.i^f. 
 
 i&!^n\vimin She WihtAtv 
 
GIFT OF 
 
 "y^r^/r; 22/e yV/?eG./e>r^^ 
 
10 
 
THE 
 
 Sequence of Tenses in Latin 
 
 BY 
 
 WILLIAM GARDNER HALE 
 
 Professor of the Latin Language and Literature in Cornell University 
 
 Read at the Meeting of the American Philological Association 
 HELD AT Ithaca, July 1886 
 
 Reprinted from The American Journal of Philology, Vol. VII, No. 4 and VoL Vllf, No. i 
 
 BALTIMORE 
 
 PRESS OF ISAAC FRIEDENWALD 
 1887 
 
d'V 
 
2^ so 
 
 I 2^1 
 
 THE SEQUENCE OF TENSES IN LATIN. 
 
 I. 
 
 The doctrine of the Sequence of Tenses has two sides, a 
 theoretical and a pedagogical. It is the purpose of the present 
 and the succeeding paper to examine it with some fullness on the 
 former side, and briefly on the latter. 
 
 The doctrine is stated in various ways, which deal with the sub- 
 ject, some more, some less, externally. For an example of the 
 former this will serve : In subordinate clauses the tenses of the 
 subjunctive conform to the following rule : principal tenses depend 
 upon principal tenses, historical upon historical — a form of state- 
 ment which contents itself with tabulation, and does not touch 
 ground. Deeper-reaching is the statement that the choice of the 
 tense in each sentence is determined by the tense of the verb on 
 which the sentence in question depends. 
 
 A modified form of the doctrine will be discussed later. At the 
 outset our concern is with the prevailing view. ^ 
 
 For convenience' sake, we may state that view, with justice to 
 all parties thus far included, in some such way as this : The tense 
 of the subordinate clause is found to be under the influence of the 
 tense of the main clause, or, as Engelmann puts it (Schneider's 
 translation, p. 308), "A subjunctive clause is, in regard to its tense, 
 dependent on the principal sentence." 
 
 To this statement the literature offers exceptions, some speci- 
 mens of which (mostly confined to clauses of result) are given in 
 the grammars. Our examination starts from a scrutiny of these 
 exceptions, beginning with the so-called primary tenses. 
 
 I. In Co7isecutive Clauses after UT. 
 
 a. The present : 
 
 Nam priores ita regnarunt ut haud inmerUo omnes deinceps 
 conditores partium certe urbis, quas novas ipsi sedes ab se auctae 
 multitudmis addiderunt, numerentur. Liv. 2, i, 2. For the pre- 
 decessors of Tarquin the Proud reigned in such a manner that 
 
 4274 la 
 
.• •• • 
 
 • • • •• , 
 
 • • 
 
 we very properly regard ihem all as founders of the city, etc. The 
 verb regnarunt belongs, according to the traditional chronology, 
 to the years 753-534 ; while the verb numerentiir belongs, just as 
 an indicative manerantur would, to the age of Livy. 
 
 XoX/}j; uKparov iioctii eieci ; statim ita sum levatus tit mihi deus 
 aliquis medicinam fecisse videatur. Cic. Fam. 14, 7, i. . . . in 
 an histant I was so relieved t/iat the cure has the look of a miracle. 
 
 In eodem {^Luculld) tanta prudentia fuit in constituendis tempe- 
 randisque civitatibus, taiita aequitas, ut hodie stet Asia Luciilli 
 institutis servandis et quasi vestigiis persequendis. Cic. Acad. 2, 
 I, 3. Lucullus took so long a look ahead in establishing forms of 
 government, and had such a sense for justice, that to-day Asia 
 stands by holding to his arrangements and following, so to speak, 
 in his tracks. 
 
 . . . ifi provincia Sicilia, quain iste per trienniwn ita vexavit ac 
 perdidit ut ca restiiui in antiquum statuni nullo modo possit, vix 
 autein per multos annos imiocentisque praeiores aliqua ex parte 
 recreari aliquando posse videatur. Cic. Verr. Act. Pr. 4, 12. 
 For three years this fellow so harried and ruined Sicily that there 
 is no possible way of restoriiig her to her old condition, etc. 
 
 The comment of Allen and Greenough (p. 201) upon the pas- 
 sage is as follows : " Here the present is ut-ed in describing a state 
 of things actually existing " ; which of course means at the time 
 when Cicero made the speech. The modally dependent /<?>5-^// and 
 videatur, then, mean, so far as tense alone goes, precisely the same 
 thing as would /<?/^^/ and videtur. Barring the formal expression 
 of degree and result, Cicero might equally well have said Siciliam 
 iste per triennium vexavit ac perdidit ; neque ea restiiui in anti- 
 quum statum ullo modo potest, etc. We may then state as a 
 formula for this particular case: mood apart, possit :=. potest, 
 V ide alter -=, videtur. And, governed by this and an abundance of 
 similar cases, we are obliged, whatever our prepossessions may be, 
 to lay down the statement that after secondary tenses the present 
 subjunctive in consecutive ^//-clauses expresses, shows that the 
 speaker means, time present to his speaking ; that, in other words, 
 the present subjunctive in consecutive 2^/-sentences after a secondary 
 verb is wholly independent of any Sequence of Tenses. 
 
 But if the present has this power after secondary tenses, it 
 clearly cannot be asserted not to have it after primary tenses, as 
 for example in the following : Nam sociorum auxilia propter 
 acerbitatem atque iniurias imperii nostri aut ita imbecilla sunt 
 
ut non muUum nos hivare possint, aut ita alienata a nobis tit 
 neque exspectandicin ab its neque com^nittendutn Us quicqua7n 
 videatur. Cic. Fam. 15, i, 5. It is idle to hold that the present 
 subjunctives of posse and videri, though able to express a certain 
 temporal idea in the teeth of the Sequence of Tenses, as in the 
 former example, have not the ability to express the same meaning 
 in ordinary use, as in the latter example — that, in other words, 
 they can by some mysterious accession of power set the law at 
 defiance and become tense-expressing, having, nevertheless, no 
 power to express tense. 
 
 On this point — this argument from the power clearly seen to be 
 present in the tense in the unusual combination (or, in common 
 phraseology, in the exception) to the meaning of the tense in all 
 constructions of the kind, whether unusual or usual — the whole 
 matter hinges. I must therefore insist upon it and emphasize it. 
 
 We find, in consecutive 2^/-sentences, the present subjunctive in 
 combination with a preceding present, etc., and in combination 
 with a preceding aorist, etc. Now, how much do we absolutely 
 know of the force of the subjunctive present in any one of these 
 result-clauses ? We absolutely know that in one of the combina- 
 tions it has the force, conveys the meaning, of the present — a 
 meaning precisely the same, mood apart, as that of the present 
 indicative. Next, is there anything to detain us fivom supposing 
 (as we should at once naturally proceed to do) that this mean- 
 ing, which indubitably exists in the one set of cases, exists in: 
 the other? Is there anything to indicate that the speaker, 
 using precisely the same expression in the two sets of cases, 
 meant one thing in the one set, and another thing in the other? 
 Nothing whatever. Very good. The rational interpretation of 
 the entire field of phenomena accordingly is that the speaker 
 attaches a present result, now to a cause temporally near, now to 
 a cause remote. But why (for we are naturally curious to under- 
 stand the one point remaining) do we find, as we read Latin 
 literature, that there are a great many examples of a present result 
 attached to a cause temporally near, and comparatively few exam- 
 ples of a present result attached to a remote cause ? The reason 
 is the simple fact, familiar to everybody, that very few results abide 
 long, or, at any rate, are appreciated as long-abiding. There are, 
 in the year 1887, abiding results of the victory of the Greeks over 
 Xerxes, and we may therefore say at ihe battle of Salamis the 
 Greeks routed their enemies^ in consequence of which victory our 
 
^modern civilizaHon is essentially Greek, not Oriental ; but as we 
 read Greek and Roman history, or any other, we find that results 
 are mostly immediately attendant upon their causes, and have no 
 visible effect upon the state of things at the time when the history 
 was written ; and, consequently, after the expression of a past 
 cause we find a great many expressions of a past result,* and very 
 few expressions of a present result. All this falls in perfectly with 
 the way things are in this world. But to set up a theory which, 
 to apply it to perfectly analogous phenomena in English, would 
 maintain that I say " am " in the sentence / hurt my leg badly tefi 
 years ago, so that /am finable to walk, because I mean now, while 
 in the sentence / have hurt my leg badly, so that /am unable to 
 walk, I don't say ''am'' because I mean now, but say it (not 
 meaning anything temporal by the word itself) because my mind 
 is under the control of the tense of the main verb — to set up such 
 a theory as this is to see things only as they appeared in the early 
 part of the present century, in the world of certain' grammarians 
 
 ' The history of the common use of the imperfect to indicate results seen in 
 temporal connection with the past will be given later. 
 
 2 Some of the grammarians of the time dealt with the matter in a very reason- 
 able way. Kriiger, for example (Untersuchungen aus dem Gebiete der lat. 
 Sprache, II Theil, 1821), has an admirable statement of the force and uses of 
 the subjunctive tenses ; and though in his grammar as edited by Grotefend in 
 1842 (I have not his own edition at hand) the formulae in fashion at the 
 present day are given, yet the true grounds of the phenomena are rightly 
 sketched in §617, note i. The doctrine of Wenck (Lat. Sprachlehre, 1798, 
 §§26, 165-9, ^"^ particularly §168 with note) is so much juster than that of the 
 school-grammars of the present day that I must allow myself to quote a few 
 lines, with italics of my own: " Doch muss bei dieser Kegel auf die eigentliche 
 Bedeutung der Temportim^ folglich auf die Sache selbst, Riicksicht genommen 
 werden. Es versteht sich z. B. von selbst, dass, wenn nach einem Praesens 
 wieder ein Praesens folgen soil, von einer gegenwartigen Sache geredet werden 
 mlisse." That which, as Wenck correctly says, versteht sich von selbst, has un- 
 happily, through the pedagogical stiffening and congealing of modes of expres- 
 sion in the grammars, become a matter which we of to-day have to set ourselves 
 to prove by formal reasoning. 
 
 It is interesting and cheering to note that we have quite outlived a " common 
 rule" of a similar nature, which had some vogue at the time when the doctrine 
 of the Sequence was growing up, namely, that " these coujunctions " {et, ac, atque, 
 etc.) '^ connect the same tenses'''; a doctrine which Schelling (Walker's transla- 
 tion, 1825, II, p. 185) sets himself to controvert, taking the very natural ground 
 that they " connect the same tenses only so far as when the same tense is in- 
 tended " — a ground directly applicable, mutatis mutandis, to the doctrine of the 
 Sequence. 
 
5 
 
 who failed to feel the play of human thought, and saw their phe- 
 nomena only by inventories.^ 
 
 Further,* such a theory lands one at once in a plain absurdity. 
 Let us apply it to the expression of the results of the administra- 
 tion of Asia under Lucullus, taking Cicero's word, in the example 
 above, for the character of that administration. It lasted, we will 
 say, from the year 74 to the year 66. Cicero wrote the Academica 
 21 years later. The advocate of the Sequence of Tenses, then, 
 is bound to hold that in the year 45 the present tense stet was in 
 itself competent, just as stat would have been in the paratactical 
 construction, to express the then-existing result of Lucullus's 
 activity 21 to 29 years earlier, but that on the day following 
 Lucullus's return in 66 the very same word would have been 
 incompetent to do anything of the kind ! The tense, it would 
 appear, must be put away and allowed, like new wine, to ferment, 
 before it can have any power to express itself. But such a view, 
 to speak very temperately, seems to a plain mind a more difficult 
 doctrine than the doctrine that the tense has everywhere, in a 
 given construction, the meaning which it is absolutely known to 
 have, in that construction, in a good many cases. 
 
 We shall therefore have to amplify our former statement, and to 
 lay down the following: In consecutive sentences after ut, and 
 after all tenses, whether secondary or primar}^, the present tense 
 expresses time present to the speaker; or, in other words, the 
 present tense of the subjunctive in consecutive 2^/-sentences is 
 altogether free from, wholly independent of, in no way concerned 
 even with the existence of, the doctrine of the Sequence of Tenses. 
 
 b. The perfect definite : 
 
 Quamquam enijn adeo excellebat Aristides absiineniia, ut untis 
 post hominum memoriam, quern quidem 7ios audierwtus, co- 
 gnomine Justus sit appellatus, tamen a Themisiocle collabef actus 
 testula ilia exsilio decern annorum multatus est. Nep. Arist. i, 2. 
 For, though Aristides was so pre-eminent for his respect for other 
 7nen's rights that he is the only man who has been named the Just, 
 yet he was ostracized, etc. 
 
 Ardebat autem cupiditate dicendi sic, ut in nullo umqtiam 
 
 ' Hiibner's general bibliography of the subject (Grundriss, II Theil, § 39) 
 may be interestingly supplemented, from the point of view of a contemporary 
 of the rise of the doctrine, by a list of titles given by Stallbaum in his edition 
 of Ruddimann's Institutiones Grammaticae Latinae (Leipzig, 1823, Vol. II, 
 P- 341). 
 
Jlagrantms stiidmm viderim. Cic. Brut. 88, 302. He {Hor(ensius) 
 was possessed with such a passion for speaking that /have never 
 seen in anybody a more burning ardor. 
 
 In these examples the perfect subjunctive is used, though after 
 an imperfect, as a perfect definite, meaning, so far as tense is con- 
 cerned, precisely the same thing as would the perfect indicative. 
 Barring the formal subordination of result to cause, Cicero might 
 have said, with precisely the same temporal force : Horte?isiiis 
 ardebat dicendi cupiditate ; nee in ullo umquam flagrantius 
 studium vidi. 
 
 We must therefore lay down the statement that in consecutive 
 ^^/-sentences after secondary tenses the perfect definite of the sub- 
 junctive has a time-expressing power of its own — is under no law 
 of a Sequence of Tenses. 
 
 But if the perfect definite has this power after secondary tenses, 
 then, by the same reasoning as in the case of the present above, 
 it is idle to maintain that it has not the same power after primary 
 tenses, as in Cic. Div. in Caecil. 1,1: Si quis vestrum, indices, 
 forte 7niratur me, qui tot ajinos in causis iudiciisqiie publicis ita 
 sim versatus ut defenderim niultos, laeserim 71 niine^n, etc. The 
 perfect subjunctive (^defenderim, laeserim) is here doing after a 
 primary tense just what we saw it doing above after secondary 
 tenses, and to grant a power of expression in the former case 
 while denying it in the latter is, as we have seen, to set up 
 distinctions founded on no differences, and involving gross 
 absurdities. 
 
