae in commune expostulantur: privatam gratiam statim mereare,s tatim recipias; Tac. Ann. i. 28. 21.

Nam si velis quod nondum vetitum est, timeas ne vetere: at si pro- hibita impune transcenderis,n eque metus ultra neque pudor est; Tac. Ann. iii. 54. 10.

Quid in amicitia fieri oportet, quae tota veritate perpenditur? In qua nisi, ut dicitur, apertum pectus videas tuumque ostendas, nihil exploratum habeas; Cic. Am. 26. 97.

Mediocriternec ausis nocent, cum . . . . adversarioruma diumenta confirmant? . . . . Si, quae sunt in iis invidiosa, . . . . invidi- osiora faciunt, quantum est in eo tandem mali? Quid, si .... contumeliosius invehare, nonne a te iudices abalienes? Quid, si . ... non intellegas te in iudices invehi, mediocre peccatum est? Quid, si . . . . causam relinquas, nihilne noceas 2. In ipsis autem argumentis si quid posueris . . . . aperte falsum .. . .nihilne noceas? Cic. De or. ii. 75, 303-306.

Liber captivos avis ferae consimilis est: Semel fugiendi si datast occasio, satis est, Numquam postilla possis prendere; Plaut. Capt. 116.

Nulla potest oculorum acies contenta tueri, Nec porro quae cumque aevo macieque senescunt, Nec, mare quae inpendent, vesco sale saxa peresa Quid quoque amittant in tempore cernere possis; Lucr. i. 324. Similarly ii. 220, 763, 768; iii. 856, 1024; iv. 572, 1231; vi. 113. Cf. the formulae vtidere licet, vi. 79; licet cognoscere, vi. 167, etc.

Hostes alienigenae aut oppressi serviunt aut recepti in amicitiam beneficio se obligatos putant: qui autem ex numero civium, dementia aliqua depravati, hostes patriae semel esse coeperunt, nec vi coerceren ec beneficio placare possis; Cic. Cat. iv. 10. 22.

Sed tamen est decorus senis sermo quietus et remissus, facitque persaepe ipsa sibi audientiam diserti senis composita et mitis oratio. Quam si ipse exsequi nequeas, possis tamen Scipioni praeciperee t Laelio; Cic. Sen. 9. 28.

Omnino si quicquam est decorum, nihil est profecto magis quam aequabilitasc um universae vitae, tum singularum actionum, quam conservare non possis (really an independent sentence) si aliorum naturam imitans omittas tuam; Cic. Off. i. 31. 111.

Ut non omnem frugem neque arboremi n omni agro reperire possis, sic non omne facinus in omni vita nascitur; Cic. Sex. Rosc. 27. 75.

Matronaep raeter faciem nil cernerep ossis, Cetera, ni Catia est, demissa veste tegentis; * * * * * * * Altera nil obstat: Cois tibi paene videre est, Ut nudam, ne crure malo, ne sit pede turpi; Metiri possis oculo latus. An tibi mavis Insidias fieri pretiumque avellier ante Quam mercem ostendi? Hor. Sat. i. 2. 94, 95, and 101-105.

Vix credere possis Quam sibi non sit amicus; Hor. Sat. i. 2. 19.

Accipe nunc, victus tenuis quae quantaque secum Adferat. In primis valeas bene. Hor. Sat. ii. 2. 70.

Quanta cuiusque animo audacia natura aut moribus inest, tanta in bello patere solet. Quem neque gloria neque pericula excitant, nequiquam hortere; timor animi auribus officit; Sall. Cat. 58. 2.

Non votis neque suppliciis muliebribus auxilia deorum parantur: vigilando, agundo, bene consulundo prospera omnia cedunt: ubi socordiae te atque ignaviae tradideris, nequiquam deos implores: irati infestique sunt; Sall. Cat. 52. 29.

Nam cetera maleficia tum persequare, ubi facta sunt; hoc, nisi provideris ne accidat, ubi evenit, frustra iudicia implores: capta urbe nihil fit reliqui victis; Sall. Cat. 52. 4.

Siquoi mutuom quid dederis, fit pro propriop erditum: Quom repetas, inimicum amicum invenias beneficio tuo; Plaut. Trin. 1051.

Neque ego homines magis asinos numquam vidi: ita plagis costae callent: Quos quom ferias, tibi plus noceas. Eo enim ingenio hi sunt fla- gritribae, * * * * * * * At faciem quom aspicias eorum,h au mali videntur; Plaut. Pseud. 136.

Sed eccum incedit: at quom aspicias tristem, frugi censeas; Plaut. Cas. 562. Similarly cum . . . . perventum est, tum intellegas; Cic. Div. ii. 21. 48.

Tene asymbolum venire unctum atque lautum e balineis, Otiosum ab animo, quom ille et cura et sumptu absumitur! Dum tibi fit quod placeat, ille ringitur: tu rideas, Prior bibas, prior decumbas; cena dubia adponitur; Ter. Ph. 339. 24 Abs quivis homine, quom est opus, beneficium accipere gaudeas; Verum enim vero id demum iuuat, si, quem aequomst facere, is bene facit; Ter. Ad. 254.

Minus placet magis quod suadetur: quod dissuadetur placet. Quom inopiast, cupias. quando eius copiast, tum non velis; Trin. 670.

CH. Vide modo ut hominem noveris. SY. Tamquam me: fieri istuc solet; Quod in manu teneas atque oculis videas, id desideres; Trin. 913.

Lippi illic oculi servos est simillumus. Si non est, nolis esse neque desideres: Si est, abstinere quin attingas non queas; Plaut. Bacch. 913.

De mortuis loquor, qui nulli sunt; nos, qui sumus, num aut cornibus caremus aut pinnis? Ecquis id dixerit? Certe nemo. Quid ita? Quia, cum id non habeas quod tibi nec usu nec natura sit aptum, non careas, etiamsi sentias te non habere; Cic. Tuse. i. 36. 87.

Unum quom noris, omnis noris; Ter. Ph. 265.

Licentiam des linguae, cum verum petas; Publil. 308. Sinmilarly [1] 10, 78, 81, 88,117, 185, 232, 233, 252, 285, 331, 361, 382, 394, 427, 468, 512, 530, 592, 635, 645; also, in perf. Subj., 126, 364.

Assem habeas, assem valeas; habes, habeberis; Petron. 77. [2]

[1] By the generally accepted readings. In a few cases, the MSS are divided.

[2] Habes is given in all of Bticheler's editions as the reading of the single MS (Parisin. 7989) which has this passage, and is reaffirmed by Btucheleri n a private letter. Omont has been good enough to report directly from the MS, to the same effect. The readings habeas and habe, respectively from the passage as repeated in Fried- lander's notes (habes in his text) and the Segebade and Lommatzsch LexciconP etro- nianum, are errors, as I am assured by Friedlitnder and Lomrnatzsch. But of course it is the habeberis that has value for the present purpose.

We proceed to test, in the light of this collection, the explana- tions that have thus far been given for this and that example taken by itself.

The construction can not be that of the Oratio Obliqua. This explanation was forced, even in the case of the first example. To many of the others, it can not be applied at all. Thus 3 can not mean "you are thought to be dead," 4 "you are thought to re- ceive at once," etc.

The Subjunctive can not be one of Command. Blase's ex- planation of velis in 25 seems to be due to a passing error of in- terpretation, and certainly can not be accepted. Even if it would there fit, however, it would not, by any stretching, fit in 1, 2, 32, and others. Lucilius, Horace, and Petronius do not mean to com- mand a man to be worth just what he possesses.

The only possible theory that could make these constructions to be of "Jussive" origin would be that they began as ironical commands, like "cry for what you don't have," and (probably) the second verb in "go further and fare worse." But Blase pre- sumably did not have this explanation in mind. And, in any case, it would be impossible, from such an origin, to account for the negative, except by supposing the construction to have lost every vestige of its initial feeling, and to have become a statement of fact. This would bring us to the actual force for which I am con- tending, though by a very improbable route.' [1]

[1] In Plut. irepZ q5\o7rXovTLas VII, p. 5260, Kgp5acVe Kat q5CeoV Kal TOO-o06Tovi v6 e cavrT?v &tLoiv, aozv& v EXes the advice is not ironical, but serious.

The Subjunctive can not be Permissive. The independent Subjunctive, if expressing the idea at all, would express Permission granted by the speaker. Such can not be the meaning in the ex- ample (23) for which Elmer gives this explanation. But, even if it were, it would be impossible to find the force in 1, 2, 3, etc. The meaning of 3, e. g., can not be "when you are dead, you are permitted to be dead in fact."

The Subjunctive can not be Potential, in the true sense. The true Potential is, to be sure, common enough in the second person singular indefinite. It indicates what lies within one's power, or within the possibilities of one's experience. Thus in Liv. xxii. 50. 9, cuneo quidemn hoc laxum atque solutunm agmen ut si nihil obstet disicias means, "by the use of the wedge-formation it lies within one's power to split this loose and disorganized line as if there were nothing in the way" (cf. w7AXla0-6AVX X7Xv. Kcat KceV &oiCTEv6etaEaP 102); and in Iuv. xiv. 41, Catilinam quocumque inpopulo videas, quocumque sub axe means "one may see (it lies within the possibilities of any man's experience to see) a Catiline in any nation, under any sky." It is also quite true that perse- qutare of 19 might by itself be interpreted, with Greenough and Daniell, as Potential (" it lies within one's power to punish other ill deeds at once") and 4, with Blase, op. cit. p. 142, in a similar way. But such a translation would be forced for frustra implores (in the same passage, 19, with per-sequare), making it mean "it lies within one's power to implore in vain," or "one might per- haps implore in vain; " while it would be completely out of the question for such examples as 1, 2, and 3, making them mean "it lies within your power to be worth as much as you possess," "it lies within your power to be dead in fact," etc.

A special word needs to be said with regard to the examples with possis. In such cases, the grammarians and commentators, as Blase op. cit., p. 142, upon our examples 8 and 9, and Munro upon 9, are too ready to give the explanation "Potential," con- fusing the meaning which the verb has in itself with the meaning of the mood employed in the particular instance. In videas of Verg. Georg. 1. 387, rightly cited by Blase in the same passage, the verb itself expresses the idea of seeing, while the mood conveys the idea of the possibility of that seeing. Possis, correspondingly, if the mood is to be really potential, would have to mean either "you have the power to have the power" or "it may possibly happen that you should have the power,"-meanings obviously out of the question here, and probably, in fact, nowhere occurring. It should be remembered, too, that by Roman idiom (apart from the usage which I am endeavoring to establish) the natural mood to use for possurm in saying " one may " would be the Indicative, potes.

The only possible way in which to deal with possis by any of the categories hitherto recognized in our grammars would be by interpreting it as meaning "one would be able." This is the Sub- junctive of Ideal Certainty, to which we pass.

The construction can not be one of IdealCertainty ("you would"). Such an explanation would, in the absence of any other, tolerably explain valeas of 16 ("you would in that case be well "), and might, with some forcing, be held to account for sev- eral of the examples with possis, and also for 17 ("it would be use- less to urge"), and for 221 ("you would think him good for some- thing"); but it would be unreasonable for 1, 2, 3, and many others. The natural meaning can not be "one would be worth what one should possess," etc. Still less reasonable would it be for 21. The speaker is actually flogging the slaves as he speaks. He is in no frame of mind to say, "if one should flog them, one would hurt one's self worse than them." And, finally, the explanation is obviously impossible for rideas, bibas, and decumbas of 23. The meaning cannot be, "his face is drawn with anxiety; you would laugh, you would drink before him, you would take your place at table before him."

The same reasoning would destroy any theory that should re- sort to the modified form of the Subjunctive of Ideal Certainty known as the "Subjunctive of Softened Assertion." This might, though not without strain, be adopted for velis of 25 and nolis of 27 (making them correspond to velim and nolim), and in conse- quence be resorted to for cupias of 25, and even for desideres of 26 and 27. But it would not fit the equivalent of desideres, namely, careas of 28; and it obviously would be impossible to find such a meaning in any other of the examples.

The construction in all the examples except 29 is obviously not an Indicative one. Dziatzko and Elmer would therefore not be through with their difficulty after explaining the ambiguous form noris as a Future Perfect.

These six explanations are all that suggest themselves as pos- sible to attempt, with the categories that we now have in the grammars. They prove to be insufficient. None of them will ac- count for all the examples. Their extreme difficulty, and the un- satisfactoriness of the whole method of procedure, become very obvious, if one sums up one's impressions by recalling that dif- ferent writers have given different explanations for the same ex- amples, that the same writer has at different times given different explanations for the same example, that certain writers have been willing to admit either of two explanations for the same example, and that certain writers (as Blase in the case of velis of 25 and valeas of 31) have put under different explanations examples obviously similar.

We must then set up a new tenet of some kind. In reaching this, three factors are to be taken into the account.

The natural interpretation of the passage in each case. It is upon such interpretations that all grammatical categories must ultimately be based. In the present set of examples the natural interpretation is sure. If our minds were grammatically dispas- sionate, a judgment founded on the reasonable demands of mean- ing in such verbs as sis in 1, habearis in 2, sis in 3, des in 30, and many others, would see in these examples plain statements of a general fact. The only reasonable interpretations are: "a man is valued according to his possessions" (1, 2); "when you are dead, dead you are" (3); "when you ask for the truth, you give a man a chance to say anything he pleases" (30); etc. And it is obviously an interpretation of this kind that Porphyrio had in mind in writing, in his comment upon iubeas, etc., of example 1, relinque eum, inquit, qui miser est, quonziam deduci ab hac opinione non potest, quo minus puttet tanti esse unumquemque, quantas divitias habet. The direct form of the opinio would evidently be, tanti est unusquisque, quantas divitias habet.

If anywhere else we can find other modes of expression obviously conveying the same general class of ideas as any of our Subjunctives under examination, we should consider their suggestions.

Now it happens that many of our examples are sententiae. One of these sententiae is the very familiar one (appearing three times in our collection) upon the relation between a man's pos- sessions and the estimation in which he is held. It can easily be matched.

Xp,qaa, XrapqLaT' LaVrp (scil. LcT7T), os fa K7TEaVWV 0dl/ca XEL40iELS Ka't At'Xw"v;'M oney, money (is) the man,' he said, 'when robbed at once of goods and friends;"' Pind. Isthmn2. . 11. OL yOZV 7rarTEpES rOVrt 7rp(LrOV rOLs 4(AETE'pOLtV LEtE0 7rapaLVOVOLVv,E 7rELoLV Ets T?V -'XLK'cV TaXffTa L3 COVTcLL TOV '7&) cfPOVELV, US 8OKOVcTL, cKO7TELV OOEV 7rX61`t0Lt EcTOVTaL, (S) aaV Et EX-7%Sa cL$ 0o TOV E tIEv 8 v6 Oi3vo' Eryx. Pseudoplat. 396 C.