 We shall accordingly be obliged a second time to amplify a 
 statement, and to lay down the following : In consecutive ut-sen- 
 tences, and after all tenses, whether primary or secondary, the 
 perfect definite of the subjunctive conveys the idea that the act 
 indicated by it is completed at the time of speaking ; or, in other 
 words, the perfect definite is altogether free from, wholly inde- 
 pendent of, in no way concerned with the existence of, the doctrine 
 of the Sequence of Tenses. 
 
 c. The aorist : 
 
 Barbarus . . . adeo angiisto mart conflixit ut eius multitudo 
 navium expiicari non potuerit. Nep. Them. 4, 4. Xerxes en- 
 gaged his enemy in such a narroiv strait that he could not bring 
 the great mass of his ships into actio?i. 
 
 XXV. indices ita fortes tamen fuerunt utsiwimo proposito periculo 
 vet perire maluerint qnam perdere omnia. Cic. Att. i, 16, 5. 
 
Twenty-five of the judges ^ however y were so bold that they pre- 
 ferred the risk of utter destructioji to the risk of losing all. 
 
 The phenomenon is a very familiar one. The aorist is used in 
 precisely the same temporal sense as that which is conveyed by 
 the aorist indicative. 
 
 We must therefore lay down the statement that after secondary 
 tenses the aorist subjunctive in consecutive 2//-clauses expresses, 
 mood apart, the same idea as the aorist indicative — has, in other 
 words, a power of its own, and is under no law of a Sequence of 
 Tenses. 
 
 The idea to be conveyed by an aorist of result after a primary 
 tense can exist only when the main verb states a cause that always 
 exists or has always thus far existed, and the result-clause cites an 
 historical case illustrating the working of that cause. Such an 
 example (which might be illustrated in English by the sentence 
 The lust of power is so great that even the Fouyider of Rome slew 
 his own brother) naturally occurs rarely, and I have had the bad 
 luck to lose one which I had found. Still, it is clear that, in the 
 few examples that may occur, the force of the aorist in consecutive 
 ^/-clauses is the same, mood apart, as that of the aorist indicative. 
 And the omission of an example of this kind cannot count against' 
 my case, because such examples are recognized by the law of the 
 Sequence as regular. 
 
 We must therefore again amplify what we have said, and assert 
 that in consecutive 2//-clauses, no matter whether after secondary 
 or after primary tenses, the aorist subjunctive conveys, of its own 
 power, an idea of time (the same, mood apart, as that of the aorist 
 indicative), and is, consequently, under no law of a Sequence of 
 Tenses ; and further, summing up what has been shown under a, 
 b, and c, we must lay down the larger statement that in consecu- 
 tive 2/^clauses the present, the perfect, and the aorist have in 
 themselves a tense-expressing force — owe their use, not to the 
 dictation of a preceding verb, but to their own power to convey 
 the temporal meaning which the speaker has in his mind — in other 
 words, are entirely unconcerned with any law of the Sequence of 
 Tenses. 
 
 2. In Consecutive Relative Sentences, 
 
 a. The present : 
 
 Erat no7i studiorum tantum verum etiam studiosorum amanlis- 
 simus^ ac prope cotidie ad audiendos quos tunc ego frequentabam 
 Quintilianum, Niceten Sacerdotem ventitabat, vir alioqui clarus et 
 
8 
 
 gravis et qui prodesse filio memoria sui debeat. Plin. Ep. 6, 6, 3. 
 He was extremely fond, not only of literary pursuits but of literary 
 people, and used to go nearly every day to the lectures of Quin- 
 tilian and Nicetes Sacerdos {with whom /was at that time taking 
 courses')— a man of distinction and weight, who ought to be of 
 assistance to his son through the memories he has left behind him. 
 
 Note that after the tenses of a past activity, erat, frequeniabam, 
 veniitabat, the characterizing clause expresses with perfect ease 
 and certainty, merely through the force of the tense, a now-existing 
 state of affairs. Debeat is simply a subjunctive debet. 
 
 Hi fere fuerunt Grace ae gentis duces qui memoria digni vide- 
 antur, praeter reges. Nep. de Reg. i. These, we may say, were 
 the generals of the Greek race, outside of royalty, who seem to be 
 worthy of a place in history. 
 
 Examples of this sort could easily be produced by scores, but 
 our limits of space make economy necessary.' Furthermore, I shall 
 not, after the present set, take space to treat present, perfect, and 
 aorist separately, nor to treat imperfect and pluperfect separately, 
 since that which holds for a part of a set holds for the rest also ; 
 and I shall no longer repeat the arguments by which, under i, I 
 showed that a temporal power conceded to a primary tense after a 
 secondary tense must also be conceded to it after a primary, and 
 vice versa. 
 
 We must then lay down the statement that after all tenses, 
 whether primary or secondary, the present of the subjunctive in 
 consecutive relative clauses conveys of itself the force of a present; 
 or, in other words, the present subjunctive in these sentences is in 
 no way concerned with the existence of a doctrine of the Sequence 
 of Tenses. 
 
 b. The perfect definite : 
 
 Quis tum fuit Syracusis quin audierit, quin sciat has Timar- 
 chidi pactiones sepulturae cum vivis etiam illis esse factas 9 Cic. 
 Verr. 5, 45, 120. Who that was at Syracuse at the time has not 
 heard, does 7iot know, etc. 
 
 ^ The giving of an abundance of examples would have the good effect of 
 showing, with a cumulative influence upon the reader's mind, the entire free- 
 dom with which, in the great mass of constructions at any rate, the Roman 
 said in his subordinate verb that which would express his meaning, without 
 paying any consideration whatever to anything that he had previously said in 
 another verb. In this part of my paper, and in many other parts, I regret that 
 the case to be presented, in the face of the traditional doctrine, must be so 
 curtly stated. 
 
In consecutive relative clauses, consequently, the perfect definite, 
 whether after secondary or after primary tenses, is in no way the 
 product of the dictation of a preceding verb, but has in itself tem- 
 poral expression— in other words, is unconcerned with the existence 
 of the doctrine of the Sequence of Tenses. 
 
 c. The aorist : 
 
 Nulla domus in Sicilia locuples fuit, ubi iste no?i texirinum 
 instituerit. Cic. Verr. 4, 26, 58. There wasn't a well-to-do house 
 in Sicily where he didn't set people to weaving. 
 
 Fuerunt quos fames magis guam fama commoverit. Cic. Att. 
 I, 16, 5. There were some over whom famine had more power 
 than fame. 
 
 The doctrine of the Sequence of Tenses, therefore, has nothing 
 to do with the aorist subjunctive in consecutive relative sentences. 
 
 From the results of our examination under «, b, and <:, then, we 
 learn that in consecutive relative clauses, after whatsoever tenses, 
 the present, the perfect definite, and the aorist have in themselves 
 the power of temporal expression ; and that, consequently, the 
 doctrine of the Sequence of Tenses is not for them. 
 
 3. In Causal Sentences. 
 
 Turn ille ** locabatur," inquit, ''Catulus, praesertim cum ita 
 dicat ipse ut a^nbrosia alendus videatur.^' Ci«j. de Or. 2, 57, 234. 
 Then spoke up Crassus : '' Catulus vf2iS certainly joking when he 
 said that, for he himself is such ayi orator that it seems as if he 
 must live on a diet sent from Heaven ^ 
 
 Non ego ig7iar2is quid responsurus facturusve esses quaesivi, 
 quippe cujii prae te feras temptare te magis qua^n coyisulere se7ia- 
 tum. Liv. 28, 45, 3-4. In asking my question I was not in 
 doubt what your answer and your course of action would be, for 
 you show very plainly that you are tryi72g to find out the feeling of 
 the Senate instead of formally asking its vote. 
 
 Sed nee eiusmodi est ut a pluribus confusa videatur ; unus enim 
 sonus est totius oratiotiis et idem stilus ; nee de Persia reticuisset 
 Gracchus, cum ei Fanjiius de Menelao Maratheno et de ceteris 
 obiecisset, praesertim cum Fannius numqua^n sit habitus elinguis. 
 Cic. Brut. 26, 100. (An allusion has been made to the author- 
 ship of the Oratio de Sociis ascribed to Fannius, but thought by 
 some to have been written by Persius.) But it hasn't the look of 
 a co7nposite ; for the whole oration rings like one and keeps up a 
 uniform style ; nor would Gracchus have held his tongue in regard 
 
lO 
 
 to Persius when Fannius gibed him about Mefielaus of Mara- 
 this, to say nothing of the still stronger consider atio7i that Fannius 
 has 7iever been regarded as a man who couldn't speak. 
 
 Fuit eniyn inirijica vigilantia, qui suo toto consulatu soninuni 
 non viderit. Cic. Fam. 7, 30, i. He was a tremendously wide- 
 atvake ma7i,for during his entire consulship he didn't know what 
 sleep was. 
 
 Ille vero ante decemviros non fuit, quippe qui aedilis curulis 
 fuerit, qui magistratus multis annis post decemviros institutus est. 
 Cic. Att. 6, I, 8. He didn't live before the decemvirs, for he was 
 curule aedile, and that office was not created till long after the 
 time of the decemvirate. The tense is as free in fuerit as in histi- 
 tutus est. 
 
 The primary tenses of the subjunctive in causal sentences, then, 
 are in themselves expressive, and are exempt from any law of a 
 Sequence of Tenses. 
 
 4. In Concessive Sentences. 
 
 Nam cum apud Graecos antiquissimum e doctis genus sit pocta- 
 rum, si qtiidem Homer us fuit et Hesiodus ante Roma?n conditam, 
 Archilochus regnante Roniulo, serius poetic am nos accepimus. 
 Annis fere CCCCCX post Romam conditajn Livius fabulam dedit 
 C. Claudio Caeci filio M. Tuditaiio consulibus anno ajite natum 
 Ennium : sero igitur a nostris poetae vel cogniti vel recepti. Cic. 
 Tusc. I, I, 3. Though in Greece poets are the oldest class of 
 literary men . . . we Romans took to poetry later. . . . So our 
 nation either became acquainted with the poets late, or took to 
 them late. Sit is simply a subjunctive est. 
 
 Nam primum, id quod dixi, cum in ceteris coloniis Hviri appel- 
 lentur, hi se praetores appellari volebant. Cic. Leg. Agr. 2, 34, 
 93. For, to begin with, though in all other colonies S2ich officers 
 are called duimiviri, these people were desirous of being called 
 praetors. 
 
 Nam hoc toto proelio, cum ab hora septima ad vesperum pug- 
 natum sit, aversum hostem videre 7iemo potuit. Caes. B. G. i, 26, 
 I. Though the battle lasted //// eve7ii7ig, 7iobody could catch sight 
 of a7i e7iemys back. 
 
 Quae cum 077t7iia facta sint, ta7nen unam sola77i scitote esse 
 civitate77t Mamerti7iajn quae publice legatos qui istiwi laudarent 
 miserit. Cic. Verr. 2, 2, 5, 13. Though all this was done, stilly 
 you 77tust under stand y there is only one state that sent a delegation 
 to whitezvash him. 
 
II 
 
 The primary tenses of the subjunctive In concessive sentences, 
 then, are In themselves expressive, and are exempt from any law 
 of a Sequence of Tenses. 
 
 5. Li the Indirect Discourse, etc, 
 
 Haec in omnibus Ebtiromini partibus gerebanttir, diesque 
 adpetebat sepfimus, quern ad diem Caesar ad i'tnpedinienta legio- 
 ne?nque reverti constiiuerat. Hie, quantum in bello for tuna possit 
 et quanios adferat casus, cognosci potuit. Caes. B. G. 6, 35, 1-2. 
 . . . at this juncture it was possible to recognize what a power 
 Fortune Is in war, and what ups and downs she brings about. The 
 reflection is put as a general one, called up by the recital of the 
 story. Adferat Is simply a subjunctive adfert. 
 
 Quamobrem aute^n in hoc provinciali dclectu spem habeatis 
 aliquam, causa 7iulla est : neque multi sunt et diffugiunt qui stmt 
 metu oblatu ; et, quod genus hoc militum sit, iudicavit vir fortis- 
 simus M. Bibulus in Asia, qui, cum vos ei permisissetis, dilectum 
 habere noluerit. Cic. Fam. 15, i, 5. There is 710 reason for 
 your basing any hopes on the levy in this proviiice : there are few 
 men here, and the few that there are ru7i away as soon as they 
 meet with anything to be afraid of ; 07i the questio7i what kind of 
 soldiers they make, Bibulus expressed his opinion iii refusing to 
 hold a levy, etc. Sit is a general present precisely parallel to szpit 
 and diffugiu7it, differing from them in no respect whatever except 
 in that it is put indirectly. 
 
 Quae quantimi in provincia valeant, vellem expertus essem, sed 
 tame7i suspicor. Cic. Fam. 13, 6a, 4. /could wish /had learned 
 fro7n experience hoiv far these things count iyi a provi7ice, but even 
 as it is I have my suspicions. 
 
 Docui, cum desertum esse dicat vadimo7iiu7}i, om7iino vadi- 
 mo7iiur}t 7iullum fulsse : quo die hunc sibi pro77iisisse dicat, eo die 
 7ie Rornae quide77i eum fulsse. Cic. Quint. 28, 86. / showed thai, 
 whereas he claims that the recognizance had been forfeited, there 
 never was aiiy recog7iiza7ice in the case: that 07i the day on which 
 he claims that Quinctius gave it, Quinctius waswV even i7i town, 
 Dicat (the second) differs from dicit only in being indirectly put. 
 
 Postea recitavi edictum, quod aperte do)7ii7t7im de praedio det7'udi 
 vetaret : i7i quo constltit Naeviti7n ex edicto 7ion possedisse, C2im 
 confiteretur ex praedio vi detrustmi esse Qui7ictiu77i. 077i7iino 
 autem bona posse ssa 7ion esse constltui, quod bo7ioru77i possessio 
 spectetur 7i07i in aliqua parte, sed in U7iiversis, quae teneri et 
 
12 
 
 possideri possint. Cic. Quinct. 29, 89. Next I x^2.^ the prohibitory 
 edict . . . /established the point that his goods had not been in pos- 
 session, for the reason that the test 0/ possession 0/ goods X\es i?i 
 the field of the entire property, not a part of it, etc. The quod 
 spectetur is given as a universal principle, applied to a particular 
 case in the past. 
 
 Audire vie meniini ex senioribus visum saepius inter manus 
 Pisonis libellum, quern ipse 7ion vulgaverit. Tac. Ann. 3, 16, i. 
 / remember hearing from men older than myself that Piso was 
 seen a number of times to have a note-book in his hands, which he 
 did 7iot make public. Vulgaverit is simply an indirect vulgavit, 
 with precisely the same temporal force, the difference concerning 
 nothing but the mood. 
 
 Cur abstinuerit spectaculo ipse, varie trahebant. Tac. Ann. i, 
 76, 6. To the question why he stayed away from the show hifn- 
 self, people at the time gave all sorts of answers (as if we should put 
 it why did he stay away f that was the question^. 
 
 Quae fuerit hesterno die C. Pompei gravitas in dicendo . . . 
 perspicua admiratione declarari videbatur. Cic. Balb. i. What 
 a weighty affair Pompeys speech of yesterday was, was clearly 
 shown at the time by the evident admiration of his hearers. Mood 
 apart, the tenses mean the same as if Cicero had written gravis 
 fuit hesterno die or alio C. Pompei, ut perspicua admiratione 
 declarari videbatur. 
 
 Id quantae saluti fuerit universae Graeciae, bello cognitum est 
 Persico. Nep. Them. 2, 4. What a salvation it was to the whole 
 of Greece was seen in the Persian war (=r saluti fuit, ut bello cogni- 
 tum est Persico). 
 
 In the indirect discourse, etc., then, the primary tenses of the 
 subjunctive convey a temporal meaning, and are under no law of 
 any Sequence of Tenses. 
 
 6. In Conditions. 
 
 Si hodie bella sint, quale Etruscum fuit, cum Porsinna lani- 
 culum insedit, quale Gallicum modo, cum praeter Capitoliurn 
 atque arcein omnia haec hostium erant, et consulatum cum hoc M. 
 Furio et quolibet alio ex patribus L. ille Sextius peteret, pos- 
 setisne ferre Sextium hand pro dubio coyisulem esse, Camillum 
 de repulsa dimicare? Liv. 6, 40, 17. If i?i our own times there 
 should be wars like the Etruscaii . . . or the Gallic . . . , and Sextius 
 were running for the consulship, could you endure, etc. ? The 
 
13 
 
 first condition looks to an entirely possible contingency, and the 
 second, with its conclusion, to something hardly conceivable. 
 
 The whole set of so-called mixed conclusions and conditions 
 containing a secondary conclusion falls under this head. For it is 
 a part of the doctrine of the Sequence that the imperfect and plu- 
 perfect subjunctive in conclusions contrary to fact are regularly 
 followed, by the secondary tenses. 
 
 The primary tenses of the subjunctive, then, in conditional 
 sentences, are in themselves expressive, and are exempt from any 
 Law of a Sequence. 
 
 7. In Conclusions^ Softened Statements, etc, 
 
 Sed post aliquanto propter has amplitudines sepulcrorum, quas 
 in Ceramico vidimus, lege sanctum est ne quis sepulcrum faceret 
 operosius quam quod decem homines effecerint triduo. Cic. Leg. 
 2, 26, 64. But somewhat later, on account of the great scale on 
 which the tombs we have seen in the Ceramicus were built, it was 
 enacted that no one should construct a tomb more elaborate than ten 
 men could make and finish up in three days. The mechanism of our 
 English tongue fails to show that effecerint is put by Cicero in the 
 generalizing form (as if he had said ten men would accomplish a 
 certain amount of work in three days ; and that amount, it was 
 provided by law, was not to be exceeded). x 
 
 Quid? tu me hoc tibi mandasse existimas, ut mihi gladiatorum 
 compositio7ies, ut vadimonia dilata et Chresti conpilatione?n mit- 
 teres et ea, quae nobis, cum Romae sumus, narrare nemo audeat ? 
 Cic. Fam. 2, 8, i. Is it your understanding, my dear fellow, that 
 my instructions to you were that you should send me news of 
 matches of gladiators, of postponements of cases, and Chrestus's 
 bundle of gossip, arid things which, whe7i I am in town, no one 
 would venture to tell me f Sumus and audeat are alike free in 
 tense. 
 
 In conclusions, softened statements, etc., then, the primary 
 tenses of the subjunctive are in themselves expressive, and are 
 exempt from any Law of a Sequence. 
 
 8. In Final Clauses, 
 
 Nam, ne vos falsa opinio teneat, iniussu meo Albani subiere ad 
 montem, nee imperium illud meum, sed consilium et imperi simu- 
 latio fuit, ut nee . , . et terror ac fuga iniceretur. Liv. i, 28, 5. 
 
14 
 
 For, 7iot to leave you in error {lest you may misimderstand) , it was 
 not at my bidding that the Albans went up the hill, nor was it a 
 command of mine, but a device to throw the enemy into a panic {in 
 order that the enemy might be thrown i?ito a pafiic). The tense- 
 less phrase in order to, used ahke for present and past purposes in 
 Enghsh, fails to convey the temporal ideas conveyed by the Latin 
 present and imperfect subjunctive. 
 
 It will not do to answer that such subjunctives depend upon 
 omitted verbs. The question is, do the subjunctives of themselves 
 convey to us temporal ideas ? To concede that they tell us that a 
 verb is omitted, and that they tell us, moreover, just what kind of a 
 tense that verb would be in, if expressed, is to concede to them 
 very great temporal significance. 
 
 In final clauses, then, the primary tenses of the subjunctive are 
 expressive of temporal relations, and owe their choice to that fact, 
 and not to any Sequence of Tenses. 
 
 The examination has now covered the ground of the dependent 
 present, perfect definite, and aorist subjunctive, outside of a very 
 few constructions. We may accordingly, and for the last time in 
 this field, bring together our statements into the following (re- 
 serving, for the present, the few constructions alluded to) : 
 
 In the great mass of constructions, the present, perfect definite, 
 and aorist of the subjunctive directly express the temporal aspect 
 of the act conveyed, as it appears to the speaker's mind at the 
 moment of the utterance of the verb in question. They have 
 nothing to do with any Sequence of Tenses. If there be a control 
 exercised by main verbs over dependent verbs, its field must be 
 sought for on other ground. 
 
 We pass to the remaining tenses of the subjunctive. Have they 
 by some freak of linguistic fate fared differently ? 
 
 I. In Consecutive Clauses after UT. 
 
 Haec enim {philosophid) una nos cimi ceteras res omnis, tiun, 
 quod est difficillijnum, docuit, ut nosniet ipsos nosccrernus : cuius 
 praecepti tanta vis et tanta se7itentia est, ut ea non honihii cuipiam, 
 sed Delphico deo tribueretur. Cic. Leg. i, 22, 58. . . . the pith and 
 force of which precept are so great, that it was attributed 7iot 
 to any mortal 7nan, but to the god of Delphi. The cause still exists : 
 the effect instanced lies in the past, as a subjunctive tribuebatur. 
 
 Quid si magnitudine pecimiae persuasu7n est? Veri si7nile 7i07i 
 est, ut ilk homo tai7i locuples^ ta7n honcstus religio7ii suae 77i07iu7?ie7i- 
 
15 
 
 tisqiie maiorum pecunimn anteponeret Cic. Verr. 4, 6, 1 1. . . //f isn't 
 a 7'easonable thing to suppose that this man, so rich, so honorable, 
 would proceed to put money above religio7i and above the memo- 
 rials of his ancestors. The sentence has followed a video . . . 
 vendituricm non fuisse. Anteponeret is put from the same point 
 of time, namely, in the past, as if we were to say // isn't likely that 
 he would at that time prefer, etc. 
 
 Nimis iracunde hoc quidem et valde intempera7iter : cuius enim 
 malefcii tanta ista poena est, ut dicere in hoc ordine auderet se 
 publicis operis disturbaturum publice ex se7iatus sente7itia aedifi- 
 cata7n domumf Cic. Phil, i, 5, 12. For what wrong-doi7ig 
 deserves such a punishment, that he should venture to say in this 
 body ? etc. The auderet is thought of in the succession of events 
 in the past, while the main question is made general. 
 
 Ve7''i si7nile 7ion est, ut, que77i in secimdis rebus, quern in otio 
 sc7nper secu77i habuisset, himc iii adversis et in eo tu?7tultu que7n 
 ipse comparabat, ab se dimitteret. Cic. Sull. 20, 57. It isn't likely 
 that, after having had his frie7id with him C07istantly in prosperity 
 a7id in quiet ti7nes, he ^o\j\^ pack Jmn off i7i adversity and in a 
 disturba7ice of his own getti7ig-up. The thought of the speaker 
 as he says dimitteret is back at the time of in adversis, etc., to 
 which dimitteret stands related as dimittat would stand related to 
 the present ; while in veri si77iile 71071 est Cicero gives the present 
 look of the matter. 
 
 Ac si nos, id quod ?7iaxi77ie debet, 7iostra patria delectat, cuius 
 rei tanta est vis \ac tanta\ natura ut Ithacavt illa7n in asperri7nis 
 saxulis tamqua77t 7iidulum adfixam sapientissimus vir i7imortali- 
 tati anteponeret, quo amore tande77i . . . Cic. de Or. i, 44, 196. 
 . . . the power of which se7itime7it\s so great that Odysseus i^re^err^d 
 (=z sapie7itissimus vir Ithaca77i illa7n anteponebat : ta7ita est 
 vis, etc.). 
 
 In consecutive ^^/-clauses, then, the secondary tenses of the sub- 
 junctive in themselves express the idea that the act stated in them 
 is put as from a point of view in the past, and are under no 
 law of any Sequence. 
 
 2. In Consec7itive Relative Se7ite7ices. 
 
 Video igitur causas esse permultas quae istimi impellerent. 
 Videa77ius 7iunc ecqtiae facultas suscipie7idi 77ialeficii fuerit. Ubi 
 occisus est Sex, Rosciusf Ro77iae. Quid? tu, Rosci, ubi ttmc 
 eras? Ro7nae. Cic. Rose. Am. 33, 92. I recognize the existence 
 
i6 
 
 of a 7iumber of causes of such a nature as at that time to be 
 pushing him on. In the main sentence the thought is in the 
 present ; in impellerent it is in the past, precisely as in eras at the 
 end of the passage quoted. 
 
 The secondary verb in the consecutive relative sentence, then, 
 conveys in itself the temporal impression which the speaker de- 
 sires to give, and does not accept its tense at the hands of the main 
 verb. 
 
 3. In Causal Sentences. 
 
 Equidem, cum tuis omnibus negotiis interessem, ^nemoria teneo 
 qualis T. Ligarius quaestor urbanus fuerit ergo te et dignitatem 
 tuam. Cic. Lig. 12, 35. Since I was habitually conceriied in all 
 that you did, I have not forgotten how Ligarius treated you, 
 Interessem is simply a subjunctive intereram. 
 
 In the field of the secondary tenses, then, the verb of the causal 
 sentence conveys of itself the desired temporal meaning, and is free 
 of the Sequence of Tenses. 
 
 4. In Concessive Sentences. 
 
 Ilia {epistolci) fuit gravis et plena rerum, quam mihi M. Paccius^ 
 hospes tuus, reddidit. Ad eam rescribam igitur, et hoc quidem 
 primum : Paccio et verbis et re osteyidi, quid tua cojnmendatio 
 ponderis haberet ; itaque in intimis est meis, cum a^itea 7iotus non 
 fuisset. Cic. Att. 4, 16, i. . . . I showed Paccius, alike in word arid 
 171 deed, what iv eight your good opinion carried; arid consequently 
 he is now one of my inti?nate friends, though previously to that 
 we had been strangers. Est lies in the present, while the point of 
 view {ov fuisset is seen in ante a. 
 
 In the field of the secondary tenses, then, the temporal aspect 
 of the speaker's thought in concessive sentences is conveyed di- 
 rectly by the tense employed, and no control is exercised by the 
 preceding verb. 
 
 5. In the Indirect Discourse, etc. 
 
 Laudantur oratores veteres, Crassi illi et Antonii, quod crimina 
 diluere dilucide, quod copiose reorum causas defendere solerent. 
 Cic. Verr. 2, 78, 191. The orators of the old school are praised 
 because it was their way to defend their clients without stintirig 
 time, etc. Solerent is simply an indirect solebant. 
 
 Quern amicum tuom ais fuisse istum, explana inihi^ et qui 
 
17 
 
 cognaium ?ne sibi esse diceret. Ter. Phorm. 380-1. Explain to 
 vie who, according td your story, this friend of yours was, and 
 what manner of relationship with me he used to claim. Diceret is 
 a subjunctive dicebat, and echoes Phor7nios statement in vss. 
 365-6 : Saepe hiterea mihi se7iex narrabat se hu7ic neglegere co- 
 g7iatum suo7n. 
 
 Acta quae essent usque ad a. d. VIII Kal. lunias cognovi ex 
 tuis litteris, Cic. Att. 3, 10, i. /am informed by your letters what 
 had taken />/(2<:^ before and up to May 2^th. 
 
 Quid est aliud de eo referre no7i audere, qui contra se consulem 
 exercitum duceret, nisi se ipsum hostem iudicare f Necesse erat 
 e7iim alterutrum esse hostem; nee poterat aliter de adversariis 
 iudicari ducibus. Cic. Phil. 3, 8, 21. What is the difference 
 betwee7i lacki7ig courage to raise the question in regard to a man 
 who was leading an army agai7ist you, and passing sentence on 
 yourself as a public ene7ny ? For one of the two was, m the nature 
 of things, a public ene7ny ; there was no other possible ivay of 
 regarding generals who were facing each other under arms. The 
 question quid est aliud is put without reference to the special occa- 
 sion (just as in the English), and the verb duceret (as a subjunc- 
 tive ducebat, corresponding exactly to erat following) alone gives 
 the time at which, when he comes to give the special occasion, the 
 speaker's mind is engaged. 
 
 nihil enimfuit clarius ; non quo quisquam aliter putasset, sed 
 nihil de insignibus ad laude7n viris obscure mmtiari solet. Cic. 
 Fam. 3, 1 1, 1. For 7iothing has attracted 77iore attention ; 7iot that 
 anybody had expected a different result, but people never talk in a 
 closet about men of marked positio7i. 
 