Non quare et iinde, quid habeas, tantum rogant.

Ubique tanti quisque quantum habuit fuit; Sen. Ep. Mor. 115. 14. from the lost Danae of Euripides (Nauck. T. G. F; p. 324).

Pauper ubique iacet; Ov. Fast. 1. 218.

Dat census honores; Ov. Amr.I II. 8. 55 and Fast. 1. 217.

Quisquis habet nummos, secura navigat aura; Petron. 137.

Quantumq uisque sua nummorums ervat in arca,

Talntumh abet et fidei; Iuv. iii. 143.

Unde et illud proverbium: Quantum habebis, tantus eris; Augustin. De disc. Christ. 11. 12.

The second Greek example is an especially striking one, since, though the main sentence (a general one) is cast in the third person plural, the general second singular indefinite is retained in the dependent clause, and is in the Indicative. The phrase illutd proverbium in the last example should also be noted.

It would seem, then, that the natural idea to come into the mind in expressing the relation between a man's money and the estimation in which the world holds him is that of a general statement of fact.

Several of the examples refer to individual persons. Now among them are certain ones in which the ideas, down to the end of the introductory clause, look alike. Thus at quom aspicias tristem of 22 is substantially like atfaciem quomn aspicias eorum of 21. It would seem somewhat probable, then, that the con- cluding clauses would express similar ideas. These clauses are frugi censeas and hau mali videntur. They prove, as expected, to be alike in general force. It is probable, then, that the mood- force is the same, namely, that of a statement of fact.

These are the first factors in the determination of the nature of the idiom under discussion. The most natural interpretation to put upon these Subjunctives, whether in and by themselves, or in the light of the expression of the same general ideas by other means, is that they are statements of fact.

The third factor in the determination of the force of the mood lies in the study of the context of such of the examples as possess a closely attached one, of the same general bearing. This factor is of the same essential nature as the second, but is more decisive. It is, in fact, completely sufficient, by itself alone, to determine the answer to our question. It happens that in many cases the Subjunctive under examination is put into exact paral- lelism with a general statement of fact in the Indicative. Thus, in 8, possis is the amplification of satis est (not sit); in 5, timeas is balanced against neque metus . . . . est, and, in 18, pros- pera omnia cedunt against nequiquam deos implores; in 24, gautdeas is set up as the antithesis of iuvat; in 23, rideas, bibas, and decutmbas are hemmed in between the two general statements ringitur and adponitur, with the parallelism still more strongly brought out by the antithesis ille . .. . tu . . . in 19, frustra iudicia implores is justified by the general state- ment fit, and, in 20, the general statement fit by invenias; in 14, matronae . . . . nil cernere possis is set over against altera nil obstat, and videre est is carried out by metiri possis ocutlo; in 9, nulla potest oculorum ac es . . . . tueri is clearly parallel to nec cernere possis, and, in 13, ut non . . in omni agro reperire possis to sic non .... in omni vita nascitur; while in 17, not only is hortere balanced against a general statement officit, but the two together form the unfolding of the contents of the very significant Indicative solet, "it is regularly the case that." Note similarly solet . . . . desideres in 26. In 28, careas is balanced against caremus. In 7, abalienes and nihilne noceas (bis) are parallel with mediocriterne nocent, quantum est mali, and mediocre peccatum est. And, finally, in 31, the correspond- ence between valeas and habeberis is so obvious that I am almost ashamed to have printed anything of the present paper except the title and this example.

The condition of things being what has been seen, there can be no reasonable doubt that the second person singular of the Latin Subjunctive may be employed to make independent state- ments of a general fact.

Such a conclusion should afford a sensible relief. The variety of explanations adduced above as having been actually applied to one and another of these examples, the heavily forced character of some of them, even for the individual cases to which they are applied, the contradictory explanations made by different writers for the same examples, the shift of ground on the part of one writer,-all show how hard beset grammarians and commentators have been.

In the collections employed, only examples that are surely to be classified as statements of fact are given. Once established, however, the category fits many other examples more naturally than does the category of Possibility, or that of Ideal Certainty. The more important step has been taken. It remains to see whether the second is possible,-whether a probable origin can be found for the Subjunctive construction. I have long taught the following:

The extremely common use of the Subjunctive of the second singular indefinite in generalizing assumptions (relative or tem- poral clauses, or express conditions) might well, in time, affect the mood of the main sentence. To express the generalizing idea in the dependent clause, the verb in any person or number except the second singular indefinite was in the Indicative, but if it oc- curred in this form, then the mood was, by fixed force of habit, the Subjunctive. The Subjunctive would then seem to be a natural mood to use wherever the generalizing force was intended, if the second singular indefinite were employed. Just as else- where, in the expression of this idea, Indicative went with Indica- tive, so here it would seem natural that Subjunctive should go with Subjunctive. If, saying quod quis in manu tenet, one goes on with id desiderat, then, saying quod in manu teneas, it would be natural to go on with id desideres.

This influence would be all the stronger from the fact that, whatever may have been the original force of the Subjunctive in these dependent clauses [1], that force could no longer be clearly felt. The Subjunctive of the second singular indefinite had be- come merely the sign of universality. Why, then, should it re- niain confined to the subordinate clause?

[1] I believe the construction to have originated in a Volitive of the imagination, as in "lose your money, and all your friends desert you" (not, as seems everywhere as- sumed, in the Potential Subjunctive); and I have so classed it in the syntax of the Hale-Buck Grammar, § 504. 2. But a treatment of this question would properly involve a discussion of the Greek generalizing Subjunctive (which I hold to be of parallel origin), and a discussion of my difference from Delbruick with regard to an impor- tant group of Greek constructions, for which this paper affords no room.

This natural tendency toward extension would probably be fur- thered by the fact that the same second person singular indefinite happens to be in frequent use in a number of true independent Subjunctive idioms. Thus it is found in the Subjunctive Question of Perplexity, as in quid agas? "what is one to do?" Iuv. iv. 14; in the Question of Obligation, Propriety, or Reasonableness (to avail myself of a classification and name from own teaching) [2], as in cur tamen corripias? "Why should one chide her?" Iuv. x. 291; and in the form of statement which lies in the neutral terri- tory of Natural Likelihood, Possibility, and Ideal Certainty, as in crederes victos, "one might well think them conquered," "one might think them conquered," "one would think them conquered," Liv. ii. 43. 9.

[2] The Subjunctive of Obligation, Propriety, or Reasonableness, first set up as a general category in my teaching at Cornell, is made one of the principal families of the mood in my general scheme, and several of the hitherto difficult types of dependent clauses seem to find full solution in it (Hale-Buck Grammar, ?§ 512,5 13; also Classical Journal I. 1. 1905, "An Experiment in the Teaching of First and Second Year Latin," where a conspectus is given of the uses of the Latin Subjunctive, classified according to origins). Not only is the mood-use a frequent one, but,,in my opinion, it furnishes the missing link between the true Optative and the Optative (Latin Subjunctive) of Natural Likelihood, Possibility, or Ideal Certainty. (See my "Leading Mood-Forces in the Indo-European Parent Speech," Proceedings of the American Philological Association XXXII, p. cxx [1901].)

The combination of these influences with the one already mentioned might well have led the Romans to feel that, whatever mood-feeling they had to convey, the Subjunctive was the natural mood to use, if the person was the second singular indefinite. But, most of all (after the primary cause), the natural tendency to extend the idiom would be strengthened by the very frequent occurrence of a construction which, though evidently of true Potential origin, clearly suggests and implies fact. Thus Juvenal's Potential in Catilinam quocumque in poputlo videas, quocumque sub axe," "one may see a Catiline in any nation, under any sky," xiv. 41, easily suggests "one sees . . ." That which is said to be anywhere and everywhere possible is easily understood as anywhere and everywhere taking place. The number of ex- amples of this type (videas, cernas, etc.) is very large. An especi- ally instructive set is afforded by reperias. I give three of these in sequence:

Quamvis malam rem quaeras, illic reperias; Plaut. Trin. 554.

Per pol quam paucos reperias meretricibus Fidelis evenire amatores, Syra; Ter. Hec. 58.

Ut enim sunt, quem ad modum supra dixi, qui urbanis rebus bellicas ante- ponant, sic reperias multos, quibus periculosae t calidac onsilia quietis et cogitatis splendidiora et maiora videantur; Cic. Off. i. 24. 82.

In the first of these, "one may find it there," may suggest "one finds it there." In the second, "how few one can find," strongly suggests "how few one |finds;" while in the third, reperias qui- bus . . . . is put into formal parallelism with the general state- ment sunt qui . . . . , "there are people who . . . ." Reperias accordingly is completely ready, at any rate, to convey the idea of fact.

So much for the construction under examination. An inde- pendent use of the Subjunctive to express statements-assertions of fact-in the second person singular indefinite, exists; and it probably is historically an extension from other constructions in which the Subjunctive had, originally, true-mood forces.[1]

[1] Hence in my grammatical scheme, referred to above, this use is placed under the distinct heading of "Subjunctive Constructions Due to the Influence of Other Constructions."

But did the construction become universal, where the idea of a general statement of fact was to be conveyed, and the second singular indefinite was to be used? It seems not. The older usage, which of course must have been that of the Indicative mood, still survived. In the future, the idea had necessarily to be expressed by the Indicative if expressed at all, since the in- dependent Anticipatory Subjunctive had disappeared, wholly or almost wholly, from the language. But there are also examples of the present.

Examples of the Present Indicative [2]

Nescis quid vesper serus vehat; Varro Sat. Men. ap. Gell. xiii. 11. 1; similarly, but without serus, in Macrob. i. 7. 12, and ii. 8. 2. (It would seem that nescis must be general, since in all three passages the sentence to which it belongs is given as the title of a book.)

Bis peccas, cum peccanti obsequium commodas; Publil. 52.

Cum das avaro praemium ut noceat rogas; Publil. 122. Similarly 658. [3]

Ita, si silenda feceris, factum ipsum (scil. culpatur), si laudanda non sileas, ipse culparis; Plin. Ep. i. 8. 15.

[2] Examples are not given from the satiric poets, since they appear often to have been addressing, either through a long passage or for the moment, an imagined indi- vidual (cf. our "Ig entle reader "), and the exact discrimination of force is therefore uncertain for the Indicative mood. See, for instance, in 14 above, and cf. Lebreton Etudes sur la langue et la grammaire de Ciceron, pp. 349 ff.

[3] Publilius's Indicatives in the dependent second person singular indefinite are in- teresting, and will be treated, as well as Blase's view that there is no second person singular Subjunctive of Indefiniteness in dependent Clauses (Studien u. Kritiken II, pp. 5-8, 1905), in a later paper already indicated.

Examples of the Future Indicative

Quantumvisq uare sit macer invenies; Catull. 89. 6.

Nulli facilius quam malo invenies parem; Publil. 395. [4]

Facile invenies qui bene faciant, cum qui fecerunt coles; Publil. 664. [4] divitiae et opes, quas facilius invenies qui vituperet quam qui fastidiat; Tac. Dial. 8.

[4] In the light of the other examples given, W6lfflin's " emendation" to invenias is clearly unnecessary.

Ideo autem (apes) pungunt, quia ubicumque dulce est, ibi et acidum invenies; Petron. 56.

Assem habeas, assem valeas: habes, habeberis; Petron. 77.

Unde et illud proverbium: Quantum habebis, tantus eris; Augustin. De disc. Christ. 11. 12.

My collections for all these idioms are the result of chance happenings upon examples, not of methodical search over the whole ground. They make it fairly clear, however, that the Subjunctive is distinctly common, the Present Indicative, outside of Publilius and the satiric poets (see the last footnote but two), very rare. The statement in the Hale-Buck Grammar is accord- ingly not strong enough. For "sometimes," "regularly" should be substituted. It should be noted also that, of the various verbs employed in the Subjunctive idiom, possis, " one can," seems to be in largest use.

Finally, can any definition be made of the difference between the Subjunctive and Indicative idioms? The question may be answered as follows:

In the Subjunctive construction, the mood is a sign of gener- ality. In the Indicative construction, the mood is not a sign of generality. One feels generality, if the context demands it; but it is not hinted at by the mood itself. The difference in the idioms is like the difference between English "you" and "one." " You " may turn out to be either individual or general. "One " is immediately and necessarily general. Compare the "Tacit causal qui-clause," Indicative, and the "Explicit causal qui- clause," Subjunctive (Hale, Cum-constructions, pp. 97, 112 f., or 107, 127 f. of the German translation).

An Unrecognized Construction of the Latin Subjunctive

An Unrecognized Construction of the Latin Subjunctive

The Second Person Singular in General Statements of Fact

Author: William Gardner Hale

Source: Classical Philology I, January 1906, pp. 21-42.

The Second Person Singular in General Statements of Fact

Author: William Gardner Hale

Source: Classical Philology I, January 1906, pp. 21-42.

The Second Person Singular in General Statements of Fact

The Second Person Singular in General Statements of Fact

My paper deals with such Subjunctives as the independent ones in the familiar passages: A. Nil satis est, inquit, quia tanti quantum habeas sis; Hor. Sat. i. 1. 62. B. Assem habeas, assem valeas; Petron. 77. C. Ubi mortuos sis, ita sis ut nomen cluet; Plaut. Trin. 496. D. Quom inopiast, cupias: quando eius copiast, tum non velis; Plaut. Trin. 671. E. Quod in manu teneas atque oculis videas, id desideres; Plaut. Trin. 914. F. Unum quom noris, omnis noris; Ter. Ph. 265. G. Accipe nunc, victus tenuis quae quantaque secum Adferat. In primis valeas bene; Hor. Sat. ii. 2. 70. H. Dum tibi fit quod placeat, ille ringitur: tu rideas, Prior bibas, prior decumnbas; Ter. Ph. 341. I. Nam cetera maleficia tum persequare, ubi facta sunt, hoc, nisi provi- deris ne accidat, ubi evenlit, frustra iudicia implores; Sall. Cat. 52.4. J. Quem neque gloria neque pericula excitant, nequiquam hortere; timor animi auribus officit; Sall. Cat. 58. 2. K. Nec porro quaecumque aevo macieque senescunt, Nec, mare quae inpendent, vesco sale saxa peresa Quid quoque amittant in tempore cernere possis; Lucr. i. 325. L. Quae si ipse exsequi nequeas, possis tamen Scipioni praecipere et Laelio ("a Scipio and a Laelius "); Cic. Sen. 9. 28.

These look alike. If there were a comfortable category in the grammars into which to put them all, one would not think of separating them. As the title of the paper implies, I am brought to regard such Subjunctives as expressing general statements of fact,-as corre- sponding, in effect, to Indicatives in the same second person singu- lar indefinite. I wish to translate example A by, "because you are worth just what you possess ;" B by, "have a penny, you are worth a penny ;" C by, "when you are dead, dead you are," etc. In accordance with this view, § 542 of the Hale-Buck Grammar reads: "A general statement of fact is sometimes expressed by a Subjunctive of the Second Person Singular Indefinite;" and to our example C, there given, is added the statement, "the second sis has the force of es."

The construction itself will prove to be a simple one to solve, but the exhibition of previous opinion upon the matter must be a somewhat tangled affair, since grammars, grammatical treatises, notes in commentaries written by professional grammarians, and notes by commentators who are not professional grammarians, have all to be taken into the account. It is also obviously impracti- cable to refer to every note ever written upon any of these pas- sages. The selections, however, will fairly cover the ground.

So far as I know, no grammar or grammatical paper had rec- ognized the existence of the construction for which I am contend- ing. The statements regularly made (where any are made) with regard to the range of the Subjunctive of the second person singu- lar indefinite are of a kind to exclude such an interpretation. Thus Madvig, § 370, says: "Diese Form findet sich in bedingter Rede, in hypothetischenAussagen und in Fragen uber das, was geschehen wird und kann, in Nebensatzen mit Conjunctionen, in Relativsat- zen (mit qui oder einem unbestimmten Relativ), und in Vor- schriften und Verboten." To this is added the Anmerkung: "Ein solcher conjunctivischer Bedingungssatz fuhrt nicht den Con- junctiv im Hauptsatze herbei." Quite recently (1905), Nutting, Studies in the Si-Clause, pp. 84, 85, has distinctly shown the conviction that there is no Subjunctive in a main sentence corre- sponding to the common second singular indefinite in subordinate clauses. "One looks in vain," he says, "for a clear case of such leveling in Plautus." And Elmer, in his note on F (as will be seen below), in effect says that a general statement of fact in the second singular indefinite must be expressed by the Indicative.

The grammarians of course all provide for the Subjunctive second person singular indefinite in conditional or corresponding relative clauses which, if in any other person or number, would be in the Indicative, as in memoria minuitur, nisi eam exerceas, Cic. Sen. 7. 21; but they know nothing of the power for which this paper contends. If anywhere they chance to make use of an ex- ample of the type before us, it is placed under some already recog- nized and entirely different category, as we shall see below.

Whatever the grammarians may do, the commentators, it would seem, would frequently have occasion to discuss such examples. The most striking feature of the case is, however, that they gen- erally pass them without comment. Sometimes, indeed, they actually give an explanation of the dependent verb in such com- binations, and pass in silence the far more difficult independent one. Thus Greenough, in his edition of the De senectute (1873 and in various reprints), says of our example L, "si . . . . nequteas, gen. cond. (§ 309 a)," but says nothing of possis. Occasionally the dependent verb and the main verb are dealt with together and without distinction, with a reference to some rule, such as the grammars afford. This reference, as is natural under the circum- stances, never fits. Thus Gudeman, dealing with example I in his edition of Sallust, says, " persequare . . . . provideris inplores: Subjunctive of indefinite second person singular," and refers to the Allen and Greenough Grammar, 266a. This section, when one turns to it, proves to deal with independent "EIortatory" Subjunctives, and to read (266), " The Hortatory Subjunctive is used to express an exhortation, a command, a con- cession, or a condition;" and (a), "The Second Person is used only of an indefinite subject, except in prohibition, in early Latin, and in poetry." Not only the general headings "independent" and "'h ortatory," but the examples given (e. g., isto bono utare dum adsit, Cic. Sen. 10. 33), show the type meant. But nisi provideris is evidently not independent; and it is also difficult to believe that Sallust means to direct people to implore in vain, unless, indeed, the editor conceives the idiom to be one of ironical command (not provided for in the reference), in which case the explanation should certainly be given. More frequently, if any- thing at all is said about the main Subjunctive, it is simply classed with the dependent one, and the Subjunctive idiom in the second singular indefinite in " conditional " clauses is referred to or assumed. Thus Munro, in his comment upon our example K, puts possis with si non relinquas, Lucr. i. 515, quibus careas, ii. 4, etc. Thus, again, Greenough, in the note in his edition of the Satires and Epistles of Horace, 1888, says of our example A, "tanti sis (you are r ated att), etc. The subjunctive is the regular one of the second person with indefinite subject." But in his Gram- mar Greenough (the same is true of his revisers) makes no pro- vision for an independent Subjunctive of the second person singular indefinite, except as hortatory or potential; while the treatment of the dependent member of the example in question comes in § 309 a (518 of the new edition) under the head of "General Conditions," to which it is obvious that sis does not belong. It is clear, also, that general conditions are really meant; for the rule reads: "The Subjunctive is sometimes used in the second person singular to denote the act of an indefinite Subject (you =any one). Here the Indicative of a general truth may stand in the apodosis."'

Yet the tendency naturally produced by the sight of the com- bination of dependent and independent Subjunctives in the same indefinite second person is a good one, and has in a number of instances led, even in the absence of an admitted grammatical category, to the solid interpretation. Thus Wagner, in the note in his Trinummus (1875) upon our example E, says, " id desid- eres, 'that one is apt to forget': the second person subjunctive expresses generality." Greenough (1888), as we have seen, translates tanti sis of A by " you are rated." Friedlander, in his Petronii Centa Trimalchionis (1891) translates our example B by 4"G laubt mir, habe einen As, so giltst du einen As." Morris, in his edition of the Captivi and Trinummus of Plautus (1898), paraphrasing example C, says, " when you are dead, you are dead in the full sense of the term." He makes no grammatical com- ment. Rolfe, in his edition of the Satires and Epistles of Horace

[1] The word " sometimes " is not happy. The Subjunctive is regularly used, there being but a very small number of exceptions in the whole range of the literature. Hardly any usage in Latin is more fixed. Or, if the intended meaning is " one some- times finds examples of . " then, by the same token, the word " sometimes" ought to be added to every rule in the Grammar. Moreover, the phrase " may stand" in the last line should read " generally stands."

(1902) puts the main and subordinate clauses in example A to- gether, rightly translating, " because you are rated by the amount of your possessions, " and adding, "habeas and sis are subjunctive because of the indefinite second person singular." No grammati- cal reference is given. The procedure of Wagner, Greenough, Friedlander, Morris, and Rolfe seems to imply a subconsciousness of the force naturally to be attributed to the independent con- struction; but these excellent opportunities which Greenough and Morris, as grammarians, had for calling attention to a much mis- understood, and very variously interpreted, construction, and of enunciating its true character, pass unused. Note also, under 4 below, p. 27, a discordant explanation of the exactly similar per- sequtare of example I (namely as Potential) in the Greenough and Daniell Sallust.

The only comment I have noticed which makes a sound ex- plicit statement about the independent construction is that of Sch-itz (1881), who, in his note upon our passage A, says "sis, nicht es, weil allgemein, nicht an H. direct gerichtet; 'man gilt."' To this he adds, "Im Indicativ wulrde man eher die 1. Person verlaingen, weil der Geizige doch sich selbst damit entschuldigen will." He is right in his general feeling, but not right in imply- ing that in the second singular indefinite the Indicative could not be used. See below, at the end of this paper.

The explanations which clearly attempt to bring one or arnother of the independent Subjunctives in sentences A-L under one or another of the categories provided for in the grammars are of six kinds. I shall first merely state them, without argument.

The Subjunctive is that of the Oratio Obliqua. This is the favorite older explanation, and is still frequently found. Thus Wusteman, revision of Heindorf (1843), says of our example A, "man bemerke den im Latein. gew6hnlichen Uber- gang von der orat. recta in die obliqua;" H. Duntzer (1869) says "sis, Ubergang in die abhangige Rede;" Kirchner (1885) says "im Lateinischen geht die Rede aus der directen Construc- tion (nil satis est) in die abhangige (quia tanti . . . . sis) uiber, was auch in Prosa nicht selten vorkommt;" and G. T. A. Kruger in various editions (as in 1860 and 1866), by a still bolder device, says "sis, statt es, als ware vorhergegangen: nil satis esse ait. Also auf einer Anakoluthie beruhend." G. Krufger says the same, in re-editing his father's edition in 1876. Among the later editors, Kiessling (1886) says the same in the sentence, "der Konjunktiv sis = esse putaris." Palmer (my edition is of 1888) seems to mean the same by his phrase "general opinion" in saying, "the subjunctive sis is used because it is the general opinion, indefinite." And Harper's Lexicon (1879) means the same, in putting example A under the Subjunctive illustration of the statement, "construed with indicative in asserting a fact; with subjunctive in stating an assumed reason, or one entertained by another mind."

The Subjunctive is that of Command or Exhortation ("Imperative"). Thus Blase, in his recently published "Tem- pora u. Modi" in the Historische Grammatik d. lat. Sprache, p. 136, puts our D as his first example under the head of the Jussive, Hortative, or (true) Optative, and accounts for non as modifying, not the sentence as a whole, but the single word velis. The same explanation seems to be meant, p. 123, for C, which is translated by "so bald man tot ist, soll man es ganz sein." And the same is explicitly suggested by Antoine and Lallier as possible for I. See below, at the end of 4.

The Subjunctive is Permissive. Thus Elmer, on example H, says, "'rideas, bibas, etc.: subj. of permission, involving the yielding of somebody's will; to be carefully distinguished from the potential."

The Subjunctive is Potential. It is by the Potential formula ("you can laugh, drink your wine before him," etc.) that Morris H. Morgan, in his translation of the Phormio (1904), renders these same Subjunctives which Elmer tells us are "to be carefully distinguished from the potential." The express word "Poten- tial" is used of one or another of our examples by many commen- tators. Thus Dillenburger (my edition is of 1867) both affirms this explanation and denies 1 above, in the note: "sis non per anacoluthiam positum est pro es, quasi praecederet nil satis esse ait, qutia, sed est potentialis modus, ut apud Lucilium." Likewise G. Krttger, who in the edition of 1876 explained sis in A as due to Oratio Obliqua, explains it in the edition of 1889 as Potential. This same explanation is given by Lucian Miller (1891) for A and B, by Blase op. cit., p. 142, for B, by SchtVtz op. cit. for G, and by Schmalz Bell. Cat. (my edition is of 1885) for J.

Unfortunately this word "Potential," as used in our gram- mars, is ambiguous. It includes meanings so far apart, at their extremes, as "may possibly" or "can," and "would certainly." The former idea alone should bear the name Potential, while the other (seen most easily in the common Subjunctive conclusion) should bear some name like that of Ideal Certainty, which I have given it in my own grammatical writing. [1] It is the true Potential use, apparently, that Kruger and Miller have in mind, and likewise Freeman and Sloman upon our example D. This, at any rate, is what is meant by Greenough and Daniell in their Sallust, in the note upon persequare in our example I. The translation given is "one may prosecute," and the reference (Allen and Greenough Grammnar 311a) is to the statement, "The Potential Subjunctive is used to denote an action not as actually performed, but as possible." Similarly Antoine and Lallier, Salluste, Co7ijuration de Catilina (1888), write in their commentary: "persequare pourrait etre un subjonctif imp6ratif, puisque le sujet est indetermin6; mais en réalité, comme le note M. P. Thomas, c'est le subjonctif potentiel: 'on peut ne les pour- suivre que lorsqu'ils sont accomplis.'"

[1] 'Delbruiek Verg. Syntax II, p. 371, and Brugmann Kurze vergl. Gramm. II, p. 584, while accepting my term, have misunderstood my meaning.

The Subjunctive is that of Ideal Certainty. This is doubt- less what Roby means in placing our example J under the "hypo- thetical Subjunctive," § 1544 (cf. § 1528 and 3). The corre- sponding English for our examples would then be "one would." This is also substantially what Kiesslin0 means in saying, upon example G, "valeas, Konj. der gemilderten Behauptung, ὑγιαίνοις ἄν," L. Muller in saying, upon the same example, "valeas; wie oft in Satiren und Episteln, urbane Milderung der Behauptung," Heinze, new edition (1906) of Kiessling's Scttires of Horace, in saying, "valeas scil. si parvo vivas," Rolfe op. cit., in slaying, "v aleas: potential subjunctive with an implied protasis, si tenui vietu utaris," Palmer op. cit., in saying, "valeas, a potential or modified future," Wickham Satires, Epistles, and De arte poetica of Horace, 1891, in saying, "valeas, potential," and Bennett Cato itior, 1897, in his note nipon L, "if you should be unable to practice this." This must likewise be what Wickham had in mind in saying of A, "the mood is probably the same as in the original, sis5-' aVlv S;7" though I fancy that this is a restored original, not a known one. And this is also probably what Heinze means, op. cit., in saying of the same example, "der potentiale Konjunktiv sis wie bei Petron. 77 asserm hctbeas assem valeas." (It will be noticed that Heiutze has abandoned Kiessling's explanation.)

The independent verb of example F, unum quom noris, omnis nor is, is explained by at least two of the commentators as not a Subjunctive at all, but an Indicative of the future perfect. Thus Dziatzko says, "noris ist beidemal Fut. exact;" and Elmer gives the same view, with the reasoning which brings him to it, as follows; "noris: in both cases fut. perf. ind. equivalent to fut. If the first noris were the perf. subj. of an indefinite second person, we should have omnis novisti, instead of omnis noris."

The error of procedure in all these explanations is of a kind that shows itself again and againi in the history of grammatical study. Those who have dealt with one or another example of a type have, in repeated instances, explained the example which they had before them, without stopping to make a collectiont of examples sufficient in number to enable a scientific inference of any kind to be drawn. The perfect method is, first, to make a complete collection, and theni to study every example, in the light of all. Human life being, however, what it is, one must, until our work is better organized, be content with a fairly large col- lection. In making this, one must throw aside grammatical pre- possessions, and admit all examples that would naturally seem to belong together. Further, it would be worth while to look, at least, at any sentences that might be found, in which ideas of the same general kind seemed to be expressed, but by a different mechanism. Further, and without any doubt, as much of the context of the Stibjunctive examples as could possibly be illumi- nating should be added for each. Here again our prevailing method is sadly defective. In a number of examples of the type which we are studying, the context immediately and irresistibly overthrows the explanations given. The material being thus gath- ered, every explanation that might account for one or more of the mass of examples, but would not account for all, or that in any example would fail to fit the context, should be rejected. The result of this process in the present instance will be the overthrow of all six of the explanations given for the examples cited above.

Assuming this for the moment to be the case, the next step should be to seek for an explanation that will explain all the examples, and that will be in harmony with the context in all. If one is thereby brought to believe in the existence of a con- struction not recognized by the grammars, one need not be alarmed, if one has confidence in the dispassionateness of the pro- cedure that has brought him to his conclusion.