 Sed quaero a te cur C. Come Hum no7i defenderem : 7ium lege7n 
 aliquam Cornelius contra auspicia tulerit, etc. Cic. Vatin. 2, 5. 
 /want _y<??/ to answer the questio7i : why was I not to defend Gaius 
 Cor7ielius f Cur non defe7idere77i is, in the dependent form as in 
 the independent, a deliberative question placed at a point in past 
 time. The tense tells its own story (compare it with that of 
 tulerit^, and has an inherent and inalienable meaning of its own, 
 quite distinct from that of any other tense. 
 
 In the field of the secondary tenses, then, dependent verbs in the 
 indirect discourse, etc., of themselves express the desired temporal 
 aspect of the act, and owe their tense to that fact and to no outside 
 influence. 
 
i8 
 
 6. hi Conditions. 
 
 Equidem tibi potissimum velim, si idem ilia vellet. Cic. Att. ii, 
 24, 2. /should like the will to be put into your hands rather than 
 into those of any one else, if her wish were the same. 
 
 In conditions, then, the secondary tenses of the subjunctive are 
 used because they express the idea which the speaker desires to 
 convey, and not because of any influence exerted by the main verb. 
 
 7. In Conclusions, Softened Statements, etc. 
 
 Nemost quern ego nunciam magis cuperem videre quam ie. 
 Ter. Eun. 561. There's nobody whom at the present moment I 
 should rather see tha7i you. 
 
 . . . quia tale sit ut, vel si ignorarent id homines vel si obmutu- 
 issent, sua tamen pulchritudine esset specieque laudabile. Cic. 
 Fin. 2, 15, 49. . . . because it is such that if meii did not know it, 
 or if they had never breathed a word about it, still it would be 
 praiseworthy for its inherent beauty and loveliness, 
 
 Opinor, tuum testimonium, quod in aliena re leve esset, id in 
 tua, quoniam contra te est, gravissimum debet esse. Cic. Quinct. 
 24, 76. Your evidence, which, where another person is concerned 
 would be of light weight, ought, / dare say, to be of great weight 
 in a case that concerns yourself , inasmuch as it is against you. 
 
 Non est credibile, quae sit perfidia iii istis principibus, 2^^ volunt 
 esse et ut essent, si quicquam haberent fidei. Cic. ad Att. 4, 5, i. 
 // is incredible what treachery there is in these leaders as they 
 desire to be, and as they would be, if they could get anybody to 
 t7'ust them. 
 
 In dependent conclusions, etc., then, the secondary tenses of the 
 subjunctive in themselves express the same meaning as in inde- 
 pendent constructions, and owe their use, accordingly, to the fact 
 that they convey that which the speaker desires to say, and not to 
 any influence of the main verb. 
 
 8. hi Final Clauses. 
 
 Explicavi, inquit, sententiam meam, et eo quidem consilio, tuum 
 indicium ut cognoscerem. Cic. Fin. i, 21, 72. / have now 
 developed my views to you, said he, and my purpose in doing so 
 was to get your J7cdgme7it in the ^natter. The act of the main 
 verb is completed at the moment of speaking, and carries with it 
 the idea of the state of affairs now reached (=: habes sententiam 
 
19 
 
 meant), while the purpose of the act operated as an aim from the 
 beginning of the explicatio {id consilium erat ut, etc.). 
 
 Cum ille aut vestra aut sua culpa manserit apud hostem — suas, 
 si metum simulavit, vestra, si periculum est apud vos vera referen- 
 tibus — ego, ne ignoraretis esse aliquas et salutis et pads vobis 
 condiciones, pro vetusto hospitio quod mihi vobiscum est ad vos 
 veni. Liv. 21, 13, 2. While he has stayed in the enemy's camp — 
 whether the fault be his or yours — / have come to you (am here 
 now) ; for I was desirous that you should not overlook the possi- 
 bilities of preservation and peace. F'^/e/ would issue in an adsum ; 
 but to it is attached the hope which existed in the speaker's mind 
 at starting, as well as afterward. 
 
 Ut filius cum ilia habitet apud te, hoc vestru7n consilium fuit. 
 Ter. Phorm. 933-4. Yoii want my son to live with her at your 
 house — that v^ 2.^ your plan. The diOnst fuit goes back to the time 
 of quom repudiiim alter ae remiserim quae dotis tantumdem dabat, 
 a few lines before, while the purpose is put as still entertained. 
 
 Sed senattcs consulta duo iam facta sunt odiosa, quod in consulem 
 facta putantur, Catone et Domitio postulayite, unum, ut aptcd 
 magistratus inquiri liceret, alterum, cuius domi divisor es habitarent, 
 adversus rem public am. Cic. ad Alt. i, 16, 12. But a couple of 
 odious decrees have been passed, which are thought to aim at the 
 consul. . . , one io the effect that an examination before magistrates 
 should be permitted, the other, etc. The ia7n facta simt looks 
 upon the decrees as being now law ; the 2^/-clause looks at the 
 aim with which they were passed. 
 
 " Ut me omnes,'^ inquit, ^^ pater, tuo sanguine ortu7n vere ferrent, 
 provocatus equestria haec spolia capta ex hoste caeso porto." Liv. 
 8, 7, 13. That all men might say with justice, father, that I am of 
 your blood, I bring these spoils taken from the dead body of 7ny 
 challenger. The motive ut ferrent (probably) goes back to the 
 beginning of the act of porto, while the act of porto itself still 
 goes on. 
 
 In final clauses, then, the secondary tenses of the subjunctive 
 are chosen when and because they will express the speaker's 
 meaning, and not because of a Law of Sequence. 
 
 The case for the Law of the Sequence of Tenses seems to be 
 in a curious state. It would be supposed that a law laid down as 
 this is would cover a considerable range of facts. But a detailed 
 examination has shown us, first, that in nearly all the dependent 
 constructions of which the subjunctive is capable, the present, 
 
20 
 
 perfect, and aorist are absolved from the Law, the tense being 
 used (just as if there were no law) because it expresses that which 
 the speaker has, in the particular verb in question, to say; and 
 secondly, that in nearly all the dependent constructions of which 
 the subjunctive is capable the imperfect and pluperfect are likewise 
 absolved from the Law, the tense being used (precisely as if there 
 were no law) because it expresses that which the speaker has, in 
 the particular verb in question, to say ; in short, a detailed examin- 
 ation has shown that in the great mass of the dependent sub- 
 junctive constructions possible to the Roman language, the present, 
 the perfect, the aorist, the imperfect, and the pluperfect are exempt 
 from the Law. But these are all the tenses that the Romans had. 
 Clearly, then, this kingdom ruled by the Sequence of Tenses is 
 under strong suspicion of being a kingdom in dream-land. 
 
 Under this condition of affairs, we shall be obliged, in our second 
 paper, to set up and examine the hypothesis naturally suggested 
 by the negative results thus far reached, namely, the exact oppo- 
 site of the doctrine formulated by Engelmann ; which will then be 
 as follows : ^'A subjunctive clause is, in regard to its tense '^ not 
 " dependent upon the principal sentence ".• in dependent as in 
 independeyit subjunctives^ the tense conveys meanings and owes its 
 .choice to that fact. 
 
11. 
 
 In our former paper, the examination of the received doctrine 
 of the Sequence of Tenses betrayed the fact that the great mass 
 of constructions are exempt from the supposed Law ; and accord- 
 ingly forced us to set up, for examination, the opposite hypothesis, 
 namely, that there exists no such law, — that, in a word, the tense 
 of the dependent subjunctive conveys temporal meaning, precisely 
 as does the tense of an independent subjunctive or indicative. 
 
 To the support of this hypothesis every proof which we have 
 seen militating against the other doctrine will at once repair, and 
 will form a powerful force. The sole question then remains, Can 
 phenomena be produced which the hypothesis cannot fairly 
 explain ? If so, then there exists, as yet, no tenable hypothesis, 
 and we Latinists must go to work to study the facts and reach a 
 sound doctrine on this important and very" practical subject; if 
 there exist no such phenomena, then we may at once label our 
 hypothesis doctrine^ formulate it in a shape suitable for tender 
 minds, and teach it. 
 
 Now, a search over the whole field will be able to produce 
 six objections with which to challenge the hypothesis, namely: 
 
 1. The fact that in the great majority of cases the tenses are such 
 as would be in conformity to the supposed Law of the Sequence ; 
 
 2. The fact that exceptions do not occur in the antequam and dum 
 group of constructions ; 3. The common use of the imperfect in 
 so-called clauses of result after secondary tenses ; 4. The use of a 
 secondary tense of the subjunctive, in connection with a main verb 
 in the past, to express ideas corresponding to facts known to be 
 true at the time of speaking, or even universally true ; 5. The use 
 of a secondary subjunctive, in connection with conditions and con- 
 clusions contrary to fact, to express ideas corresponding to facts 
 known to be true at the time of speaking, or even universally true ; 
 6. The common use of the forms -turtis fuerit^ etc., after primary 
 tenses to represent conclusions contrary to fact. 
 
22 
 
 I. The first objection need detain us but a moment. As has 
 been already suggested in connection with the use of the present 
 in consecutive clauses, the phenomena are what they are because 
 the ideas which they express are such as come oftenest, owing to 
 the constitution of things in this world, into the human brain. 
 To take up, for example, the " exceptions," there are not many 
 present results of remote causes ; while, on the other hand, the 
 only result in the past that can follow a present cause is an his- 
 torical example of a cause universally existing, or, at least, always 
 existing up to the present time. A characterizing present is 
 rarely found in connection with a remote main act, because the 
 facts or tendencies which characterize a person concerned in a 
 past activity are mostly facts or tendencies that were neighbors of 
 the main activity/ When, however, a present fact or tendency 
 can be found that characterizes a person concerned in a past 
 activity, as in the case cited from Pliny the Younger, the Latin, 
 like other languages, simply says in the dependent verb what there 
 is to say. Similarly, the reasons which governed past acts or 
 judgments mostly lie in, and are thought in connection with, the 
 past, and can lie in, or be thought in connection with, the present, 
 only if they are general or habitual facts on which a past act or 
 judgment was based, or past facts on which a present judgment is 
 based; and, in the same way, a present act or judgment is com- 
 monly brought about by a reason near at hand, though that reason 
 may occasionally be an act habitual in the past, etc. The same 
 things are true of acts that stand to each other in an adversative 
 relation. As regards final sentences, the purpose of a remote act 
 must itself have lain in the past, and the purpose of a present act 
 cannot, in strictness, lie in the past ; though a past purpose may 
 be associated, as in Livy's utferre?ity porto, with an activity begun 
 in the past and still going on (cf. John I 31 : but that he should be 
 made manifest unto Israel, therefore am I come baptizing with 
 water), and a past activity may be associated with a purpose ex- 
 pressed as now existing. And, in general, the theory that the 
 Romans used the tenses they did in the subordinate clauses because 
 in each case the particular tense used expressed precisely what 
 they wanted to say, is entirely consistent, as I believe, with the 
 facts, encountering not a single exception in the whole range of 
 Latin literature ; while the theory of an outside determination of 
 
 ^ I am obliged at this point, in default of the explanation of the meanings 
 of the tenses in Part IV, to content myself with a defective phraseology. 
 
23 
 
 the tense of the dependent clause is confronted ^at every turn by- 
 obstacles that cannot be disposed of, to the satisfaction of a thinking 
 mind, by the magic of any such phrase as " exception to the rule." 
 
 2. The second objection likewise need detain us but a moment, 
 for the facts on which it rests have already been explained in the 
 answer to objection i. 
 
 For brevity's sake we will speak of the set of constructions (after 
 antequam, priusqiiam, dum, do7iec, and quoad) as the aniequam set. 
 
 The indicative construction gives us two facts, main and subor- 
 dinate, arranged in the order of their occurrence, and with a clear 
 mark of that order in the shape of the adverb + relative, ante- 
 quam. The subjunctive construction gives us no fact. The act 
 pictured in it may have taken place later, or may never have taken 
 place at all. All that makes no difference whatever. The act 
 itself is simply represented as existing in somebody s brain at a 
 certain time which the narrator has in mind. That time, of course, 
 is the time of the main act. In Liv. i, 26, i : priusquam iiide 
 digrederentur, roganti Metiio ex foedere icto quid imperaret, 
 imperat Tullus, we are not told in the least that Mettius and Tullus 
 went away, but only that they had a going-away in mind, and that, 
 with that departure in vieiv, the one asked, and the other gave, 
 instructions. Digrederentur, then, is not an historical incident 
 noted by the narrator, an event (indicative) serving as a terminus 
 ante quern for the main act, but the thought of Mettius and Tullus ; 
 and, in general, these subjunctive constructions present an act as the 
 thought of some important person in the main sentence, generally 
 the principal actor. Now, if the actor acted in the past, then the 
 thought in the light of which he acted must have been in the past; 
 while if the actor is acting as I speak, his thought must lie in the 
 present, etc. So, then, the very simple reason why the verb of the 
 main clause (act asserted) and the verb of the subordinate clause 
 (act thought) are always of the same order of tenses (both in the 
 past, or both in the present, or both in the future) is that, in the 
 constitution of things, nobody's mind can conceive any other kind 
 of combination ; and what we cannot conceive, we naturally have 
 no occasion to express. Precisely as under i, then, the explanation 
 of the grammatical phenomena is found to be, not a mechanical 
 dependence, but the constitution of this world. And it would be 
 as unreasonable to explain the observed facts by a supposed in- 
 fluence of tense upon tense, as it would be to explain in the same 
 way the fact that we do not find in Latin literature any imperfect 
 
24 
 
 subjunctive of result after a future cause, instead of recognizing 
 that there never has been, nor ever can be, a past result of an act 
 that has not yet taken place. 
 
 3. The common use of the imperfect in clauses of result attached 
 to causes lying in the past. 
 
 In occasional instances, it might be claimed that the imperfect 
 subjunctive has, so far as tense goes, the same force as the imper- 
 fect indicative, portraying e. g. a resulting state of affairs, just as 
 the imperfect indicative does in Greek after wore. But in the great 
 majority of cases in the narrative style there can be no doubt that 
 the independent form of expression of the result would be the 
 aorist indicative, just as in abundant cases in Greek. If, then, the 
 tense tells its own story, why not the aorist subjunctive ? 
 