Finally, one may properly seek for a probable origin of the usage thus reached, out of some construction or constructions known actually to exist. The two steps, however,-the establish- ment of the existence of a construction, and the search for its origin,-may be entirely disconnected. Failure to take the second step successfully will not of itself invalidate the first. A number of constructions are surely known to exist of which the origins are not surely known. In the present case, however, both steps can, I think, be securely taken.

In the collection which follows, the examples are arranged by groups of ideas, or for convenience in reference:

At bona pars hominum,d ecepta cupidine falso 'Nil satis est,' inquit, 'quia tanti quantum habeas sis.' Quid facias illi? Iubeas miserum esse libenter Quatenus id facit. Hor. Sat. i. 1. 61.

Aurum adque ambitio specimen virtutis viriquest. Quantum habeas, tantum ipse sies tantique habearis; Lucil. Schol. ad Iuv. iii. 143.

Mirum quin tu illo tecum divitias feras. Ubi mortuos sis, ita sis ut nomen cluet; Plaut. Trin. 495.

Tarda sunt quae in commune expostulantur: privatam gratiam statim mereare,s tatim recipias; Tac. Ann. i. 28. 21.

Nam si velis quod nondum vetitum est, timeas ne vetere: at si pro- hibita impune transcenderis,n eque metus ultra neque pudor est; Tac. Ann. iii. 54. 10.

Quid in amicitia fieri oportet, quae tota veritate perpenditur? In qua nisi, ut dicitur, apertum pectus videas tuumque ostendas, nihil exploratum habeas; Cic. Am. 26. 97.

Mediocriternec ausis nocent, cum . . . . adversarioruma diumenta confirmant? . . . . Si, quae sunt in iis invidiosa, . . . . invidi- osiora faciunt, quantum est in eo tandem mali? Quid, si .... contumeliosius invehare, nonne a te iudices abalienes? Quid, si . ... non intellegas te in iudices invehi, mediocre peccatum est? Quid, si . . . . causam relinquas, nihilne noceas 2. In ipsis autem argumentis si quid posueris . . . . aperte falsum .. . .nihilne noceas? Cic. De or. ii. 75, 303-306.

Liber captivos avis ferae consimilis est: Semel fugiendi si datast occasio, satis est, Numquam postilla possis prendere; Plaut. Capt. 116.

Nulla potest oculorum acies contenta tueri, Nec porro quae cumque aevo macieque senescunt, Nec, mare quae inpendent, vesco sale saxa peresa Quid quoque amittant in tempore cernere possis; Lucr. i. 324. Similarly ii. 220, 763, 768; iii. 856, 1024; iv. 572, 1231; vi. 113. Cf. the formulae vtidere licet, vi. 79; licet cognoscere, vi. 167, etc.

Hostes alienigenae aut oppressi serviunt aut recepti in amicitiam beneficio se obligatos putant: qui autem ex numero civium, dementia aliqua depravati, hostes patriae semel esse coeperunt, nec vi coerceren ec beneficio placare possis; Cic. Cat. iv. 10. 22.

Sed tamen est decorus senis sermo quietus et remissus, facitque persaepe ipsa sibi audientiam diserti senis composita et mitis oratio. Quam si ipse exsequi nequeas, possis tamen Scipioni praeciperee t Laelio; Cic. Sen. 9. 28.

Omnino si quicquam est decorum, nihil est profecto magis quam aequabilitasc um universae vitae, tum singularum actionum, quam conservare non possis (really an independent sentence) si aliorum naturam imitans omittas tuam; Cic. Off. i. 31. 111.

Ut non omnem frugem neque arboremi n omni agro reperire possis, sic non omne facinus in omni vita nascitur; Cic. Sex. Rosc. 27. 75.

Matronaep raeter faciem nil cernerep ossis, Cetera, ni Catia est, demissa veste tegentis; * * * * * * * Altera nil obstat: Cois tibi paene videre est, Ut nudam, ne crure malo, ne sit pede turpi; Metiri possis oculo latus. An tibi mavis Insidias fieri pretiumque avellier ante Quam mercem ostendi? Hor. Sat. i. 2. 94, 95, and 101-105.

Vix credere possis Quam sibi non sit amicus; Hor. Sat. i. 2. 19.

Accipe nunc, victus tenuis quae quantaque secum Adferat. In primis valeas bene. Hor. Sat. ii. 2. 70.

Quanta cuiusque animo audacia natura aut moribus inest, tanta in bello patere solet. Quem neque gloria neque pericula excitant, nequiquam hortere; timor animi auribus officit; Sall. Cat. 58. 2.

Non votis neque suppliciis muliebribus auxilia deorum parantur: vigilando, agundo, bene consulundo prospera omnia cedunt: ubi socordiae te atque ignaviae tradideris, nequiquam deos implores: irati infestique sunt; Sall. Cat. 52. 29.

Nam cetera maleficia tum persequare, ubi facta sunt; hoc, nisi provideris ne accidat, ubi evenit, frustra iudicia implores: capta urbe nihil fit reliqui victis; Sall. Cat. 52. 4.

Siquoi mutuom quid dederis, fit pro propriop erditum: Quom repetas, inimicum amicum invenias beneficio tuo; Plaut. Trin. 1051.

Neque ego homines magis asinos numquam vidi: ita plagis costae callent: Quos quom ferias, tibi plus noceas. Eo enim ingenio hi sunt fla- gritribae, * * * * * * * At faciem quom aspicias eorum,h au mali videntur; Plaut. Pseud. 136.

Sed eccum incedit: at quom aspicias tristem, frugi censeas; Plaut. Cas. 562. Similarly cum . . . . perventum est, tum intellegas; Cic. Div. ii. 21. 48.

Tene asymbolum venire unctum atque lautum e balineis, Otiosum ab animo, quom ille et cura et sumptu absumitur! Dum tibi fit quod placeat, ille ringitur: tu rideas, Prior bibas, prior decumbas; cena dubia adponitur; Ter. Ph. 339. 24 Abs quivis homine, quom est opus, beneficium accipere gaudeas; Verum enim vero id demum iuuat, si, quem aequomst facere, is bene facit; Ter. Ad. 254.

Minus placet magis quod suadetur: quod dissuadetur placet. Quom inopiast, cupias. quando eius copiast, tum non velis; Trin. 670.

CH. Vide modo ut hominem noveris. SY. Tamquam me: fieri istuc solet; Quod in manu teneas atque oculis videas, id desideres; Trin. 913.

Lippi illic oculi servos est simillumus. Si non est, nolis esse neque desideres: Si est, abstinere quin attingas non queas; Plaut. Bacch. 913.

De mortuis loquor, qui nulli sunt; nos, qui sumus, num aut cornibus caremus aut pinnis? Ecquis id dixerit? Certe nemo. Quid ita? Quia, cum id non habeas quod tibi nec usu nec natura sit aptum, non careas, etiamsi sentias te non habere; Cic. Tuse. i. 36. 87.

Unum quom noris, omnis noris; Ter. Ph. 265.

Licentiam des linguae, cum verum petas; Publil. 308. Sinmilarly [1] 10, 78, 81, 88,117, 185, 232, 233, 252, 285, 331, 361, 382, 394, 427, 468, 512, 530, 592, 635, 645; also, in perf. Subj., 126, 364.

Assem habeas, assem valeas; habes, habeberis; Petron. 77. [2]

[1] By the generally accepted readings. In a few cases, the MSS are divided.

[2] Habes is given in all of Bticheler's editions as the reading of the single MS (Parisin. 7989) which has this passage, and is reaffirmed by Btucheleri n a private letter. Omont has been good enough to report directly from the MS, to the same effect. The readings habeas and habe, respectively from the passage as repeated in Fried- lander's notes (habes in his text) and the Segebade and Lommatzsch LexciconP etro- nianum, are errors, as I am assured by Friedlitnder and Lomrnatzsch. But of course it is the habeberis that has value for the present purpose.

We proceed to test, in the light of this collection, the explana- tions that have thus far been given for this and that example taken by itself.

The construction can not be that of the Oratio Obliqua. This explanation was forced, even in the case of the first example. To many of the others, it can not be applied at all. Thus 3 can not mean "you are thought to be dead," 4 "you are thought to re- ceive at once," etc.

The Subjunctive can not be one of Command. Blase's ex- planation of velis in 25 seems to be due to a passing error of in- terpretation, and certainly can not be accepted. Even if it would there fit, however, it would not, by any stretching, fit in 1, 2, 32, and others. Lucilius, Horace, and Petronius do not mean to com- mand a man to be worth just what he possesses.

The only possible theory that could make these constructions to be of "Jussive" origin would be that they began as ironical commands, like "cry for what you don't have," and (probably) the second verb in "go further and fare worse." But Blase pre- sumably did not have this explanation in mind. And, in any case, it would be impossible, from such an origin, to account for the negative, except by supposing the construction to have lost every vestige of its initial feeling, and to have become a statement of fact. This would bring us to the actual force for which I am con- tending, though by a very improbable route.' [1]

[1] In Plut. irepZ q5\o7rXovTLas VII, p. 5260, Kgp5acVe Kat q5CeoV Kal TOO-o06Tovi v6 e cavrT?v &tLoiv, aozv& v EXes the advice is not ironical, but serious.

The Subjunctive can not be Permissive. The independent Subjunctive, if expressing the idea at all, would express Permission granted by the speaker. Such can not be the meaning in the ex- ample (23) for which Elmer gives this explanation. But, even if it were, it would be impossible to find the force in 1, 2, 3, etc. The meaning of 3, e. g., can not be "when you are dead, you are permitted to be dead in fact."

The Subjunctive can not be Potential, in the true sense. The true Potential is, to be sure, common enough in the second person singular indefinite. It indicates what lies within one's power, or within the possibilities of one's experience. Thus in Liv. xxii. 50. 9, cuneo quidemn hoc laxum atque solutunm agmen ut si nihil obstet disicias means, "by the use of the wedge-formation it lies within one's power to split this loose and disorganized line as if there were nothing in the way" (cf. w7AXla0-6AVX X7Xv. Kcat KceV &oiCTEv6etaEaP 102); and in Iuv. xiv. 41, Catilinam quocumque inpopulo videas, quocumque sub axe means "one may see (it lies within the possibilities of any man's experience to see) a Catiline in any nation, under any sky." It is also quite true that perse- qutare of 19 might by itself be interpreted, with Greenough and Daniell, as Potential (" it lies within one's power to punish other ill deeds at once") and 4, with Blase, op. cit. p. 142, in a similar way. But such a translation would be forced for frustra implores (in the same passage, 19, with per-sequare), making it mean "it lies within one's power to implore in vain," or "one might per- haps implore in vain; " while it would be completely out of the question for such examples as 1, 2, and 3, making them mean "it lies within your power to be worth as much as you possess," "it lies within your power to be dead in fact," etc.

A special word needs to be said with regard to the examples with possis. In such cases, the grammarians and commentators, as Blase op. cit., p. 142, upon our examples 8 and 9, and Munro upon 9, are too ready to give the explanation "Potential," con- fusing the meaning which the verb has in itself with the meaning of the mood employed in the particular instance. In videas of Verg. Georg. 1. 387, rightly cited by Blase in the same passage, the verb itself expresses the idea of seeing, while the mood conveys the idea of the possibility of that seeing. Possis, correspondingly, if the mood is to be really potential, would have to mean either "you have the power to have the power" or "it may possibly happen that you should have the power,"-meanings obviously out of the question here, and probably, in fact, nowhere occurring. It should be remembered, too, that by Roman idiom (apart from the usage which I am endeavoring to establish) the natural mood to use for possurm in saying " one may " would be the Indicative, potes.

The only possible way in which to deal with possis by any of the categories hitherto recognized in our grammars would be by interpreting it as meaning "one would be able." This is the Sub- junctive of Ideal Certainty, to which we pass.

The construction can not be one of IdealCertainty ("you would"). Such an explanation would, in the absence of any other, tolerably explain valeas of 16 ("you would in that case be well "), and might, with some forcing, be held to account for sev- eral of the examples with possis, and also for 17 ("it would be use- less to urge"), and for 221 ("you would think him good for some- thing"); but it would be unreasonable for 1, 2, 3, and many others. The natural meaning can not be "one would be worth what one should possess," etc. Still less reasonable would it be for 21. The speaker is actually flogging the slaves as he speaks. He is in no frame of mind to say, "if one should flog them, one would hurt one's self worse than them." And, finally, the explanation is obviously impossible for rideas, bibas, and decumbas of 23. The meaning cannot be, "his face is drawn with anxiety; you would laugh, you would drink before him, you would take your place at table before him."

The same reasoning would destroy any theory that should re- sort to the modified form of the Subjunctive of Ideal Certainty known as the "Subjunctive of Softened Assertion." This might, though not without strain, be adopted for velis of 25 and nolis of 27 (making them correspond to velim and nolim), and in conse- quence be resorted to for cupias of 25, and even for desideres of 26 and 27. But it would not fit the equivalent of desideres, namely, careas of 28; and it obviously would be impossible to find such a meaning in any other of the examples.

The construction in all the examples except 29 is obviously not an Indicative one. Dziatzko and Elmer would therefore not be through with their difficulty after explaining the ambiguous form noris as a Future Perfect.

These six explanations are all that suggest themselves as pos- sible to attempt, with the categories that we now have in the grammars. They prove to be insufficient. None of them will ac- count for all the examples. Their extreme difficulty, and the un- satisfactoriness of the whole method of procedure, become very obvious, if one sums up one's impressions by recalling that dif- ferent writers have given different explanations for the same ex- amples, that the same writer has at different times given different explanations for the same example, that certain writers have been willing to admit either of two explanations for the same example, and that certain writers (as Blase in the case of velis of 25 and valeas of 31) have put under different explanations examples obviously similar.

We must then set up a new tenet of some kind. In reaching this, three factors are to be taken into the account.

The natural interpretation of the passage in each case. It is upon such interpretations that all grammatical categories must ultimately be based. In the present set of examples the natural interpretation is sure. If our minds were grammatically dispas- sionate, a judgment founded on the reasonable demands of mean- ing in such verbs as sis in 1, habearis in 2, sis in 3, des in 30, and many others, would see in these examples plain statements of a general fact. The only reasonable interpretations are: "a man is valued according to his possessions" (1, 2); "when you are dead, dead you are" (3); "when you ask for the truth, you give a man a chance to say anything he pleases" (30); etc. And it is obviously an interpretation of this kind that Porphyrio had in mind in writing, in his comment upon iubeas, etc., of example 1, relinque eum, inquit, qui miser est, quonziam deduci ab hac opinione non potest, quo minus puttet tanti esse unumquemque, quantas divitias habet. The direct form of the opinio would evidently be, tanti est unusquisque, quantas divitias habet.

If anywhere else we can find other modes of expression obviously conveying the same general class of ideas as any of our Subjunctives under examination, we should consider their suggestions.

Now it happens that many of our examples are sententiae. One of these sententiae is the very familiar one (appearing three times in our collection) upon the relation between a man's pos- sessions and the estimation in which he is held. It can easily be matched.

Xp,qaa, XrapqLaT' LaVrp (scil. LcT7T), os fa K7TEaVWV 0dl/ca XEL40iELS Ka't At'Xw"v;'M oney, money (is) the man,' he said, 'when robbed at once of goods and friends;"' Pind. Isthmn2. . 11. OL yOZV 7rarTEpES rOVrt 7rp(LrOV rOLs 4(AETE'pOLtV LEtE0 7rapaLVOVOLVv,E 7rELoLV Ets T?V -'XLK'cV TaXffTa L3 COVTcLL TOV '7&) cfPOVELV, US 8OKOVcTL, cKO7TELV OOEV 7rX61`t0Lt EcTOVTaL, (S) aaV Et EX-7%Sa cL$ 0o TOV E tIEv 8 v6 Oi3vo' Eryx. Pseudoplat. 396 C.

Non quare et iinde, quid habeas, tantum rogant.

Ubique tanti quisque quantum habuit fuit; Sen. Ep. Mor. 115. 14. from the lost Danae of Euripides (Nauck. T. G. F; p. 324).

Pauper ubique iacet; Ov. Fast. 1. 218.

Dat census honores; Ov. Amr.I II. 8. 55 and Fast. 1. 217.

Quisquis habet nummos, secura navigat aura; Petron. 137.

Quantumq uisque sua nummorums ervat in arca,

Talntumh abet et fidei; Iuv. iii. 143.

Unde et illud proverbium: Quantum habebis, tantus eris; Augustin. De disc. Christ. 11. 12.

The second Greek example is an especially striking one, since, though the main sentence (a general one) is cast in the third person plural, the general second singular indefinite is retained in the dependent clause, and is in the Indicative. The phrase illutd proverbium in the last example should also be noted.

It would seem, then, that the natural idea to come into the mind in expressing the relation between a man's money and the estimation in which the world holds him is that of a general statement of fact.

Several of the examples refer to individual persons. Now among them are certain ones in which the ideas, down to the end of the introductory clause, look alike. Thus at quom aspicias tristem of 22 is substantially like atfaciem quomn aspicias eorum of 21. It would seem somewhat probable, then, that the con- cluding clauses would express similar ideas. These clauses are frugi censeas and hau mali videntur. They prove, as expected, to be alike in general force. It is probable, then, that the mood- force is the same, namely, that of a statement of fact.

These are the first factors in the determination of the nature of the idiom under discussion. The most natural interpretation to put upon these Subjunctives, whether in and by themselves, or in the light of the expression of the same general ideas by other means, is that they are statements of fact.

The third factor in the determination of the force of the mood lies in the study of the context of such of the examples as possess a closely attached one, of the same general bearing. This factor is of the same essential nature as the second, but is more decisive. It is, in fact, completely sufficient, by itself alone, to determine the answer to our question. It happens that in many cases the Subjunctive under examination is put into exact paral- lelism with a general statement of fact in the Indicative. Thus, in 8, possis is the amplification of satis est (not sit); in 5, timeas is balanced against neque metus . . . . est, and, in 18, pros- pera omnia cedunt against nequiquam deos implores; in 24, gautdeas is set up as the antithesis of iuvat; in 23, rideas, bibas, and decutmbas are hemmed in between the two general statements ringitur and adponitur, with the parallelism still more strongly brought out by the antithesis ille . .. . tu . . . in 19, frustra iudicia implores is justified by the general state- ment fit, and, in 20, the general statement fit by invenias; in 14, matronae . . . . nil cernere possis is set over against altera nil obstat, and videre est is carried out by metiri possis ocutlo; in 9, nulla potest oculorum ac es . . . . tueri is clearly parallel to nec cernere possis, and, in 13, ut non . . in omni agro reperire possis to sic non .... in omni vita nascitur; while in 17, not only is hortere balanced against a general statement officit, but the two together form the unfolding of the contents of the very significant Indicative solet, "it is regularly the case that." Note similarly solet . . . . desideres in 26. In 28, careas is balanced against caremus. In 7, abalienes and nihilne noceas (bis) are parallel with mediocriterne nocent, quantum est mali, and mediocre peccatum est. And, finally, in 31, the correspond- ence between valeas and habeberis is so obvious that I am almost ashamed to have printed anything of the present paper except the title and this example.

The condition of things being what has been seen, there can be no reasonable doubt that the second person singular of the Latin Subjunctive may be employed to make independent state- ments of a general fact.

Such a conclusion should afford a sensible relief. The variety of explanations adduced above as having been actually applied to one and another of these examples, the heavily forced character of some of them, even for the individual cases to which they are applied, the contradictory explanations made by different writers for the same examples, the shift of ground on the part of one writer,-all show how hard beset grammarians and commentators have been.

In the collections employed, only examples that are surely to be classified as statements of fact are given. Once established, however, the category fits many other examples more naturally than does the category of Possibility, or that of Ideal Certainty. The more important step has been taken. It remains to see whether the second is possible,-whether a probable origin can be found for the Subjunctive construction. I have long taught the following:

The extremely common use of the Subjunctive of the second singular indefinite in generalizing assumptions (relative or tem- poral clauses, or express conditions) might well, in time, affect the mood of the main sentence. To express the generalizing idea in the dependent clause, the verb in any person or number except the second singular indefinite was in the Indicative, but if it oc- curred in this form, then the mood was, by fixed force of habit, the Subjunctive. The Subjunctive would then seem to be a natural mood to use wherever the generalizing force was intended, if the second singular indefinite were employed. Just as else- where, in the expression of this idea, Indicative went with Indica- tive, so here it would seem natural that Subjunctive should go with Subjunctive. If, saying quod quis in manu tenet, one goes on with id desiderat, then, saying quod in manu teneas, it would be natural to go on with id desideres.

This influence would be all the stronger from the fact that, whatever may have been the original force of the Subjunctive in these dependent clauses [1], that force could no longer be clearly felt. The Subjunctive of the second singular indefinite had be- come merely the sign of universality. Why, then, should it re- niain confined to the subordinate clause?

[1] I believe the construction to have originated in a Volitive of the imagination, as in "lose your money, and all your friends desert you" (not, as seems everywhere as- sumed, in the Potential Subjunctive); and I have so classed it in the syntax of the Hale-Buck Grammar, § 504. 2. But a treatment of this question would properly involve a discussion of the Greek generalizing Subjunctive (which I hold to be of parallel origin), and a discussion of my difference from Delbruick with regard to an impor- tant group of Greek constructions, for which this paper affords no room.

This natural tendency toward extension would probably be fur- thered by the fact that the same second person singular indefinite happens to be in frequent use in a number of true independent Subjunctive idioms. Thus it is found in the Subjunctive Question of Perplexity, as in quid agas? "what is one to do?" Iuv. iv. 14; in the Question of Obligation, Propriety, or Reasonableness (to avail myself of a classification and name from own teaching) [2], as in cur tamen corripias? "Why should one chide her?" Iuv. x. 291; and in the form of statement which lies in the neutral terri- tory of Natural Likelihood, Possibility, and Ideal Certainty, as in crederes victos, "one might well think them conquered," "one might think them conquered," "one would think them conquered," Liv. ii. 43. 9.

[2] The Subjunctive of Obligation, Propriety, or Reasonableness, first set up as a general category in my teaching at Cornell, is made one of the principal families of the mood in my general scheme, and several of the hitherto difficult types of dependent clauses seem to find full solution in it (Hale-Buck Grammar, ?§ 512,5 13; also Classical Journal I. 1. 1905, "An Experiment in the Teaching of First and Second Year Latin," where a conspectus is given of the uses of the Latin Subjunctive, classified according to origins). Not only is the mood-use a frequent one, but,,in my opinion, it furnishes the missing link between the true Optative and the Optative (Latin Subjunctive) of Natural Likelihood, Possibility, or Ideal Certainty. (See my "Leading Mood-Forces in the Indo-European Parent Speech," Proceedings of the American Philological Association XXXII, p. cxx [1901].)

The combination of these influences with the one already mentioned might well have led the Romans to feel that, whatever mood-feeling they had to convey, the Subjunctive was the natural mood to use, if the person was the second singular indefinite. But, most of all (after the primary cause), the natural tendency to extend the idiom would be strengthened by the very frequent occurrence of a construction which, though evidently of true Potential origin, clearly suggests and implies fact. Thus Juvenal's Potential in Catilinam quocumque in poputlo videas, quocumque sub axe," "one may see a Catiline in any nation, under any sky," xiv. 41, easily suggests "one sees . . ." That which is said to be anywhere and everywhere possible is easily understood as anywhere and everywhere taking place. The number of ex- amples of this type (videas, cernas, etc.) is very large. An especi- ally instructive set is afforded by reperias. I give three of these in sequence:

Quamvis malam rem quaeras, illic reperias; Plaut. Trin. 554.

Per pol quam paucos reperias meretricibus Fidelis evenire amatores, Syra; Ter. Hec. 58.

Ut enim sunt, quem ad modum supra dixi, qui urbanis rebus bellicas ante- ponant, sic reperias multos, quibus periculosae t calidac onsilia quietis et cogitatis splendidiora et maiora videantur; Cic. Off. i. 24. 82.

In the first of these, "one may find it there," may suggest "one finds it there." In the second, "how few one can find," strongly suggests "how few one |finds;" while in the third, reperias qui- bus . . . . is put into formal parallelism with the general state- ment sunt qui . . . . , "there are people who . . . ." Reperias accordingly is completely ready, at any rate, to convey the idea of fact.

So much for the construction under examination. An inde- pendent use of the Subjunctive to express statements-assertions of fact-in the second person singular indefinite, exists; and it probably is historically an extension from other constructions in which the Subjunctive had, originally, true-mood forces.[1]

[1] Hence in my grammatical scheme, referred to above, this use is placed under the distinct heading of "Subjunctive Constructions Due to the Influence of Other Constructions."

But did the construction become universal, where the idea of a general statement of fact was to be conveyed, and the second singular indefinite was to be used? It seems not. The older usage, which of course must have been that of the Indicative mood, still survived. In the future, the idea had necessarily to be expressed by the Indicative if expressed at all, since the in- dependent Anticipatory Subjunctive had disappeared, wholly or almost wholly, from the language. But there are also examples of the present.

My paper deals with such Subjunctives as the independent ones in the familiar passages: A. Nil satis est, inquit, quia tanti quantum habeas sis; Hor. Sat. i. 1. 62. B. Assem habeas, assem valeas; Petron. 77. C. Ubi mortuos sis, ita sis ut nomen cluet; Plaut. Trin. 496. D. Quom inopiast, cupias: quando eius copiast, tum non velis; Plaut. Trin. 671. E. Quod in manu teneas atque oculis videas, id desideres; Plaut. Trin. 914. F. Unum quom noris, omnis noris; Ter. Ph. 265. G. Accipe nunc, victus tenuis quae quantaque secum Adferat. In primis valeas bene; Hor. Sat. ii. 2. 70. H. Dum tibi fit quod placeat, ille ringitur: tu rideas, Prior bibas, prior decumnbas; Ter. Ph. 341. I. Nam cetera maleficia tum persequare, ubi facta sunt, hoc, nisi provi- deris ne accidat, ubi evenlit, frustra iudicia implores; Sall. Cat. 52.4. J. Quem neque gloria neque pericula excitant, nequiquam hortere; timor animi auribus officit; Sall. Cat. 58. 2. K. Nec porro quaecumque aevo macieque senescunt, Nec, mare quae inpendent, vesco sale saxa peresa Quid quoque amittant in tempore cernere possis; Lucr. i. 325. L. Quae si ipse exsequi nequeas, possis tamen Scipioni praecipere et Laelio ("a Scipio and a Laelius "); Cic. Sen. 9. 28.

These look alike. If there were a comfortable category in the grammars into which to put them all, one would not think of separating them. As the title of the paper implies, I am brought to regard such Subjunctives as expressing general statements of fact,-as corre- sponding, in effect, to Indicatives in the same second person singu- lar indefinite. I wish to translate example A by, "because you are worth just what you possess ;" B by, "have a penny, you are worth a penny ;" C by, "when you are dead, dead you are," etc. In accordance with this view, § 542 of the Hale-Buck Grammar reads: "A general statement of fact is sometimes expressed by a Subjunctive of the Second Person Singular Indefinite;" and to our example C, there given, is added the statement, "the second sis has the force of es."

The construction itself will prove to be a simple one to solve, but the exhibition of previous opinion upon the matter must be a somewhat tangled affair, since grammars, grammatical treatises, notes in commentaries written by professional grammarians, and notes by commentators who are not professional grammarians, have all to be taken into the account. It is also obviously impracti- cable to refer to every note ever written upon any of these pas- sages. The selections, however, will fairly cover the ground.

So far as I know, no grammar or grammatical paper had rec- ognized the existence of the construction for which I am contend- ing. The statements regularly made (where any are made) with regard to the range of the Subjunctive of the second person singu- lar indefinite are of a kind to exclude such an interpretation. Thus Madvig, § 370, says: "Diese Form findet sich in bedingter Rede, in hypothetischenAussagen und in Fragen uber das, was geschehen wird und kann, in Nebensatzen mit Conjunctionen, in Relativsat- zen (mit qui oder einem unbestimmten Relativ), und in Vor- schriften und Verboten." To this is added the Anmerkung: "Ein solcher conjunctivischer Bedingungssatz fuhrt nicht den Con- junctiv im Hauptsatze herbei." Quite recently (1905), Nutting, Studies in the Si-Clause, pp. 84, 85, has distinctly shown the conviction that there is no Subjunctive in a main sentence corre- sponding to the common second singular indefinite in subordinate clauses. "One looks in vain," he says, "for a clear case of such leveling in Plautus." And Elmer, in his note on F (as will be seen below), in effect says that a general statement of fact in the second singular indefinite must be expressed by the Indicative.

The grammarians of course all provide for the Subjunctive second person singular indefinite in conditional or corresponding relative clauses which, if in any other person or number, would be in the Indicative, as in memoria minuitur, nisi eam exerceas, Cic. Sen. 7. 21; but they know nothing of the power for which this paper contends. If anywhere they chance to make use of an ex- ample of the type before us, it is placed under some already recog- nized and entirely different category, as we shall see below.