 It is a pretty question, and all the more so because it is neces- 
 sarily interlinked with another pretty question, namely : in clauses 
 of strict result, which are necessarily clauses of fact, why the 
 subjunctive at all ? Why not the indicative, as in Sanscrit, Greek, 
 German, English, etc., etc. ? What is the history of this purely 
 Roman form of expression ? ^ 
 
 The solution is not difficult. The so-called result-clause is in 
 its origin no clause of resulting fact. It makes no unlimited asser- 
 tion — no assertion that a certain thing has taken place, is taking 
 place, or will take place. The assertion which it makes is a limited 
 one, an assertion that keeps within the limits of the visions of the 
 mind. In Latin such an assertion is expressed by the subjunctive, 
 in Greek by the optative. In Latin the limited nature of the 
 assertion is marked by the mood only ; in Greek it is marked not 
 only by the mood, but also by the use of the admirable litde label 
 av. Where the Latin says scias—yo2i' d know — the Greek says 
 yvolr]^ av. Still, cvcn without the «j/, the Roman is able to make a 
 clear difference between assertions like you'd zindei'stand easily, 
 facilei7itellegas,?ind assertions like you understand easily, facile 
 intellegis ; between assertions like you saw, videbas, and asser- 
 tions \\\i& you'd see, videres. 
 
 Now, the source of all the consecutive sentences, whether after 
 qui or after ut — a perfectly definite and concrete source — is this 
 expression of independent limited assertion.'^ In the coordinating 
 
 ^ From Anglo-Saxon the manuals cite a few examples of the subjunctive in 
 a resulting fact-clause, of which I find no discussion. 
 
 2 1 dissent materially from Dahl (Die lat. Partikel VT, pp. 153-164), finding 
 his treatment too metaphysical, and not resolutely historical. 
 
25 
 
 form you may say, e.g.^ of a given theory j^^z^W easily understand 
 it : it is a very simple thing. The first clause is just as indepen- 
 dent a statement as the second ; it has a construction entirely of 
 its own. So much is evident at the outset. Looking further at its 
 contents, we see that the statement, in point of fact, throws a certain 
 light upon the nature of the thing in question ; it is practically 
 only another way of saying a very simple thing ; it is, in other 
 words, an independent characterizing statefnent. Reverse the 
 order, putting the subjunctive statement last — it is a very simple 
 thing : you'd easily understand it — and it of course remains pre- 
 cisely what it was before, an indepejtdenty an unattached^ charac- 
 terizing statement. 
 
 Wait now for the use of the relative to develop, and then, since 
 the same object of thought occurs in both sentences, tie them 
 together by using which instead of it in the second, and you have, 
 without any change in the grounds of the mood, the dependeyit^ 
 the attached, characterizing statement, common to all languages. 
 
 I could a tale unfold whose lightest word would harrow up thy 
 soul. — Shak. Hamlet, i, 5. 
 
 Who is here so base that would be a bojidman^ — Shak. Jul. 
 Caes. 3, 2. 
 
 Tcoj/ /xei' ovv ras TrarpiKas /SacriXetas' Trapaka^ovToav rls ov< av rovs 'Evayopov 
 
 Kiv8vpovs 7TpoKpii;€uv; (independent characterizing question, corre- 
 sponding to a characterizing predication). OuSe)? ydp ianv oZtco 
 
 paOvfios ocrri^ av de^airo napa rcov npoyovcou rrjv apxrjv ravrrjv TrapaXn^eiv 
 fxaXKov rj KTrjcrdfxevos coarr^p e/cavoy toIs naicrl to7s avrov KaTaXiTreiv ', (de- 
 pendent characterizing predication). Isocr. Evag. 35 : ... who 
 would ?iot rather choose . . . ? There is no one so easy-going that 
 {he) would prefer, etc. 
 
 Tiff ovv ovroas dyaOos rj tls ovTa>s laxvpbs 6? XifiS koi piyei bvvatr av 
 
 fxaxofievos a-rpar^veaOai ; Xen. Cyr. 6, I, 14. IVho is SO brave or who 
 so strong thai {lie') could serve a campaign against hunger and 
 cold f 
 
 Nihil est aeque quod faciam lubens. Ter. Phorm. 565. There 
 is nothing that /should so like to do. 
 
 Vin primum hodie facere quod ego gaudeam, Nausistrata, et 
 quod tuo viro oculi doleant ? Ter. Phorm. 1052-3. Do you want 
 to begin to-day, Nausistrata, by doing something that would delight 
 me^ and would make your husband's eyes smart? 
 
 Cupio videre qui id audeat dicer e. Cic. Phil. 5, 2, 6. I ivant to 
 see the man that would venture to say that. 
 
26 
 
 By a precisely similar development the relative locative uti (ut), 
 which differs from qui only in having a sentence for an antecedent 
 instead of a single word/ introduces a characterizing statement 
 (called a result-clause) of limited predication, precisely as do the 
 corresponding Greek sentence-relative cbs (an ablative) and (uore, 
 and the English that A perfect specimen would be had by trans- 
 posing the two sentences, with ut for a connective, in the following, 
 from Tac. Ann. i, 8i, i : de comitiis consularibus, quae turn 
 primmn illo principe ac deinceps fuere, vix quicquam firmare 
 ausim : adeo diver sa no?i modo apud auctores sed in ipsius orationi-* 
 bus reperiuntur. 
 
 Bpecf)os yap rju tot iv KXvTaijxvrja-Tpas x^potV} ot i^eXcLTTOV p.e\n6pov is Tpoiav 
 i(jt>v, axTT ovK av avTov yva)pl(raifi av ticnScoj/. Eurip. Orest. 377'"9* -^'^ 
 
 was a babe in Clytemnestra^ s arms when, setti7ig off for Troy^ I 
 left 77iy roof, (whereby) so that /should not know him setting eyes 
 on him. 
 
 Quae (occupatioj etsi summa est, tamen nulla esse potest tanta 
 e^/ interrumpat iter amoris riostri. Cic. Att. 4, 2, i. Though my 
 occupations are very pressing, still fione could be so pressing as to 
 (that they would) interrupt the course of our love. 
 
 The Latin, however, alone among languages, extends this wholly 
 logical characterizing construction beyond its original bounds. 
 After Tts- ovTws ev^erjs c(tt\v vfx5)v, e. g,, the Greek distinguishes per- 
 fectly between oo-tis dyvod and oo-n? av dyuool, just as the English 
 distinguishes perfectly between that knows not and that would 7iot 
 kyiow. The Latin, however, fails to make such a distinction as 
 might have been expressed, after quis est ve strum tam stultuSy by 
 qui nesciat and qui nescit. The form of which qui nesciat is a type 
 comes to be practically a phrase of te7idency, of 7iatural direction 
 toward so7ne act, and so conveys very much the same feeling as does 
 the English S2ich as to, in which as is a relative, while the suggestion 
 of direction is conveyed by the preposition to. This may be 
 called the second stage in the history of the construction. Next 
 there intrudes into the idea conveyed by the construction, which 
 ' does not in itself deal with the world of reality, an idea that squints 
 at that world. If I say he is such a man as never to lie, I might 
 as well have said he is a man who never lies, and might, indeed, 
 very easily be quoted as having said that precise thing. In many 
 Latin sentences, in fact, it is impossible to be sure whether limited 
 
 ' 'Uti might be rudely rendered ivhereby ; for the word by likewise begins 
 with expressing a local relation, and then passes into an expression of means. 
 
27 
 
 or unlimited predication is meant. And so the thing- felt and the 
 thing said come to be confused, and the construction of the latter 
 is used to express the idea of the former ; or, in other words, in 
 the Latin language the mood of characterizing predication limited 
 becomes also, in relative sentences, the mood of characterizing 
 predication unlimited. 
 
 So much for the mood. As regards the tense, the present 
 expresses a limited predication of an act thought as future to the 
 time of speaking ; while the tense for limited predication of an 
 act thought as similarly situated relatively to a past time is the 
 so-called imperfect. An example will give a clear idea of this 
 latter use. In this way I should get at the real thing would be, in 
 Tacitean phraseology, hoc modo id incorruptuni sit. The same idea 
 put interrogatively would be q^ionam modo id inc or r upturn sit. Let 
 time pass on, and then state a past question of this sort, and you 
 have quonam modo id incorruptum foret, as Tacitus uses it in Ann. 
 2, 12, 3, In dependence upon agitabat {igitur propinquo summae 
 rei discrimine explorandos militum animos ratus, quonam id modo 
 incorruptum foret secum agitabat^. 
 
 Inasmuch as to carry a limited statement back into the past Is 
 practically to quote It, which requires the use of the infinitive, it is 
 not strange that we but rarely find examples of this imperfect sub- 
 junctive in the independent declarative form. A remarkable in- 
 stance, however, is to be found In the oratio obliqua in Caes. B. C. 
 3) 73 • contionem apud milites habuit . . . dandam omnibus operam 
 ut acceptum incommodum virtute sarciretur ; quod si esset factum 
 (future condition), detrimentum in bonum verteret (future conclu- 
 sion).^ 
 
 Now, the constructions nihil est aeque quod faciam lubens 
 {there Is nothing that I should so gladly do') become, when stated 
 again for the original situation after some time has elapsed, nihil 
 erat aeque quod facerem lubens (to translate into unfamiliar but 
 intelligible English, there was at that time nothing that I would at 
 that time so gladly do). Faciam and facerem both express 
 limited predication, each from its standpoint, and the sole differ- 
 
 * Other examples are probably /r^T/fr^r^/ and citaret in Hor. Sat. i, 3, 6 and 7. 
 It is a very illustrative fact, furthermore, that in English the corresponding 
 forms of limited predication from a past standpoint (namely, the auxiliaries 
 would, should, etc.) are the regular forms of expression in indirect quotations, 
 as, e. g., in the translation of verteret above. The construction, of course, plays 
 a large part in both languages, in a secondary stage, as the means of expression 
 for conclusions contrary to fact. 
 
28 
 
 ence lies in that standpoint. The facial of to-day becomes the 
 faceret of to-morrow's retrospect. So far, all is strictly logical. 
 And when the construction of the mood extends itself in Latin, 
 the tense and the mood go together, the former still carrying the 
 idea oi connection with a past time ; and the tense is no more the 
 product of a subtile influence exerted by the tense of the main 
 verb than is the mood the product of a subtile influence exerted 
 by the mood of the main verb.^ 
 
 4. The use of a secondary tense of the subjunctive, in connection 
 with a main verb in the past, to express ideas corresponding to 
 facts known to be true at the time of speaking, or even universally 
 true. 
 
 A universal fact may be regarded with reference to its bearing 
 upon some present act or judgment, or with reference to its bearing 
 upon some past act or judgment. In the first case it is a universal 
 truth put as now applicable^ in the second a universal truth put as 
 then applicable. That which tells whether the point of view from 
 which the universal truth is applied is that of a present judgment 
 or that of a past judgment, is the tense. Its power is seen clearly 
 in independent sentences in English ; e.g. the sentence Tyra7iny is 
 never right — the orator sdijs — eve7i if it be the tyranny of a majority 
 over a minority^ becomes, as we speak of the same utterance later, 
 Tyra7iny was never right — the orator said — even if it were the 
 tyranny of a majority over a minority? It is not even necessary, 
 in such a use, to have any word indicating an indirectness of state- 
 ment; e. g. in Guizot's Earth and Man (preface, p. vi) I find this 
 sentence : Numerous quotations and references were incompatible 
 with the form of these discourses. They remain incompatible ; 
 but the point is not the general incompatibility as recognized at 
 the time of writing the preface, but the incompatibility as recog- 
 nized and acted upon at the time of the writing of the discourses. 
 The matter may then be briefly stated as follows : general or 
 lasting facts may be put, in their larger aspect, in the general 
 present or so-called logical perfect, or, in their aspect as bearing 
 
 ^ The mood in the resulting fact-clause never freed itself from its illegiti- 
 mate origin, never became the indicative; but the tense in these clauses did 
 after a while — first in Cicero's time — partly free itself from its illegitimate 
 origin, and frequently appears, in appropriate places, as the aorist. In the 
 main, however, the old habit continues, and the aorist therefore always had the 
 power of catching sharply the attention. 
 
 *0n this point, and others connected with it, Otto Behaghel's Die Zeitfolge 
 der abhangigen Rede im Deutschen is very helpful. 
 
29 
 
 upon some past act at the time of which they Hkewise existed, in 
 the imperfect or logical pluperfect. 
 
 The same phenomena occur in Latin in the indicative in subor- 
 dinate clauses, as in Cic. Fam. 5, 2, 9 : sedtamen fieri non moleste 
 tuli atque etiam^ ut ita fieret, pro me a parte adiuvi, ut senati con- 
 sulto meus inimicus^ quia tuus /rater ersit, sud lev are tur (the lasting 
 fact, tzms f rater est, furnishes a ground for action on the past 
 occasion mentioned ; for which time, of course, the statement must 
 be, not est, but erat). A comparison oi pertinerent and the pre- 
 cisely parallel pertinebant in the two sentences following will show 
 what the feeling of the tense of the subjunctive is : Cic. Tusc. 1,1, 
 I : ...(?/, cum omnium artium, quae ad rectam vivendi viam per- 
 tinerent, ratio et disciplina studio sapientiae, quae philosophia dici- 
 tur, contineretur, hoc mihi Latinis litteris inlustrandu7n putavi . . . ; 
 De Or. 3, 19, 72: Namque, ut ajite dixi, veteres illi usqtie ad 
 Socratem o^nnem omnium rerum, quae ad mores hominum, quae 
 ad vitam, quae ad virtutem, quae ad rem publicam pertinebant, 
 cognitionem et scientiam cum dicendi ratione iungebant. 
 
 One cannot, therefore, believe in a mechanical and unfeeling use 
 of the subjunctive in these cases, unless he is prepared also to 
 believe in a mechanical and unfeeling use of the indicative in 
 similar sentences, including independent sentences in modern lan- 
 guages. 
 
 5. The use of a secondary tense of the subjunctive, in dependence 
 upon conditions and conclusions contrary to fact, to express ideas 
 corresponding to facts known to be true at the time of speaking, 
 or even universally true. 
 