Whatever the grammarians may do, the commentators, it would seem, would frequently have occasion to discuss such examples. The most striking feature of the case is, however, that they gen- erally pass them without comment. Sometimes, indeed, they actually give an explanation of the dependent verb in such com- binations, and pass in silence the far more difficult independent one. Thus Greenough, in his edition of the De senectute (1873 and in various reprints), says of our example L, "si . . . . nequteas, gen. cond. (§ 309 a)," but says nothing of possis. Occasionally the dependent verb and the main verb are dealt with together and without distinction, with a reference to some rule, such as the grammars afford. This reference, as is natural under the circum- stances, never fits. Thus Gudeman, dealing with example I in his edition of Sallust, says, " persequare . . . . provideris inplores: Subjunctive of indefinite second person singular," and refers to the Allen and Greenough Grammar, 266a. This section, when one turns to it, proves to deal with independent "EIortatory" Subjunctives, and to read (266), " The Hortatory Subjunctive is used to express an exhortation, a command, a con- cession, or a condition;" and (a), "The Second Person is used only of an indefinite subject, except in prohibition, in early Latin, and in poetry." Not only the general headings "independent" and "'h ortatory," but the examples given (e. g., isto bono utare dum adsit, Cic. Sen. 10. 33), show the type meant. But nisi provideris is evidently not independent; and it is also difficult to believe that Sallust means to direct people to implore in vain, unless, indeed, the editor conceives the idiom to be one of ironical command (not provided for in the reference), in which case the explanation should certainly be given. More frequently, if any- thing at all is said about the main Subjunctive, it is simply classed with the dependent one, and the Subjunctive idiom in the second singular indefinite in " conditional " clauses is referred to or assumed. Thus Munro, in his comment upon our example K, puts possis with si non relinquas, Lucr. i. 515, quibus careas, ii. 4, etc. Thus, again, Greenough, in the note in his edition of the Satires and Epistles of Horace, 1888, says of our example A, "tanti sis (you are r ated att), etc. The subjunctive is the regular one of the second person with indefinite subject." But in his Gram- mar Greenough (the same is true of his revisers) makes no pro- vision for an independent Subjunctive of the second person singular indefinite, except as hortatory or potential; while the treatment of the dependent member of the example in question comes in § 309 a (518 of the new edition) under the head of "General Conditions," to which it is obvious that sis does not belong. It is clear, also, that general conditions are really meant; for the rule reads: "The Subjunctive is sometimes used in the second person singular to denote the act of an indefinite Subject (you =any one). Here the Indicative of a general truth may stand in the apodosis."'

Yet the tendency naturally produced by the sight of the com- bination of dependent and independent Subjunctives in the same indefinite second person is a good one, and has in a number of instances led, even in the absence of an admitted grammatical category, to the solid interpretation. Thus Wagner, in the note in his Trinummus (1875) upon our example E, says, " id desid- eres, 'that one is apt to forget': the second person subjunctive expresses generality." Greenough (1888), as we have seen, translates tanti sis of A by " you are rated." Friedlander, in his Petronii Centa Trimalchionis (1891) translates our example B by 4"G laubt mir, habe einen As, so giltst du einen As." Morris, in his edition of the Captivi and Trinummus of Plautus (1898), paraphrasing example C, says, " when you are dead, you are dead in the full sense of the term." He makes no grammatical com- ment. Rolfe, in his edition of the Satires and Epistles of Horace

[1] The word " sometimes " is not happy. The Subjunctive is regularly used, there being but a very small number of exceptions in the whole range of the literature. Hardly any usage in Latin is more fixed. Or, if the intended meaning is " one some- times finds examples of . " then, by the same token, the word " sometimes" ought to be added to every rule in the Grammar. Moreover, the phrase " may stand" in the last line should read " generally stands."

(1902) puts the main and subordinate clauses in example A to- gether, rightly translating, " because you are rated by the amount of your possessions, " and adding, "habeas and sis are subjunctive because of the indefinite second person singular." No grammati- cal reference is given. The procedure of Wagner, Greenough, Friedlander, Morris, and Rolfe seems to imply a subconsciousness of the force naturally to be attributed to the independent con- struction; but these excellent opportunities which Greenough and Morris, as grammarians, had for calling attention to a much mis- understood, and very variously interpreted, construction, and of enunciating its true character, pass unused. Note also, under 4 below, p. 27, a discordant explanation of the exactly similar per- sequtare of example I (namely as Potential) in the Greenough and Daniell Sallust.

The only comment I have noticed which makes a sound ex- plicit statement about the independent construction is that of Sch-itz (1881), who, in his note upon our passage A, says "sis, nicht es, weil allgemein, nicht an H. direct gerichtet; 'man gilt."' To this he adds, "Im Indicativ wulrde man eher die 1. Person verlaingen, weil der Geizige doch sich selbst damit entschuldigen will." He is right in his general feeling, but not right in imply- ing that in the second singular indefinite the Indicative could not be used. See below, at the end of this paper.

The explanations which clearly attempt to bring one or arnother of the independent Subjunctives in sentences A-L under one or another of the categories provided for in the grammars are of six kinds. I shall first merely state them, without argument.

1. The Subjunctive is that of the Oratio Obliqua. This is the favorite older explanation, and is still frequently found. Thus Wusteman, revision of Heindorf (1843), says of our example A, "man bemerke den im Latein. gew6hnlichen Uber- gang von der orat. recta in die obliqua;" H. Duntzer (1869) says "sis, Ubergang in die abhangige Rede;" Kirchner (1885) says "im Lateinischen geht die Rede aus der directen Construc- tion (nil satis est) in die abhangige (quia tanti . . . . sis) uiber, was auch in Prosa nicht selten vorkommt;" and G. T. A. Kruger in various editions (as in 1860 and 1866), by a still bolder device, says "sis, statt es, als ware vorhergegangen: nil satis esse ait. Also auf einer Anakoluthie beruhend." G. Krufger says the same, in re-editing his father's edition in 1876. Among the later editors, Kiessling (1886) says the same in the sentence, "der Konjunktiv sis = esse putaris." Palmer (my edition is of 1888) seems to mean the same by his phrase "general opinion" in saying, "the subjunctive sis is used because it is the general opinion, indefinite." And Harper's Lexicon (1879) means the same, in putting example A under the Subjunctive illustration of the statement, "construed with indicative in asserting a fact; with subjunctive in stating an assumed reason, or one entertained by another mind."

2. The Subjunctive is that of Command or Exhortation ("Imperative"). Thus Blase, in his recently published "Tem- pora u. Modi" in the Historische Grammatik d. lat. Sprache, p. 136, puts our D as his first example under the head of the Jussive, Hortative, or (true) Optative, and accounts for non as modifying, not the sentence as a whole, but the single word velis. The same explanation seems to be meant, p. 123, for C, which is translated by "so bald man tot ist, soll man es ganz sein." And the same is explicitly suggested by Antoine and Lallier as possible for I. See below, at the end of 4.

3. The Subjunctive is Permissive. Thus Elmer, on example H, says, "'rideas, bibas, etc.: subj. of permission, involving the yielding of somebody's will; to be carefully distinguished from the potential."

4. The Subjunctive is Potential. It is by the Potential formula ("you can laugh, drink your wine before him," etc.) that Morris H. Morgan, in his translation of the Phormio (1904), renders these same Subjunctives which Elmer tells us are "to be carefully distinguished from the potential." The express word "Poten- tial" is used of one or another of our examples by many commen- tators. Thus Dillenburger (my edition is of 1867) both affirms this explanation and denies 1 above, in the note: "sis non per anacoluthiam positum est pro es, quasi praecederet nil satis esse ait, qutia, sed est potentialis modus, ut apud Lucilium." Likewise G. Krttger, who in the edition of 1876 explained sis in A as due to Oratio Obliqua, explains it in the edition of 1889 as Potential. This same explanation is given by Lucian Miller (1891) for A and B, by Blase op. cit., p. 142, for B, by SchtVtz op. cit. for G, and by Schmalz Bell. Cat. (my edition is of 1885) for J.

Unfortunately this word "Potential," as used in our gram- mars, is ambiguous. It includes meanings so far apart, at their extremes, as "may possibly" or "can," and "would certainly." The former idea alone should bear the name Potential, while the other (seen most easily in the common Subjunctive conclusion) should bear some name like that of Ideal Certainty, which I have given it in my own grammatical writing. [1] It is the true Potential use, apparently, that Kruger and Miller have in mind, and likewise Freeman and Sloman upon our example D. This, at any rate, is what is meant by Greenough and Daniell in their Sallust, in the note upon persequare in our example I. The translation given is "one may prosecute," and the reference (Allen and Greenough Grammnar 311a) is to the statement, "The Potential Subjunctive is used to denote an action not as actually performed, but as possible." Similarly Antoine and Lallier, Salluste, Co7ijuration de Catilina (1888), write in their commentary: "persequare pourrait etre un subjonctif imp6ratif, puisque le sujet est indetermin6; mais en réalité, comme le note M. P. Thomas, c'est le subjonctif potentiel: 'on peut ne les pour- suivre que lorsqu'ils sont accomplis.'"

[1] 'Delbruiek Verg. Syntax II, p. 371, and Brugmann Kurze vergl. Gramm. II, p. 584, while accepting my term, have misunderstood my meaning.

5. The Subjunctive is that of Ideal Certainty. This is doubt- less what Roby means in placing our example J under the "hypo- thetical Subjunctive," § 1544 (cf. § 1528 and 3). The corre- sponding English for our examples would then be "one would." This is also substantially what Kiesslin0 means in saying, upon example G, "valeas, Konj. der gemilderten Behauptung, ὑγιαίνοις ἄν," L. Muller in saying, upon the same example, "valeas; wie oft in Satiren und Episteln, urbane Milderung der Behauptung," Heinze, new edition (1906) of Kiessling's Scttires of Horace, in saying, "valeas scil. si parvo vivas," Rolfe op. cit., in slaying, "v aleas: potential subjunctive with an implied protasis, si tenui vietu utaris," Palmer op. cit., in saying, "valeas, a potential or modified future," Wickham Satires, Epistles, and De arte poetica of Horace, 1891, in saying, "valeas, potential," and Bennett Cato itior, 1897, in his note nipon L, "if you should be unable to practice this." This must likewise be what Wickham had in mind in saying of A, "the mood is probably the same as in the original, sis5-' aVlv S;7" though I fancy that this is a restored original, not a known one. And this is also probably what Heinze means, op. cit., in saying of the same example, "der potentiale Konjunktiv sis wie bei Petron. 77 asserm hctbeas assem valeas." (It will be noticed that Heiutze has abandoned Kiessling's explanation.)

6. The independent verb of example F, unum quom noris, omnis nor is, is explained by at least two of the commentators as not a Subjunctive at all, but an Indicative of the future perfect. Thus Dziatzko says, "noris ist beidemal Fut. exact;" and Elmer gives the same view, with the reasoning which brings him to it, as follows; "noris: in both cases fut. perf. ind. equivalent to fut. If the first noris were the perf. subj. of an indefinite second person, we should have omnis novisti, instead of omnis noris."

The error of procedure in all these explanations is of a kind that shows itself again and againi in the history of grammatical study. Those who have dealt with one or another example of a type have, in repeated instances, explained the example which they had before them, without stopping to make a collectiont of examples sufficient in number to enable a scientific inference of any kind to be drawn. The perfect method is, first, to make a complete collection, and theni to study every example, in the light of all. Human life being, however, what it is, one must, until our work is better organized, be content with a fairly large col- lection. In making this, one must throw aside grammatical pre- possessions, and admit all examples that would naturally seem to belong together. Further, it would be worth while to look, at least, at any sentences that might be found, in which ideas of the same general kind seemed to be expressed, but by a different mechanism. Further, and without any doubt, as much of the context of the Stibjunctive examples as could possibly be illumi- nating should be added for each. Here again our prevailing method is sadly defective. In a number of examples of the type which we are studying, the context immediately and irresistibly overthrows the explanations given. The material being thus gath- ered, every explanation that might account for one or more of the mass of examples, but would not account for all, or that in any example would fail to fit the context, should be rejected. The result of this process in the present instance will be the overthrow of all six of the explanations given for the examples cited above.

Assuming this for the moment to be the case, the next step should be to seek for an explanation that will explain all the examples, and that will be in harmony with the context in all. If one is thereby brought to believe in the existence of a con- struction not recognized by the grammars, one need not be alarmed, if one has confidence in the dispassionateness of the pro- cedure that has brought him to his conclusion.

Finally, one may properly seek for a probable origin of the usage thus reached, out of some construction or constructions known actually to exist. The two steps, however,-the establish- ment of the existence of a construction, and the search for its origin,-may be entirely disconnected. Failure to take the second step successfully will not of itself invalidate the first. A number of constructions are surely known to exist of which the origins are not surely known. In the present case, however, both steps can, I think, be securely taken.

In the collection which follows, the examples are arranged by groups of ideas, or for convenience in reference:

1. At bona pars hominum,d ecepta cupidine falso 'Nil satis est,' inquit, 'quia tanti quantum habeas sis.' Quid facias illi? Iubeas miserum esse libenter Quatenus id facit. Hor. Sat. i. 1. 61.

2. Aurum adque ambitio specimen virtutis viriquest. Quantum habeas, tantum ipse sies tantique habearis; Lucil. Schol. ad Iuv. iii. 143.

3. Mirum quin tu illo tecum divitias feras. Ubi mortuos sis, ita sis ut nomen cluet; Plaut. Trin. 495.

4. Tarda sunt quae in commune expostulantur: privatam gratiam statim mereare,s tatim recipias; Tac. Ann. i. 28. 21.

5. Nam si velis quod nondum vetitum est, timeas ne vetere: at si pro- hibita impune transcenderis,n eque metus ultra neque pudor est; Tac. Ann. iii. 54. 10.

6. Quid in amicitia fieri oportet, quae tota veritate perpenditur? In qua nisi, ut dicitur, apertum pectus videas tuumque ostendas, nihil exploratum habeas; Cic. Am. 26. 97.

7. Mediocriternec ausis nocent, cum . . . . adversarioruma diumenta confirmant? . . . . Si, quae sunt in iis invidiosa, . . . . invidi- osiora faciunt, quantum est in eo tandem mali? Quid, si .... contumeliosius invehare, nonne a te iudices abalienes? Quid, si . ... non intellegas te in iudices invehi, mediocre peccatum est? Quid, si . . . . causam relinquas, nihilne noceas 2. In ipsis autem argumentis si quid posueris . . . . aperte falsum .. . .nihilne noceas? Cic. De or. ii. 75, 303-306.

8. Liber captivos avis ferae consimilis est: Semel fugiendi si datast occasio, satis est, Numquam postilla possis prendere; Plaut. Capt. 116.