 In complex sentences made up of a main sentence with subjunc- 
 tive verb and one or more subordinate sentences, the modal feeling 
 in the speaker's mind which expresses itself in the main sentence 
 is, in the nature of things, very likely to continue in the speaker's 
 mind in the subordinated sentence or sentences, either quite un- 
 changed or but slightly shaded. If, for example, I say in Latin, 
 Let him send whom he will, mittat quern velit, the mood in velit is 
 not a case of ''attraction" or "assimilation" at all. Velit is as 
 much a jussive as mittat is. The meaning is, Let him choose his man, 
 and send that man, or, in older English, choose he his man and send 
 him. In sei ques esent quei sibei deicerent necesus ese Bacanal 
 habere (C. I. 196), the deicerent is as much a future condition 
 (= sei ques deicerent^ as esent is. In the sentence in Mr. Howells's 
 Lemuel Barker (cap. 23), If a person heard afterwards, when I 
 
30 
 
 had made out somethijig, if I ever did, that I had been a servant^ 
 would they despise 7iie for it ?, the had made out is as much an 
 ideal state of affairs in the future as is the main condition, heard ; 
 and the Latin-speaking man would, of course, use in both of them 
 the same mood, with an unchanged feeling. But he would also, 
 of course, use a tense of the same set — not because he had used a 
 primary tense in the main sentence, but because the feeling which 
 he has to express when he gets to the second verbal idea requires 
 the same kind of a tense to convey it. Tense and mood are here 
 inseparable. Precisely like the case from Mr. Howells is Lucre- 
 tius's nee, si materiam nostrum collegerit aetas . . . pertineat quic- 
 quam tamen ad nos id quo que factum, interrupta seme I cum sit 
 repetentia nostri (3, 847-851), though Munro, with a less delicate 
 feeling than a Roman's, translates by has instead of had. Simi- 
 larly in Cicero's quod scribere, praesertim cum de philosophia 
 scriberem, non auderem, nisi idem p\3.ceret gravissimo Stoic or 21m 
 Panaetio (Off. 2, 14, 51), cum scriberem does not mean especially 
 710W that I am writing, but especially if I were writing (/should 
 7iot venture to write this, especially if writing, as now, about 
 philosophy, were // not that Payiaetius takes the same view'), as 
 Madvig recognizes (Gramm., §383, 2), though the translation fails 
 to convey what he points out. 
 
 Now, this same delicacy of feeling appears to me palpably to 
 obtain in a great many cases where we find, attached to a condition 
 or conclusion contrary to fact, a subordinate clause the contents of 
 which are known to correspond to objective reality. I do not feel 
 that in these cases the Roman verb predicates objective reality at 
 all, but rather that the thought is colored by the ideal complexion of 
 the whole feeling. InCic. N. D. i, 17, 45: si nihil aliud qn^i^re- 
 remus 7iisi 2it deos pie coleremus et 2it superstitioiie liberaremur, 
 satis erat dictum ; nam et praestaiis deorum natura hominum 
 pietate coleretur, cum et aeterna esset et beatissima . . . et . . ., the 
 aim in coleremus, whatever may be the objective facts in regard to 
 our habits of worship, is in this case an inseparable part of the 
 unreal condition si nihil alitid quaereremus ; and, in precisely the 
 same way, the cum aeterna esset is not a general ground asserted 
 as having a present bearing, but a general ground recognized as 
 one that would bear upon this ideal case (= in that case the sur- 
 passing nature of the gods would receive the pious ivorship of 
 mankind, being — still in that case — recognized as eternal, etc.). 
 
 The same thing is true, though with a still finer shade of meaning, 
 
31 
 
 in Cic. Inv. I, 2, 3 : . . . qui iandem fieri pottdt^ nisi homines ea 
 quae ratione invenlssent, eloquentia persuadere potuissent . . . ; 
 how could all this have taken place ^ had not rnen — supposing them 
 to have made discoveries — also had the gift of commending them 
 by fair words f Quae invenissent is not an independent assertion, 
 though such an assertion might, of course, be made, but an 
 assumption forming a part of an ideal sum total. 
 
 So far under this head, I trust my readers are still with me, and 
 are disposed, after these examples, to look for a modal feeling, 
 rather than an entire absence of both modal and temporal feeling, 
 in constructions of this sort in general ; recognizing, too, that our 
 own language is less fine in expression — which means that our 
 feeling itself is less fine — than that of the Roman, as is exemplified 
 by Munro's " has " in the rendering of the passage cited from 
 Lucretius. This being so, we will now examine a case of the same 
 kind, presenting as great a difficulty as can be summoned up, the 
 passage Cic. Tusc. i, 5, 9 : nam si solos eos diceres miseros quibus 
 moriendum esset, nemiiiem tu quidem eorum qui viverent exci- 
 peres. A rough rendering would be : For if you applied the riame 
 wretched to such alone as were doomed to die,yo7i could not make 
 an exception for such as breathed the breath of life, no, not for 
 one. Even in the English translation, I cannot feel that the verbs 
 were dooined and breathed are merely perverted assertions — verbs 
 not only tenseless but modeless — but rather that, though corre- 
 sponding to facts which every one knows, they are here set up in 
 the imagination as an integral and indivisible part of the sum 
 total of the ideal condition and conclusion ; so that it would be a 
 fair rendering of the feeling, though a bulky one, to translate as 
 follows : Supposing there were people doomed to die — and we know 
 that all men are — and supposing you called those people, and no 
 others, unhappy, then, assuming the existence of living people in 
 this world — a safe assumptio7i—you would have to call every 
 soul of them by the same word, unhappy. 
 
 The cases that fall under this head, then, are not specimens of a 
 mechanical adaptation of outward form, but of a very subtile and 
 delicate modal feeling, existing consistently alike in the main idea 
 and in subordinate ideas that form an integral part of it.^ 
 
 ' Upon this point really turns the whole battle. But that battle is no longer 
 for the saving of the Sequence of Tenses in toto; it is for the saving only of a 
 little territory covered by a part of the examples under this one head. If there 
 are any who believe that this modal feeling does not exist, and that the use of 
 
32 
 
 6. The sole remaining hope of the doctrine lies in the use of the 
 periphrastic form -tunes fuerit, etc., in certain of those cases in 
 which a conclusion contrary to fact is put as dependent — in certain 
 of them only, be it observed ; for we have to begin at once to make 
 inroads even upon this petty territory. Firstly, the matter touches 
 only conclusions contrary to fact in past time ; conclusions contrary 
 to fact in present time remain their simple selves and ignore the so- 
 called sequence (A. J. P. VII, p. 463). Secondly, the rule is almost 
 constantly violated after secondary tenses (where under its sway 
 the tense ought to remain unaffected) by the change of a pluper- 
 fect to the historical perfect fuerit with the future participle, etc. 
 The corresponding indicative form, the modus operandi of which 
 we need not recall, is in use in the independent construction, but, 
 be it observed, is there the less common construction, while in the 
 dependent form it is used in ninety-nine cases out of a hundred. 
 Now, this use^ marks a distinct preference of the Romans, and a 
 preference all the more striking because it goes against the sup- 
 
 the tense in some of these cases is purely mechanical, then they should feel 
 that we have come at last to a class of phenomena to which the doctrine of the 
 temporal expressiveness of all subjunctives does not apply. But it by no means 
 follows that they are thereby justified in holding the doctrine of the half-tense- 
 lessness (I say half-tenselessness because it is universally granted that the 
 dependent subjunctive tense retains one of the two powers of the independent 
 tense, that of distinguishing between actions complete and actions without 
 reference to completeness) of all dependent subjunctives. They should in that 
 case hold that, as there is undoubtedly a point in the stylistic development of 
 the language at which the subjunctive mood is used without modal meaning 
 (the final stylistic outcome of the common natural unity of modal feeling in a 
 succession of verbs attached to a subjunctive) — though not commonly so used — 
 in the same way, there is a point at which the tense also carries no meaning — 
 though not commonly so used. And, at the very least, it is clear, after the ex- 
 hibit in the previous paper of the great range of the " exceptions," that one who 
 believes in the Law of Sequence should believe in it as a law which nobody is 
 bound to obey — a law which, whatever it may do, never trammels a speaker or 
 writer ; for, even in the set of cases just now under discussion, a writer is perfectly 
 free to break with the supposed Law, as Cicero does in the sentence already cited 
 from Fam. 13, 6a, 4: quae quantum in provincia valeant,vellem expertus essem, 
 sed tamen suspicor. Even to add to the statement of the Law in the grammars 
 so much of a concession as this would save teachers in the preparatory schools 
 from difficulties (see page 71, below) which I have been told they now 
 experience. 
 
 J It is not quite universal ; see Livy, 2, 33, 10 : Tantumque sua laude obstitit 
 famae consulis Marcius, ut, nisi foedus cum Latinis columna aenea insctdptum 
 monumento esset, ab Sp. Cassia una, quia collega afuerat^ ictum, Postumutn Co- 
 minium bellum gessisse cum Volscis mevioria cessisset. 
 
33 
 
 posed rule.^ The feeling that probably underlies it we will not 
 take time to discuss. 
 
 We have now narrowed the ground to absolutely the last 
 phenomenon that can be claimed for a Sequence of Tenses, namely, 
 the use in conclusions contrary to fact after primary tenses only 
 {the supposed rule bei?ig nearly always violated after secondary')^ 
 and in forms referring to the past only {the supposed rule beijig 
 alzvays violated in forms referring to the present) of the peri- 
 phrastic -turus fuerit, etc., instead of the pluperfect subjunctive. 
 The amount of evidence for the Sequence of Tenses which this use 
 affords, seen in the light of the contradiction of that evidence on 
 its own territory in the habitual violation of the Sequence after 
 secondary tenses, and the universal violation of it when the con- 
 clusion refers to present time,^ and seen, further, in the light of the 
 enormous evidence presented on the other side by phenomenon 
 after phenomenon in construction after construction, is so very 
 small, that even if its use were without an exception, one would 
 not be too bold who should consider the probability to be over- 
 whelming that the ground of the use lay in a special liking for the 
 subordinated form of the future participle with fuerit (a prefer- 
 ence habitual, as we have seen, in despite of the supposed law, after 
 secondary tenses, and possibly reinforced by the constant use of a 
 similar infinitive form in main statements in indirect discourse), 
 and not in any supposed law. Biit, in point of fact, we find that 
 even in this last little shred of territory which our examination has 
 not yet stripped from the kingdom of the Sequence of Tenses, 
 Cicero finds the pluperfect subjunctive entirely competent to ex- 
 press the temporal idea he wants to convey, as in pro Sest. 29, 
 62 : quod ilk si repudiasset, dubitatis qui^i ei vis esset adlata "^ 
 Brut. 35, 126: quam ille facile tali ingenio^ diutius si vixisset^ 
 vel paternam esset vel avitam gloriam consecutus ! eloquentia 
 quidem nescio an habuisset /^r^w ne7ninem ; ibid. 41, 151 : atque 
 haud scio an par principibus esse potuisset ; after which examples 
 
 ' Even where the tense demanded by the theory of the sequence is found, 
 yet the preference is for the periphrastic form, as in Liv. 10, 45, 3 : Subibat 
 cogitatio aninium quonam modo tolerabilis futura Etruria fuisset, si quid in 
 Santnio adversi evenisset ; 2?i,ii\yi\ apparuitqtie quantum excitatura mo tern vera 
 fuisset clades ; and so frequently. 
 
 ^ To say that the imperfect has to be retained because the present would 
 express something different, namely, a future conclusion, is to attribute to the 
 tenses of the subjunctive that power of expressing temporal relations which it 
 is the purpose of this paper to claim for them. 
 
34 
 
 it needs no boldness whatever to say, as I now do without reserve, 
 that the tenses of the Latin subjunctive, alike in dependent arid in 
 independent sentences, tell their own temporal story — that no such 
 thing as is meant by the doctrine of the Sequence of Tenses exists. 
 But the conviction thus reached of the non-existence of an out- 
 side power controUing the tense of a subordinate verb is seen to be 
 not merely true, but pleasingly true, the most natural and probable 
 thing in the world, when we recall — what is now become a com- 
 monplace of the grammarians — that nearly all the dependent con- 
 structions were once simply independent constructions having 
 neighbors related in the speaker's mind but not in formal expres- 
 sion ; and that they then, as all agree, were able to tell their own 
 temporal story. For example, the paratactical Quid agerem? 
 Nesciebayn (What was I to do ? I had no idea) becomes Quid 
 agerem, nesciebam (What I was to do, I had no idea). Can any 
 one, then, seriously suppose that a Roman, using the imperfect in 
 the side-by-side construction Quid agerem ? Nesciebam because it 
 expressed his meaning, would in the composite sentence use the 
 imperfect Quid agerem because he used a secondary tense in the 
 main clause ? In the paratactical form quid agerem expressed 
 that which he had to say, while quid agam would have expressed 
 an entirely different thing, which he didn't at all want to say. 
 Now, in the composite Quid agerem, nesciebam, is it not s# 
 obvious that it is hardly conceivable that there should ever have 
 been any occasion for a paper on the subject, that quid agerem 
 was said, and not quid agam, because quid agerem expressed 
 what the speaker had to say, while quid agam would have 
 expressed an entirely different thing ? It is, in point of fact, not 
 *credible that a sweeping doctrine like that of the tenselessness of 
 all dependent subjunctives could ever have come into general 
 acceptance if it had been broached to a generation that had inter- 
 ested itself in the natural history of the subordinate constructions. 
 
 III. 
 
 The selection of examples given in the first paper showed that 
 any combination of temporal ideas (main and subordinate) that 
 may possibly enter into a healthy brain is capable of expression 
 in Latin (as would be expected), and that, when the combination is 
 an unusual one, the subordinate verb is, alone and by itself, expres- 
 sive of temporal relations as fully as an independent verb would 
 be. To this position no denial is possible. 
 
35 
 
 The highly probable explanation of the whole field of phenomena, 
 usual and unusual alike, was, as we saw, that the power indubitably 
 found to be exercised by a given tense in a given construction in 
 an unusual combination was inherent in the tense, in whatsoever 
 combination, usual or unusual. Where, as in the present case, a 
 cause the existence of which is absolutely proved will account for 
 all the phenomena observed, it is bad science to assume the exist- 
 ence of a second and entirely different cause. Further than this, 
 it was shown that an absurd conclusion would follow from the 
 adoption of the theory of a second cause in these phenomena, viz. 
 that the present subjunctive in the result-clause is incapable of 
 expressing the present result of a past activity immediately upon 
 the conclusion of that activity, but is capable of expressing the 
 present result of the very same activity ten years later. 
 