9. Nulla potest oculorum acies contenta tueri, Nec porro quae cumque aevo macieque senescunt, Nec, mare quae inpendent, vesco sale saxa peresa Quid quoque amittant in tempore cernere possis; Lucr. i. 324. Similarly ii. 220, 763, 768; iii. 856, 1024; iv. 572, 1231; vi. 113. Cf. the formulae vtidere licet, vi. 79; licet cognoscere, vi. 167, etc.

10. Hostes alienigenae aut oppressi serviunt aut recepti in amicitiam beneficio se obligatos putant: qui autem ex numero civium, dementia aliqua depravati, hostes patriae semel esse coeperunt, nec vi coerceren ec beneficio placare possis; Cic. Cat. iv. 10. 22.

11. Sed tamen est decorus senis sermo quietus et remissus, facitque persaepe ipsa sibi audientiam diserti senis composita et mitis oratio. Quam si ipse exsequi nequeas, possis tamen Scipioni praeciperee t Laelio; Cic. Sen. 9. 28.

12. Omnino si quicquam est decorum, nihil est profecto magis quam aequabilitasc um universae vitae, tum singularum actionum, quam conservare non possis (really an independent sentence) si aliorum naturam imitans omittas tuam; Cic. Off. i. 31. 111.

13. Ut non omnem frugem neque arboremi n omni agro reperire possis, sic non omne facinus in omni vita nascitur; Cic. Sex. Rosc. 27. 75.

14. Matronaep raeter faciem nil cernerep ossis, Cetera, ni Catia est, demissa veste tegentis; * * * * * * * * Altera nil obstat: Cois tibi paene videre est, Ut nudam, ne crure malo, ne sit pede turpi; Metiri possis oculo latus. An tibi mavis Insidias fieri pretiumque avellier ante Quam mercem ostendi? Hor. Sat. i. 2. 94, 95, and 101-105.

15. Vix credere possis Quam sibi non sit amicus; Hor. Sat. i. 2. 19.

16. Accipe nunc, victus tenuis quae quantaque secum Adferat. In primis valeas bene. Hor. Sat. ii. 2. 70.

17. Quanta cuiusque animo audacia natura aut moribus inest, tanta in bello patere solet. Quem neque gloria neque pericula excitant, nequiquam hortere; timor animi auribus officit; Sall. Cat. 58. 2.

18. Non votis neque suppliciis muliebribus auxilia deorum parantur: vigilando, agundo, bene consulundo prospera omnia cedunt: ubi socordiae te atque ignaviae tradideris, nequiquam deos implores: irati infestique sunt; Sall. Cat. 52. 29.

19. Nam cetera maleficia tum persequare, ubi facta sunt; hoc, nisi provideris ne accidat, ubi evenit, frustra iudicia implores: capta urbe nihil fit reliqui victis; Sall. Cat. 52. 4.

20. Siquoi mutuom quid dederis, fit pro propriop erditum: Quom repetas, inimicum amicum invenias beneficio tuo; Plaut. Trin. 1051.

21. Neque ego homines magis asinos numquam vidi: ita plagis costae callent: Quos quom ferias, tibi plus noceas. Eo enim ingenio hi sunt fla- gritribae, * * * * * * * * At faciem quom aspicias eorum,h au mali videntur; Plaut. Pseud. 136.

22. Sed eccum incedit: at quom aspicias tristem, frugi censeas; Plaut. Cas. 562. Similarly cum . . . . perventum est, tum intellegas; Cic. Div. ii. 21. 48.

23. Tene asymbolum venire unctum atque lautum e balineis, Otiosum ab animo, quom ille et cura et sumptu absumitur! Dum tibi fit quod placeat, ille ringitur: tu rideas, Prior bibas, prior decumbas; cena dubia adponitur; Ter. Ph. 339. 24 Abs quivis homine, quom est opus, beneficium accipere gaudeas; Verum enim vero id demum iuuat, si, quem aequomst facere, is bene facit; Ter. Ad. 254.

25. Minus placet magis quod suadetur: quod dissuadetur placet. Quom inopiast, cupias. quando eius copiast, tum non velis; Trin. 670.

26. CH. Vide modo ut hominem noveris. SY. Tamquam me: fieri istuc solet; Quod in manu teneas atque oculis videas, id desideres; Trin. 913.

27. Lippi illic oculi servos est simillumus. Si non est, nolis esse neque desideres: Si est, abstinere quin attingas non queas; Plaut. Bacch. 913.

28. De mortuis loquor, qui nulli sunt; nos, qui sumus, num aut cornibus caremus aut pinnis? Ecquis id dixerit? Certe nemo. Quid ita? Quia, cum id non habeas quod tibi nec usu nec natura sit aptum, non careas, etiamsi sentias te non habere; Cic. Tuse. i. 36. 87.

Unum quom noris, omnis noris; Ter. Ph. 265.

30. Licentiam des linguae, cum verum petas; Publil. 308. Sinmilarly [1] 10, 78, 81, 88,117, 185, 232, 233, 252, 285, 331, 361, 382, 394, 427, 468, 512, 530, 592, 635, 645; also, in perf. Subj., 126, 364.

Assem habeas, assem valeas; habes, habeberis; Petron. 77. [2]

[1] By the generally accepted readings. In a few cases, the MSS are divided.

[2] Habes is given in all of Bticheler's editions as the reading of the single MS (Parisin. 7989) which has this passage, and is reaffirmed by Btucheleri n a private letter. Omont has been good enough to report directly from the MS, to the same effect. The readings habeas and habe, respectively from the passage as repeated in Fried- lander's notes (habes in his text) and the Segebade and Lommatzsch LexciconP etro- nianum, are errors, as I am assured by Friedlitnder and Lomrnatzsch. But of course it is the habeberis that has value for the present purpose.

We proceed to test, in the light of this collection, the explana- tions that have thus far been given for this and that example taken by itself.

1. The construction can not be that of the Oratio Obliqua. This explanation was forced, even in the case of the first example. To many of the others, it can not be applied at all. Thus 3 can not mean "you are thought to be dead," 4 "you are thought to re- ceive at once," etc.

2. The Subjunctive can not be one of Command. Blase's ex- planation of velis in 25 seems to be due to a passing error of in- terpretation, and certainly can not be accepted. Even if it would there fit, however, it would not, by any stretching, fit in 1, 2, 32, and others. Lucilius, Horace, and Petronius do not mean to com- mand a man to be worth just what he possesses.

The only possible theory that could make these constructions to be of "Jussive" origin would be that they began as ironical commands, like "cry for what you don't have," and (probably) the second verb in "go further and fare worse." But Blase pre- sumably did not have this explanation in mind. And, in any case, it would be impossible, from such an origin, to account for the negative, except by supposing the construction to have lost every vestige of its initial feeling, and to have become a statement of fact. This would bring us to the actual force for which I am con- tending, though by a very improbable route.' [1]

[1] In Plut. irepZ q5\o7rXovTLas VII, p. 5260, Kgp5acVe Kat q5CeoV Kal TOO-o06Tovi v6 e cavrT?v &tLoiv, aozv& v EXes the advice is not ironical, but serious.

3. The Subjunctive can not be Permissive. The independent Subjunctive, if expressing the idea at all, would express Permission granted by the speaker. Such can not be the meaning in the ex- ample (23) for which Elmer gives this explanation. But, even if it were, it would be impossible to find the force in 1, 2, 3, etc. The meaning of 3, e. g., can not be "when you are dead, you are permitted to be dead in fact."

4. The Subjunctive can not be Potential, in the true sense. The true Potential is, to be sure, common enough in the second person singular indefinite. It indicates what lies within one's power, or within the possibilities of one's experience. Thus in Liv. xxii. 50. 9, cuneo quidemn hoc laxum atque solutunm agmen ut si nihil obstet disicias means, "by the use of the wedge-formation it lies within one's power to split this loose and disorganized line as if there were nothing in the way" (cf. w7AXla0-6AVX X7Xv. Kcat KceV &oiCTEv6etaEaP 102); and in Iuv. xiv. 41, Catilinam quocumque inpopulo videas, quocumque sub axe means "one may see (it lies within the possibilities of any man's experience to see) a Catiline in any nation, under any sky." It is also quite true that perse- qutare of 19 might by itself be interpreted, with Greenough and Daniell, as Potential (" it lies within one's power to punish other ill deeds at once") and 4, with Blase, op. cit. p. 142, in a similar way. But such a translation would be forced for frustra implores (in the same passage, 19, with per-sequare), making it mean "it lies within one's power to implore in vain," or "one might per- haps implore in vain; " while it would be completely out of the question for such examples as 1, 2, and 3, making them mean "it lies within your power to be worth as much as you possess," "it lies within your power to be dead in fact," etc.

A special word needs to be said with regard to the examples with possis. In such cases, the grammarians and commentators, as Blase op. cit., p. 142, upon our examples 8 and 9, and Munro upon 9, are too ready to give the explanation "Potential," con- fusing the meaning which the verb has in itself with the meaning of the mood employed in the particular instance. In videas of Verg. Georg. 1. 387, rightly cited by Blase in the same passage, the verb itself expresses the idea of seeing, while the mood conveys the idea of the possibility of that seeing. Possis, correspondingly, if the mood is to be really potential, would have to mean either "you have the power to have the power" or "it may possibly happen that you should have the power,"-meanings obviously out of the question here, and probably, in fact, nowhere occurring. It should be remembered, too, that by Roman idiom (apart from the usage which I am endeavoring to establish) the natural mood to use for possurm in saying " one may " would be the Indicative, potes.

The only possible way in which to deal with possis by any of the categories hitherto recognized in our grammars would be by interpreting it as meaning "one would be able." This is the Sub- junctive of Ideal Certainty, to which we pass.

5. The construction can not be one of IdealCertainty ("you would"). Such an explanation would, in the absence of any other, tolerably explain valeas of 16 ("you would in that case be well "), and might, with some forcing, be held to account for sev- eral of the examples with possis, and also for 17 ("it would be use- less to urge"), and for 221 ("you would think him good for some- thing"); but it would be unreasonable for 1, 2, 3, and many others. The natural meaning can not be "one would be worth what one should possess," etc. Still less reasonable would it be for 21. The speaker is actually flogging the slaves as he speaks. He is in no frame of mind to say, "if one should flog them, one would hurt one's self worse than them." And, finally, the explanation is obviously impossible for rideas, bibas, and decumbas of 23. The meaning cannot be, "his face is drawn with anxiety; you would laugh, you would drink before him, you would take your place at table before him."

The same reasoning would destroy any theory that should re- sort to the modified form of the Subjunctive of Ideal Certainty known as the "Subjunctive of Softened Assertion." This might, though not without strain, be adopted for velis of 25 and nolis of 27 (making them correspond to velim and nolim), and in conse- quence be resorted to for cupias of 25, and even for desideres of 26 and 27. But it would not fit the equivalent of desideres, namely, careas of 28; and it obviously would be impossible to find such a meaning in any other of the examples.

6. The construction in all the examples except 29 is obviously not an Indicative one. Dziatzko and Elmer would therefore not be through with their difficulty after explaining the ambiguous form noris as a Future Perfect.

These six explanations are all that suggest themselves as pos- sible to attempt, with the categories that we now have in the grammars. They prove to be insufficient. None of them will ac- count for all the examples. Their extreme difficulty, and the un- satisfactoriness of the whole method of procedure, become very obvious, if one sums up one's impressions by recalling that dif- ferent writers have given different explanations for the same ex- amples, that the same writer has at different times given different explanations for the same example, that certain writers have been willing to admit either of two explanations for the same example, and that certain writers (as Blase in the case of velis of 25 and valeas of 31) have put under different explanations examples obviously similar.

We must then set up a new tenet of some kind. In reaching this, three factors are to be taken into the account.

1. The natural interpretation of the passage in each case. It is upon such interpretations that all grammatical categories must ultimately be based. In the present set of examples the natural interpretation is sure. If our minds were grammatically dispas- sionate, a judgment founded on the reasonable demands of mean- ing in such verbs as sis in 1, habearis in 2, sis in 3, des in 30, and many others, would see in these examples plain statements of a general fact. The only reasonable interpretations are: "a man is valued according to his possessions" (1, 2); "when you are dead, dead you are" (3); "when you ask for the truth, you give a man a chance to say anything he pleases" (30); etc. And it is obviously an interpretation of this kind that Porphyrio had in mind in writing, in his comment upon iubeas, etc., of example 1, relinque eum, inquit, qui miser est, quonziam deduci ab hac opinione non potest, quo minus puttet tanti esse unumquemque, quantas divitias habet. The direct form of the opinio would evidently be, tanti est unusquisque, quantas divitias habet.

2. If anywhere else we can find other modes of expression obviously conveying the same general class of ideas as any of our Subjunctives under examination, we should consider their suggestions.

Now it happens that many of our examples are sententiae. One of these sententiae is the very familiar one (appearing three times in our collection) upon the relation between a man's pos- sessions and the estimation in which he is held. It can easily be matched.

Xp,qaa, XrapqLaT' LaVrp (scil. LcT7T), os fa K7TEaVWV 0dl/ca XEL40iELS Ka't At'Xw"v;'M oney, money (is) the man,' he said, 'when robbed at once of goods and friends;"' Pind. Isthmn2. . 11. OL yOZV 7rarTEpES rOVrt 7rp(LrOV rOLs 4(AETE'pOLtV LEtE0 7rapaLVOVOLVv,E 7rELoLV Ets T?V -'XLK'cV TaXffTa L3 COVTcLL TOV '7&) cfPOVELV, US 8OKOVcTL, cKO7TELV OOEV 7rX61`t0Lt EcTOVTaL, (S) aaV Et EX-7%Sa cL$ 0o TOV E tIEv 8 v6 Oi3vo' Eryx. Pseudoplat. 396 C.

Non quare et iinde, quid habeas, tantum rogant.

Ubique tanti quisque quantum habuit fuit; Sen. Ep. Mor. 115. 14. from the lost Danae of Euripides (Nauck. T. G. F; p. 324).

Pauper ubique iacet; Ov. Fast. 1. 218.

Dat census honores; Ov. Amr.I II. 8. 55 and Fast. 1. 217.

Quisquis habet nummos, secura navigat aura; Petron. 137.

Quantumq uisque sua nummorums ervat in arca,

Talntumh abet et fidei; Iuv. iii. 143.

Unde et illud proverbium: Quantum habebis, tantus eris; Augustin. De disc. Christ. 11. 12.

The second Greek example is an especially striking one, since, though the main sentence (a general one) is cast in the third person plural, the general second singular indefinite is retained in the dependent clause, and is in the Indicative. The phrase illutd proverbium in the last example should also be noted.

It would seem, then, that the natural idea to come into the mind in expressing the relation between a man's money and the estimation in which the world holds him is that of a general statement of fact.

Several of the examples refer to individual persons. Now among them are certain ones in which the ideas, down to the end of the introductory clause, look alike. Thus at quom aspicias tristem of 22 is substantially like atfaciem quomn aspicias eorum of 21. It would seem somewhat probable, then, that the con- cluding clauses would express similar ideas. These clauses are frugi censeas and hau mali videntur. They prove, as expected, to be alike in general force. It is probable, then, that the mood- force is the same, namely, that of a statement of fact.

These are the first factors in the determination of the nature of the idiom under discussion. The most natural interpretation to put upon these Subjunctives, whether in and by themselves, or in the light of the expression of the same general ideas by other means, is that they are statements of fact.

3. The third factor in the determination of the force of the mood lies in the study of the context of such of the examples as possess a closely attached one, of the same general bearing. This factor is of the same essential nature as the second, but is more decisive. It is, in fact, completely sufficient, by itself alone, to determine the answer to our question. It happens that in many cases the Subjunctive under examination is put into exact paral- lelism with a general statement of fact in the Indicative. Thus, in 8, possis is the amplification of satis est (not sit); in 5, timeas is balanced against neque metus . . . . est, and, in 18, pros- pera omnia cedunt against nequiquam deos implores; in 24, gautdeas is set up as the antithesis of iuvat; in 23, rideas, bibas, and decutmbas are hemmed in between the two general statements ringitur and adponitur, with the parallelism still more strongly brought out by the antithesis ille . .. . tu . . . in 19, frustra iudicia implores is justified by the general state- ment fit, and, in 20, the general statement fit by invenias; in 14, matronae . . . . nil cernere possis is set over against altera nil obstat, and videre est is carried out by metiri possis ocutlo; in 9, nulla potest oculorum ac es . . . . tueri is clearly parallel to nec cernere possis, and, in 13, ut non . . in omni agro reperire possis to sic non .... in omni vita nascitur; while in 17, not only is hortere balanced against a general statement officit, but the two together form the unfolding of the contents of the very significant Indicative solet, "it is regularly the case that." Note similarly solet . . . . desideres in 26. In 28, careas is balanced against caremus. In 7, abalienes and nihilne noceas (bis) are parallel with mediocriterne nocent, quantum est mali, and mediocre peccatum est. And, finally, in 31, the correspond- ence between valeas and habeberis is so obvious that I am almost ashamed to have printed anything of the present paper except the title and this example.

The condition of things being what has been seen, there can be no reasonable doubt that the second person singular of the Latin Subjunctive may be employed to make independent state- ments of a general fact.

Such a conclusion should afford a sensible relief. The variety of explanations adduced above as having been actually applied to one and another of these examples, the heavily forced character of some of them, even for the individual cases to which they are applied, the contradictory explanations made by different writers for the same examples, the shift of ground on the part of one writer,-all show how hard beset grammarians and commentators have been.

In the collections employed, only examples that are surely to be classified as statements of fact are given. Once established, however, the category fits many other examples more naturally than does the category of Possibility, or that of Ideal Certainty. The more important step has been taken. It remains to see whether the second is possible,-whether a probable origin can be found for the Subjunctive construction. I have long taught the following:

The extremely common use of the Subjunctive of the second singular indefinite in generalizing assumptions (relative or tem- poral clauses, or express conditions) might well, in time, affect the mood of the main sentence. To express the generalizing idea in the dependent clause, the verb in any person or number except the second singular indefinite was in the Indicative, but if it oc- curred in this form, then the mood was, by fixed force of habit, the Subjunctive. The Subjunctive would then seem to be a natural mood to use wherever the generalizing force was intended, if the second singular indefinite were employed. Just as else- where, in the expression of this idea, Indicative went with Indica- tive, so here it would seem natural that Subjunctive should go with Subjunctive. If, saying quod quis in manu tenet, one goes on with id desiderat, then, saying quod in manu teneas, it would be natural to go on with id desideres.

This influence would be all the stronger from the fact that, whatever may have been the original force of the Subjunctive in these dependent clauses [1], that force could no longer be clearly felt. The Subjunctive of the second singular indefinite had be- come merely the sign of universality. Why, then, should it re- niain confined to the subordinate clause?

[1] I believe the construction to have originated in a Volitive of the imagination, as in "lose your money, and all your friends desert you" (not, as seems everywhere as- sumed, in the Potential Subjunctive); and I have so classed it in the syntax of the Hale-Buck Grammar, § 504. 2. But a treatment of this question would properly involve a discussion of the Greek generalizing Subjunctive (which I hold to be of parallel origin), and a discussion of my difference from Delbruick with regard to an impor- tant group of Greek constructions, for which this paper affords no room.

This natural tendency toward extension would probably be fur- thered by the fact that the same second person singular indefinite happens to be in frequent use in a number of true independent Subjunctive idioms. Thus it is found in the Subjunctive Question of Perplexity, as in quid agas? "what is one to do?" Iuv. iv. 14; in the Question of Obligation, Propriety, or Reasonableness (to avail myself of a classification and name from own teaching) [2], as in cur tamen corripias? "Why should one chide her?" Iuv. x. 291; and in the form of statement which lies in the neutral terri- tory of Natural Likelihood, Possibility, and Ideal Certainty, as in crederes victos, "one might well think them conquered," "one might think them conquered," "one would think them conquered," Liv. ii. 43. 9.

[2] The Subjunctive of Obligation, Propriety, or Reasonableness, first set up as a general category in my teaching at Cornell, is made one of the principal families of the mood in my general scheme, and several of the hitherto difficult types of dependent clauses seem to find full solution in it (Hale-Buck Grammar, ?§ 512,5 13; also Classical Journal I. 1. 1905, "An Experiment in the Teaching of First and Second Year Latin," where a conspectus is given of the uses of the Latin Subjunctive, classified according to origins). Not only is the mood-use a frequent one, but,,in my opinion, it furnishes the missing link between the true Optative and the Optative (Latin Subjunctive) of Natural Likelihood, Possibility, or Ideal Certainty. (See my "Leading Mood-Forces in the Indo-European Parent Speech," Proceedings of the American Philological Association XXXII, p. cxx [1901].)

The combination of these influences with the one already mentioned might well have led the Romans to feel that, whatever mood-feeling they had to convey, the Subjunctive was the natural mood to use, if the person was the second singular indefinite. But, most of all (after the primary cause), the natural tendency to extend the idiom would be strengthened by the very frequent occurrence of a construction which, though evidently of true Potential origin, clearly suggests and implies fact. Thus Juvenal's Potential in Catilinam quocumque in poputlo videas, quocumque sub axe," "one may see a Catiline in any nation, under any sky," xiv. 41, easily suggests "one sees . . ." That which is said to be anywhere and everywhere possible is easily understood as anywhere and everywhere taking place. The number of ex- amples of this type (videas, cernas, etc.) is very large. An especi- ally instructive set is afforded by reperias. I give three of these in sequence:

Quamvis malam rem quaeras, illic reperias; Plaut. Trin. 554.

Per pol quam paucos reperias meretricibus Fidelis evenire amatores, Syra; Ter. Hec. 58.

Ut enim sunt, quem ad modum supra dixi, qui urbanis rebus bellicas ante- ponant, sic reperias multos, quibus periculosae t calidac onsilia quietis et cogitatis splendidiora et maiora videantur; Cic. Off. i. 24. 82.

In the first of these, "one may find it there," may suggest "one finds it there." In the second, "how few one can find," strongly suggests "how few one |finds;" while in the third, reperias qui- bus . . . . is put into formal parallelism with the general state- ment sunt qui . . . . , "there are people who . . . ." Reperias accordingly is completely ready, at any rate, to convey the idea of fact.

So much for the construction under examination. An inde- pendent use of the Subjunctive to express statements-assertions of fact-in the second person singular indefinite, exists; and it probably is historically an extension from other constructions in which the Subjunctive had, originally, true-mood forces.[1]

[1] Hence in my grammatical scheme, referred to above, this use is placed under the distinct heading of "Subjunctive Constructions Due to the Influence of Other Constructions."

But did the construction become universal, where the idea of a general statement of fact was to be conveyed, and the second singular indefinite was to be used? It seems not. The older usage, which of course must have been that of the Indicative mood, still survived. In the future, the idea had necessarily to be expressed by the Indicative if expressed at all, since the in- dependent Anticipatory Subjunctive had disappeared, wholly or almost wholly, from the language. But there are also examples of the present.

Examples of the Present Indicative

Examples of the Present Indicative

Examples of the Present Indicative [2]

Nescis quid vesper serus vehat; Varro Sat. Men. ap. Gell. xiii. 11. 1; similarly, but without serus, in Macrob. i. 7. 12, and ii. 8. 2. (It would seem that nescis must be general, since in all three passages the sentence to which it belongs is given as the title of a book.)

Bis peccas, cum peccanti obsequium commodas; Publil. 52.

Cum das avaro praemium ut noceat rogas; Publil. 122. Similarly 658. [3]

Ita, si silenda feceris, factum ipsum (scil. culpatur), si laudanda non sileas, ipse culparis; Plin. Ep. i. 8. 15.

[2] Examples are not given from the satiric poets, since they appear often to have been addressing, either through a long passage or for the moment, an imagined indi- vidual (cf. our "Ig entle reader "), and the exact discrimination of force is therefore uncertain for the Indicative mood. See, for instance, in 14 above, and cf. Lebreton Etudes sur la langue et la grammaire de Ciceron, pp. 349 ff.

[3] Publilius's Indicatives in the dependent second person singular indefinite are in- teresting, and will be treated, as well as Blase's view that there is no second person singular Subjunctive of Indefiniteness in dependent Clauses (Studien u. Kritiken II, pp. 5-8, 1905), in a later paper already indicated.

Examples of the Present Indicative [2]

Nescis quid vesper serus vehat; Varro Sat. Men. ap. Gell. xiii. 11. 1; similarly, but without serus, in Macrob. i. 7. 12, and ii. 8. 2. (It would seem that nescis must be general, since in all three passages the sentence to which it belongs is given as the title of a book.)

Bis peccas, cum peccanti obsequium commodas; Publil. 52.

Cum das avaro praemium ut noceat rogas; Publil. 122. Similarly 658. [3]

Ita, si silenda feceris, factum ipsum (scil. culpatur), si laudanda non sileas, ipse culparis; Plin. Ep. i. 8. 15.

[2] Examples are not given from the satiric poets, since they appear often to have been addressing, either through a long passage or for the moment, an imagined indi- vidual (cf. our "Ig entle reader "), and the exact discrimination of force is therefore uncertain for the Indicative mood. See, for instance, in 14 above, and cf. Lebreton Etudes sur la langue et la grammaire de Ciceron, pp. 349 ff.

[3] Publilius's Indicatives in the dependent second person singular indefinite are in- teresting, and will be treated, as well as Blase's view that there is no second person singular Subjunctive of Indefiniteness in dependent Clauses (Studien u. Kritiken II, pp. 5-8, 1905), in a later paper already indicated.

Examples of the Future Indicative

Examples of the Future Indicative

Quantumvisq uare sit macer invenies; Catull. 89. 6.

Nulli facilius quam malo invenies parem; Publil. 395. [4]

Facile invenies qui bene faciant, cum qui fecerunt coles; Publil. 664. [4] divitiae et opes, quas facilius invenies qui vituperet quam qui fastidiat; Tac. Dial. 8.

[4] In the light of the other examples given, W6lfflin's " emendation" to invenias is clearly unnecessary.

Ideo autem (apes) pungunt, quia ubicumque dulce est, ibi et acidum invenies; Petron. 56.

Assem habeas, assem valeas: habes, habeberis; Petron. 77.

Unde et illud proverbium: Quantum habebis, tantus eris; Augustin. De disc. Christ. 11. 12.

My collections for all these idioms are the result of chance happenings upon examples, not of methodical search over the whole ground. They make it fairly clear, however, that the Subjunctive is distinctly common, the Present Indicative, outside of Publilius and the satiric poets (see the last footnote but two), very rare. The statement in the Hale-Buck Grammar is accord- ingly not strong enough. For "sometimes," "regularly" should be substituted. It should be noted also that, of the various verbs employed in the Subjunctive idiom, possis, " one can," seems to be in largest use.

Finally, can any definition be made of the difference between the Subjunctive and Indicative idioms? The question may be answered as follows:

In the Subjunctive construction, the mood is a sign of gener- ality. In the Indicative construction, the mood is not a sign of generality. One feels generality, if the context demands it; but it is not hinted at by the mood itself. The difference in the idioms is like the difference between English "you" and "one." " You " may turn out to be either individual or general. "One " is immediately and necessarily general. Compare the "Tacit causal qui-clause," Indicative, and the "Explicit causal qui- clause," Subjunctive (Hale, Cum-constructions, pp. 97, 112 f., or 107, 127 f. of the German translation).

Quantumvisq uare sit macer invenies; Catull. 89. 6.

Nulli facilius quam malo invenies parem; Publil. 395. [4]

Facile invenies qui bene faciant, cum qui fecerunt coles; Publil. 664. [4] divitiae et opes, quas facilius invenies qui vituperet quam qui fastidiat; Tac. Dial. 8.

[4] In the light of the other examples given, W6lfflin's " emendation" to invenias is clearly unnecessary.

Ideo autem (apes) pungunt, quia ubicumque dulce est, ibi et acidum invenies; Petron. 56.

Assem habeas, assem valeas: habes, habeberis; Petron. 77.

Unde et illud proverbium: Quantum habebis, tantus eris; Augustin. De disc. Christ. 11. 12.

My collections for all these idioms are the result of chance happenings upon examples, not of methodical search over the whole ground. They make it fairly clear, however, that the Subjunctive is distinctly common, the Present Indicative, outside of Publilius and the satiric poets (see the last footnote but two), very rare. The statement in the Hale-Buck Grammar is accord- ingly not strong enough. For "sometimes," "regularly" should be substituted. It should be noted also that, of the various verbs employed in the Subjunctive idiom, possis, " one can," seems to be in largest use.

Finally, can any definition be made of the difference between the Subjunctive and Indicative idioms? The question may be answered as follows:

In the Subjunctive construction, the mood is a sign of gener- ality. In the Indicative construction, the mood is not a sign of generality. One feels generality, if the context demands it; but it is not hinted at by the mood itself. The difference in the idioms is like the difference between English "you" and "one." " You " may turn out to be either individual or general. "One " is immediately and necessarily general. Compare the "Tacit causal qui-clause," Indicative, and the "Explicit causal qui- clause," Subjunctive (Hale, Cum-constructions, pp. 97, 112 f., or 107, 127 f. of the German translation).