 In the present paper it has also been shown that the objections 
 which might be brought against the theory that the subjunctive 
 has temporal expressiveness disappear under examination. The 
 case would seem to me to be made out, if I were to stop here. 
 Nevertheless, I can conceive that a doubt in regard to one point 
 may remain in the minds of some of those who have held the doc- 
 trine of the Sequence of Tenses, and that a subtilized form of the 
 doctrine may throw up an intrenchment in their minds upon this 
 point ; and I desire to show that no possible point of intrenchment 
 exists. 
 
 We are obliged (they may say) to feel the living force of the 
 tense in the unusual constructions, but we do not feel it in the 
 usual constructions. The tense must clearly originally have had, 
 in at least nearly every case, the power which we see it displaying 
 in unusual combinations ; but in the usual combinations it seems 
 to us to have become a mere form — not a living tense, but a 
 speech-type. 
 
 If the doctrine of the Sequence had not been in possession of 
 the field for many years, the burden of proof would not fall 
 upon an opponent of the doctrine, but upon its supporters. There 
 are very strong antecedent objections against any form whatever 
 in which it may be held. Let us consider what the effect of sub- 
 ordination is upon tense, starting from the Sprachgefiihl of a 
 modern language. No one has a right to object to such a method, 
 for, even though the modern Sprachgefiihl be a dangerous tool to 
 handle, it is absolutely the only one that Heaven has vouchsafed 
 us. Compare now the independent deliberatives in English and 
 
36 
 
 Latin. W/ia^ am I to do is quid agam, what was I to do is quid 
 agerem. The sole difference between the first set and the second 
 hes in the time at which the question is placed. Quid agerem and 
 quid agam differ precisely as am, and was differ, and in no other 
 wise. Whatever be the history of the tenses of the Latin subjunc- 
 tive in practical use, the idea of the present lies in quid agam, as 
 fully and strongly as in what am I to do, the idea of the past in 
 quid agerem as fully and strongly as in what was / to do. Now 
 let these sentences be attached to others, and we have, e. g., I 
 don't know what I am to do, I didri't know what I was to do, — quid 
 agam nescio, quid agerem nesciebam. Has anything happened to 
 destroy the activity of the tenses of agam and agerem f The 
 absolutely identical English construction has kept its full temporal 
 power ; the am, the was have not lost their meaning. But the 
 am, the was are a part of the very nature of quid agam and quid 
 agerem,. What ground is there for supposing that constructions 
 absolutely identical in two languages, passing through absolutely 
 the same experience, should suffer absolutely opposite fates? 
 And how, there being no conceivable ground whatever for such a 
 belief, can one nevertheless swear 'tis so, when he discovers and 
 admits that this very same dependent quid agerem does retain its 
 full force, precisely as does the English what was I to do, when 
 found after a present : — IdiSkyou what /was to do, quaero a tequid 
 agerem ? The last thing to be expected is that these fully ex- 
 pressive tenses will ever become, participial-like, half tenseless.* 
 And the moment they are found with undeniably full temporal 
 meaning in any construction of the same class, it is sound logic, 
 and the only sound logic, to suppose that they have nowhere lost 
 their temporal expressiveness. 
 
 This applies fully and without reservation to a construction that 
 has remained unchanged in nature, like the deliberative. But it 
 also applies with equal force to constructions that have suffered 
 
 * The fact is that the original force of the tenses in this and that construction 
 must have been constantly preserved to the Roman mind, as it should be to 
 that of the modern reader, by certain related constructions. E. g. the use of 
 the jussive subjunctive without introductory particle in the oratio obliqiia would 
 keep fresh the temporal expressiveness of the verb in those subordinated jussive 
 forms which we call final qui- and «//-clauses ; the frequent collocation of the 
 direct jussive and the corresponding dependent interrogative form (deliberative) 
 would keep fresh the temporal expressiveness of the verb of the latter, as in 
 Plant. Merc. 624-5 (an " exception " to the Law of the Sequence, by the way) : 
 Quid ego facerem ? — Quid tu faceres, men rogas? Requireres, rogitares . . .; etc. 
 
37 
 
 some change of nature, unless there is distinct proof that this 
 change has taken place in the temporal power itself. The jussive, 
 e. g., has a future force, the present placing the commanding as 
 thought at the present moment, the imperfect as thought at a past 
 moment. The construction, once independent, in time becomes 
 very closely subordinated, but the future force is precisely that 
 part of the original jussive force which remains unimpaired. And 
 so the argument might be carried on through all the constructions 
 of subordination. It could be shown that the original temporal 
 force remained unimpaired in every construction except that of a 
 part of the consecutive clauses ; and even here it would be found 
 that, though the present had changed its force, and the imperfect 
 had changed its force, yet they had held to their power of saying, 
 the one in connection with this present, and the other in connection 
 with that past time ; and that each was ready to tell its individual 
 story in any company of main verbs whatsoever.^ 
 
 But, happily, we are not dependent upon antecedent grounds, 
 strong though they are. This subtile doctrine that the subordinate 
 tense is at one moment living, at another lifeless, even if it had 
 probabilities on its side, could be confronted with entirely suffi- 
 cient indications of its unsoundness. Of these indications, some 
 are themselves subtile, others very palpable. 
 
 I. The historical present puts a past, perhaps a very remote, act 
 as if it were going on before the eyes of us, the readers. It is as 
 
 ^ There can be no doubt that the tenselessness of many of our modern 
 idioms in subordinate clauses has done much to blunt our sensitiveness for the 
 temporal expressiveness of the corresponding Latin idioms. For the relative 
 final clause in Plaut. Trin. 740-41 : non temere dicant te benignum virgini : 
 datam tibi dotem ei quam darei eius a patre^ our common phrase would be to give 
 {they'd say a dowry had been given to you, to give in turn to her, as from her 
 father), a form which conveys no idea of the place of the plan in respect of the 
 time of its formati6n. In the Latin, however, the form is precisely as in 
 daretis and quaereret in Ter. Phorm. 296-7: non f nit necesse habere : sed id quod 
 lex iubet, dotem daretis ; quaereret alium virum. It wasn't necessary to take her 
 to wife : the thing for you to do, as the law enjoins, was to give her a dowry, and 
 the thing for her to do was to hunt up somebody else for a husband (you were to give, 
 she was to hunt up — ex post facto commands). If, now, we were to translate 
 the subordinate jussive clause (so-called final) quam dares in the Trinummus 
 by the same formula as in the independent jussive daretis in the Phormio 
 {they'd say a dozury had been given you, which you were to give to her), and if 
 we similarly everywhere used, in Roman fashion, precisely the same English 
 form for a given dependent construction and its independent form, the idea 
 would never have been tolerated that the dependent subjunctives in Latin are 
 void of temporal meaning. 
 
38 
 
 if we sat in the theatre and saw the things of long ago done upon 
 the stage. In Livy's story of what followed the death of Lucretia, 
 we first hear the solemn oath of Brutus, " By this once holy blood 
 I swear to pursue the whole brood of the Tarquins." We see him 
 hand the knife to Collatinus, to Lucretius, to Valerius. We see 
 them repeat his oath. Before our eyes they carry the body to the 
 forum. We see the gathering of men, their lamentations, and 
 their growing fury. It is not history that is given us, it is the 
 mimic stage. Now, these stage-presents are followed in a depend- 
 ent clause (say a final clause) now by a primary tense, now by a 
 secondary. What is it that tells us, as we read, whether we are to 
 keep up the fiction of the theatre, and wait to see the act of the 
 final clause, say the intended blow of a murderer, actually per- 
 formed upon the stage, or are to drop the illusion, and return to 
 the fact of sober narrative, namely, that this was once upon a time 
 a purpose ? It is nothing but the verb of the final clause itself. 
 In that verb, and in no other, lie. or do not lie, the directions. The 
 choice of the subordinate verb is itself just as perfect and complete 
 a method of communication between writer and reader as is the 
 choice, for the main verb, between the sober aorist and the stagy 
 present.^ 
 
 2. The Roman has but one word for the aorist and the present 
 perfect. As we read a complex sentence having for its main verb 
 this defectively expressive form, what is it that tells us whether the 
 writer thought aorist or thought present perfect ? It is the tense 
 of the dependent subjunctive. However it may have come to its 
 meaning, it is gifted with power to tell us the very nature of the 
 main verb. Here, then, the tense is clearly living. 
 
 3. The present perfect is capable, whife remaining its true self, 
 of being associated with either primary or, as in the final clauses 
 given on pp. 463-4, with secondary tenses. What is it that tells us 
 in such sentences whether the speaker puts his purpose as now 
 entertained, or as entertained (say) at the beginning of the action ? 
 It is, not the inflexible main verb, but the flexible verb of the sub- 
 ordinate clause. Here, then, the tense is clearly living. 
 
 4. In impassioned language the present infinitive is often used 
 in exclamations, even though the act or state thought of lies in 
 the past, as in Ter. Hec. 532 : Adeone pervicaci esse animo^ ut 
 
 ^ To say, as Roby does (and others in differing phrases), that " the historical 
 present is, in its effect on the verbs directly or indirectly dependent on it, 
 sometimes regarded as a primary, sometimes as a secondary tense," is to con- 
 tent oneself with words. 
 
39 
 
 puerum praeoptares perire. The idea of your being so obstinate 
 that you preferred that the boy should die / (the tense oi praeoptares 
 as distinctly tells us that the act lies in the past, as does the tense 
 of preferred in the translation) ; Cic. Sull. 20, 57 : iam vero illud 
 quam iucredibile, quam absurdum, qui Romae caedem facere, qui 
 hanc urbem inflammare vellet, eum familiarissimum suuni 
 dimittere ab se et amandare in ultimas terras ! Then too how 
 incredible y how absurd^ the idea of his being dismissed and packed 
 off to the end of the world by the man who wanted to butcher 
 people in Rome, who wanted to set this city on fire I 
 
 Here, again, it is not the main verb, but the subordinate verb 
 that tells the temporal story. The speaker relies wholly upon the 
 subordinate verb for the conveying of the time of the whole 
 sentence. 
 
 5. But the case is even stronger than this. For the number of 
 sentences in Latin is very great in which there is no main verb 
 whatever, and the entire burden of the expression of time falls upon 
 the subordinate verb, as in Ten Phorm. 364-7 : Saepe interea 
 mihi senex narrabat se hunc neclegere cognatum suom. At quern 
 virum ! quern ego viderim in vita optumurn. The old man used 
 now and then to tell me that this relative of his was treating him 
 shabbily. But what a ma7i ! the best I have seen in all my life ; 
 Juv. 157-8: O qualis fades et qziali digna tabella cu7n Gaetula 
 duceni portaret belua luscum. What a sight, what a subject for a 
 painting whe?t the monster from Gaetulia was carrying on his back 
 the great general — minus one eye; Cic. Quinct. 26, 80: O homi- 
 nem foriunatum, qui (see how we wait for the verb to give us our 
 temporal conception) eius modi nuntios sen potius Pegasos habeat ! 
 O happy man, that has such messengers or rather winged horses ! 
 In Cic. pro Arch. 10, 24 : O forturiate, inquit, adulescens, qui 
 tuae virtutis Homerum praeconem inveneris, the subjunctive 
 inveneris conveys the temporal idea for the whole sentence as 
 perfectly as does the indicative attulisti in Cic. Flacc. 40, 102 : O 
 nox ilia, quae paene aeternas huic urbi tenebras attulisti, and the 
 indicative viroKplveTaL in Aristoph. Acharn. 400-1 : 'G rpia-fiaKapt 
 
 'EvpiTriBrj, 06 6 bovkos ovrcoal (ro(f)S)S vTroKpii/erai. So, then, it IS not 
 
 necessary, in order that the tense should carry to the mind a dis- 
 tinct temporal meaning, that it should follow a verb in whose 
 company one is surprised to find it. The subordinate verb is 
 capable, not only of piecing out the defective temporal expression 
 of the main verb, as under 4 above, but even of getting along 
 
40 
 
 entirely without it — of doing the entire work of temporal expres- 
 sion for the whole sentence. 
 
 6. A subordinate verb which, on the theory of the Sequence, 
 accepts its tense from the main verb, nevertheless, upon that same 
 theory, may, and mostly does, force its own dependent verb 
 according to its will, and, breaking it off from all dependence upon 
 the main verb, dictate to it what its tense shall be ; in other words, 
 to use a homely but scientifically exact phrase, it is only "playing 
 dead," as in Cic. Ros. Am. 14, 141 : Quaeramus quae tanta vitia 
 fuerint in unico filio^ quare /^/^/r/displiceret. Let us inquire what 
 great faults there were in this only son, that would make him 
 obnoxious to his father. 
 
 To grant to the subordinate tense the power of expressing in 
 usual combinations the same meaning that it expresses in unusual 
 combinations is a much easier postulate than to refuse to it the 
 power to express meaning in itself, while conceding to it the power 
 to dictate to another subjunctive what its tense shall be. 
 
 7. If we can find some indicative construction which, in passing 
 into the subjunctive in the indirect discourse, would need to change 
 its tense if there is a Law of the Sequence, we can get an absolute 
 settlement of the whole question by watching its behavior. Such 
 a construction is to be found in the common use of the aorist in 
 temporal clauses introduced by ubi, ut, postquam and siniul atque, 
 as in Cic. Fam. 5, 2, 4 : Postea vero quam profectus es, velim recor- 
 derCy quae ego de te in senatu egerim, quae in contionibus dixerim, 
 quas ad te litteras miserim. 
 
 When such a clause is thrown into the indirect discourse and 
 made dependent upon a past tense, then, if the theory is true 
 that dependent verbs have no temporal expressiveness, the depend- 
 ent verb which we are watching will go into the pluperfect sub- 
 junctive, losing its peculiar individuality of expression ; whereas if 
 the theory is true that the dependent verb has an unimpaired 
 power of temporal expression in and of itself, our dependent verb 
 will be found doing in the indirect discourse precisely what it did 
 in the direct discourse, unchanged in tense, affected in no respect 
 whatever except that of mood. But everybody knows that, while 
 in perhaps one case in ten the pluperfect is found, just as it 
 is in the independent construction (cf. Liv. 43, 6, 8 : hoc etiam 
 Lampsace?ii, octoginta pondo coronam adferentes petebant, com- 
 memorantes discessisse se a Perseo, postquam Romanus exercitus 
 in Macedoniam venisset, with Liv. 44, 25, 9 : iibi ad pecuniae 
 
41 
 
 mentio7ie77t ventum erat, tbi haesltabat), in the other nine cases it is 
 the unchanged perfect that we find, as in Liv. i, i, 7 : alii proelio 
 vicium Latinuni pacem cum Aenea, deinde adfiyiitatcm iunxisse 
 tradunt ; alii, cum instructae acies cons litis senty priusquam signa 
 canerent, processisse Latinum inter primores ducemqtce adve- 
 narum evocasse ad conloquium ; percunctatum deinde, qui mor- 
 teles essent, unde aul quo casu profecli domo quidve quaerentes in 
 agru77t Latirentinum exissent, postquam audierit mullitudinem 
 
 Troianos esse, due em Aeneam, . . »fidem futurae amicitiae sanx- 
 isse. . . . others have the version that Latinus inquired who they 
 were, etc., and, when he heard (not had heard') that they were 
 
 Trojans . . . gave by the offer of his hand a solemn bond of peace 
 for the future ; Cic. Rep. 2, 2, 4 : is igitur, ut natus sit, cum 
 Remofratre dicilur ab Amulio rege Albano ob labefactandi regni 
 iimorem ad Tiberim exponi iussus esse ; Fam. 4, 3, 4 : ta7itum 
 dicam, quod te spero adprobaturum, me postea quam illi arti, cui 
 studuerain, nihil esse loci neque in curia neque in foro viderim, 
 omnem meam curam atque operam ad philosophiam contulisse. 
 Fam. 5, 8, 3 : de me sic existimes ac tibi persuadeas vehe77ie7iter 
 velim, non me repentina aliqua volu7itate aut fortuito ad tuam 
 amplitudinem meis officiis amplectendam incidisse, sed, ut primum 
 forum attigerim, spectasse semper ^ ut tibi possem qua7n maxima 
 esse coniunctus. 
 
 In the same way, when the common phrase non putaram goes 
 into the subjunctive in the oratio obliqua, it preserves its individu- 
 ahty of tense, as in Cic. Sen. 2, 4 : obrepere aiunt earn citius quam. 
 putasscnl; Att. 6, i, 6; and frequently. No more absolute proof 
 of the temporal expressiveness of the attached subjunctive verb 
 could be desired than is given by these usages.^ 
 
 » 
 IV. 
 
 The destructive part of my task has taken so much space that 
 
 *A complete survey of existing views would include a discussion of the 
 application of the doctrine of Absolute and Relative Time to the field of the 
 supposed sequence. The limits of the present paper exclude such discussion. 
 What has been said above, however, in regard to a possible subtilized theory 
 applies a fortiori to the coarser theory of Absolute and Relative Time ; nor 
 have I any fear that a reader who has agreed with me thus far will find a 
 resting-place in that doctrine. 
 
 At a future day I hope to show that the doctrine is untenable. Nevertheless 
 it has performed, especially in Germany, the good service of weakening 
 popular faith in the universal tmth of the old doctrine. 
 
42 
 
 I am obliged to state the constructive part in very summary 
 fashion. 
 
 With certain exceptions, each tense of the indicative indicates 
 to the hearer two things, the stage of advancement of the action 
 (whether it be complete, in process, or yet to be), and the position 
 in time of the point of view from which the act is regarded (whether 
 it be somewhere in the past, at the moment of speaking, or some- 
 where in the future). In each of the three verbs {domiis) aedificaia 
 erat, aedificaia est, aedificaia erii, the house is presented in a 
 completed state, the point of view alone changing. These verbs 
 are, to use a more exact nomenclature than the one in vogue, 
 respectively pasi perfect, present perfect, future perfect. In the 
 same way aedificabatur, aedificatur, aedificabitur represent an 
 action in process in the past, at the present, in the future ; or, more 
 exactly, these verbs are respectively past imperfect, present imper- 
 fect, /w/z/r^ imperfect. In the three verbs aedificaturus erat, aedi- 
 ficaiurus est, aedificaturus erit we have, similarly, a past future, a 
 present future, and a future future. 
 
 Now, these indications in themselves convey each two things 
 only : i. The point of view of the mind asserting ; 2. The stage 
 of the action at that point of view. But a third conception 
 necessarily enters in. If an act can be asserted to be in a com- 
 plete state at a certain time in the past, it is a certainty that the 
 activity had been prior to the time thought of as the standpoint. 
 The idea of the priority of the act to the standpoint is, then, 
 practically conveyed by the three perfect tenses. In the same 
 way, if an act is asserted as in process at a certain time in the 
 past, it is inevitable that the activity was contemporaneous with the 
 time thought of as the standpoint. The three imperfect tenses, 
 then, convey, in addition to standpoint and stage of action, a third 
 idea, that of contemporaneousness. Each of these six tenses thus 
 practically carries three distinct ideas to the hearer's mind : i. The 
 point of view from which the speaker puts the act ; 2. The stage 
 of advancement of the act at that point of view ; and 3. The 
 temporal relation of the activity itself to that point of view. 
 
 The subjunctive likewise is furnished with tenses which indicate 
 that the point of view from which the act is seen in imagination is 
 in the past or at the present, and that the act is seen as complete, 
 or is seen without reference to completion (the past complete and 
 present complete, the past non-complete and present non-com- 
 plete). In other words, the subjunctive tenses indicate standpoint 
 
43 
 
 and stage. So far they are like the indicative tenses. But they 
 go no farther. The idea of the temporal relation of the activity 
 to the standpoint, of its being before, or being at, or being after 
 the standpoint, cannot, in the very nature of the mood, be involved. 
 If, at the present moment, I form a picture in my brain of (say) a 
 book completed, there is absolutely nothing in the tense that can 
 fix the act at any point between the beginning of time and the end 
 of time. The vision of the completed book may be of a book 
 said to have been made long ago, or it may be of a book which I 
 hope to have in completed shape ten years hence. The point of 
 view is definite and exact ; but from the very fact that there is no 
 assertion of outward reality in the subjunctive mood, but merely an 
 imagining of an act, no exact placing of the act here or there in time 
 is possible. It follows that a form like scripius sit, e. g., which is in 
 its earliest history a parallel of neither the perfect indicative nor of 
 the future perfect indicative, but merely a vision of a finished act, 
 is used to represent what corresponds in the subjunctive to both 
 these very different forms. I may say, for example, ab Homero 
 scripta sit {suppose that Homer did write the Iliad), and, by the 
 same tense, sit deiiique inscriptum ifi f route unius cuiusque quid 
 de re public a seniiat. Cic. Cat. i, 13, 32 (be //written on every 
 man's forehead whether he is loyal or disloyal'). In the same way 
 the past non-complete subjunctive facerem and the present non- 
 complete subjunctive faciam strictly present to the mind only a 
 vision of an act without reference to completion, seen from a past 
 and a present standpoint respectively. In these tenses, however, 
 we find a certain necessary limitation. The activity is not thought 
 as lying back of the standpoint, for then the tense used would be 
 one of the perfects. But further than this there is no limit. The 
 non-complete act seen in imagination as from the present moment 
 may belong anywhere in the stretch from the present moment in- 
 clusive to the end of time, and the act similarly seen as non-com- 
 plete from a past standpoint may belong anywhere in the stretch 
 from that time on to the end of time. In other words, the sub- 
 junctive tenses of non-complete action can apply to any act present 
 to or future to the standpoint.' 
 
 * In this immediate power of application to the speaker's future lies the 
 explanation of the fact that no new and specialized subjunctives from a future 
 standpoint have arisen ; and herein also is the origin of the temporal power 
 of the so-called indicatives of the future, themselves no indicatives originally, 
 but (to speak as a Latinist) subjunctives. 
 
44 
 
 The subjunctive tenses, then, indicate, like the indicative tenses, 
 the point of view from which the act is put as pictured in the 
 brain, and the sta^o^e of advancement in which the act is represented 
 to be ; but they here part company with the indicative, and are 
 incapable of expressing the temporal relation of priority, contem- 
 poraneousness, or futurity to the standpoint. The complete tenses 
 can apply to any act seen as complete anywhere in the whole range 
 of time ; while the non-complete tenses can apply to any act seen 
 as non-complete at or after the standpoint. 
 
 So much, and only so much, is inherent in the nature of the 
 subjunctive tenses. But in their actual use in conveying this or 
 that idea, a temporal feeling inevitably grows up with regard to 
 each use of each of them. In thinking 'a realizable wish or a 
 command (the point of view being of course the speaker's present) 
 we have a mental vision of the act as lying in the future (whether 
 immediate or remote) ; and the hearer, getting our idea, naturally 
 associates futurity with the tense of the verb. In making a con- 
 cession from the present point of view, on the other hand, we 
 mostly have in mind a present act or state, or an act or state com- 
 pleted by or before the present ; and the hearer, getting our idea, 
 naturally associates contemporaneousness or priority, as the case 
 may be, with the tense of the verb. In this way there arise two 
 distinct accretions of meaning for each subjunctive tense — significa- 
 tions not inherent in the nature of the tense, but naturally involved 
 in the special kind of idea which the tense is used to convey. The 
 so-called perfect subjunctive serves as an aorist or present perfect, 
 and also as a future perfect ; the so-called present serves both as a 
 present and as a future; and, in the same way, the so-called 
 pluperfect serves as a past perfect and as a past future perfect, the 
 so-called imperfect as a past present and a past future. In other 
 words, in practical use each tense of the subjunctive is found to be 
 employed with two distinct ideas, one that which is indicated by 
 the tense of the indicative bearing the same name (as in indirect 
 questions), the other a future idea (as in the final clause, commands 
 in indirect discourse, etc.) ; so that the so-called pluperfect and 
 the so-called perfect serve, from their respective standpoints, as 
 either perfect or future perfect, and the so-called imperfect and 
 present serve from their respective standpoints as either present or 
 future. 
 
 With each subjunctive construction, then, there is in time associ- 
 ated a definite temporal meaning, seen clearly in the independent 
 
. 45 
 
 construction, and abiding in the dependent use of it. The jussive, 
 e. g., refers to a time future to the standpoint, and its dependent 
 application (the final clause) expresses a present purpose (present 
 subjunctive), or a past purpose (imperfect subjunctive). And in 
 the same way a definite temporal meaning is found to be attached 
 to each subjunctive dependent construction that has grown out of 
 an independent subjunctive construction, while in each dependent 
 subjunctive construction that is a conversion (the indirect discourse) 
 of the indicative construction, the meaning is precisely the same 
 as in the indicative ; the clause quid scripsisset^ e. g., meaning 
 precisely the same thing, so far as anything but mood goes, as 
 quid scripserat. 
 
 Now, how to bring this to bear for a beginner ? First make him 
 understand precisely what the indicatives convey to the mind. 
 Then show him, by giving him parallel examples in the direct 
 question and the indirect question, that the tenses of the subjunc- 
 tive convey precisely the same mental standpoint, or point of view, 
 as the tenses of the indicative bearing the same name. Have this 
 idea of the standpoint very clearly felt by the pupil. Then, in no 
 haste, show him by examples that each tense of the subjunctive, 
 beside the force corresponding to that of the indicative bearing the 
 same name, has a future force, as in the purpose clause, the stand- 
 point always remaining unchanged. Add to that the statement 
 that, by a peculiar development, the tenses for conditions, conclu- 
 sions, and wishes put as from a past standpoint came to convey 
 the idea of conditions, conclusions, and wishes contrary to fact, in 
 Latin as in English, and that by another peculiar development the 
 imperfect came to express past results in their temporal relation 
 with their causes, and you have a practical treatment covering the 
 entire ground. For the converse work of writing Latin, tell the 
 student to use a pluperfect or imperfect to indicate that the point 
 of view is past, i. e., if it is a past purpose, a past question, a past 
 ground of action, and so on ; and a perfect or present to indicate 
 that the point of view is present or future, i. e., if it is a present 
 purpose, a present question, a present ground of action, etc. Make 
 him see that our use of tenses in English is mostly the same, alike 
 in independent and in dependent sentences, as, e.g., in the coordi- 
 nated What was she trying to tell me f I had no idea, and the 
 subordinating / had 7io idea what she was trying to tell me. 
 
 This is all simple enough, and young children, provided they 
 have not been taught a rule that " primary tenses are followed by 
 
46 
 
 primary," etc., have, as has been proved by actual experiment per- 
 formed by other teachers under my own eyes and at a distance, no 
 difficuhy in understanding it in an entirely real and unmechanical 
 way. But no one can venture to make such a statement as regards 
 the practical working of the rule about primary tenses being 
 followed by secondary, and the rest. Understanding is precisely 
 the thing that cannot be claimed for those mental processes in 
 interpreting and writing Latin tenses which the grammars aim 
 to set up in the learner's mind. The directions which I have 
 given above, though they take the student only part way on the 
 road toward a complete theoretical understanding of the whole 
 matter, are sound as far as they go, and calculated to develop 
 understanding, needing only to be filled out at a later day ; while 
 the ordinary rules, which are founded on nothing but a count of 
 examples, are calculated to beget a self-contented mental vacuity, 
 and must be wholly swept away before any true comprehension 
 can be brought about. 
 
 If, in opposition, it is urged that students must have the rule of 
 the Sequence in order to write Latin, I should answer, first, that 
 they do not handle their tenses so successfully at present, even 
 under the help of the Law, as to justify any white lies ; secondly, 
 that a man who hits the right tense by a rule of thumb without 
 understanding or feeling, writes better Latin but is not a better 
 man; thirdly, that, though the uses are essentially the same in 
 German and French, one who should attempt to introduce a 
 doctrine of a Sequence as indispensable in learning to write those 
 languages would be derided ; and, lastly, I should call attention to 
 the fact that the rule of the Sequence very frequently betrays the 
 student. Every teacher must have had the experience of correcting, 
 under a hidden linguistic impulse, such as will sometimes rise above 
 the grammars, a Latin tense written by a student in entire confor- 
 mity to the rule, but conveying a wholly different idea from the 
 English which it is meant to represent. Suppose, for example, I 
 ask a student to express in Latin, under the " Law," what ivas the 
 character of the state at that time, and what had it been tip to that 
 time f He will write, with perfect feeling for the tenses, qualis erat 
 illo tempore civitas, et qualis antea fuerat. Suppose, now, I ask 
 him to write in Latin let us see, in Cato's own words, what was the 
 character of the state at that tivte, and what it had been before thai 
 time. He will not dare to write qualis esset illo te^jipore civitas^ et 
 antea qualis fuisset, videamus in ipsa sententia Catonis, as a student 
 
47 
 
 who knew nothing about a Sequence oT Tenses would, and as St. 
 Augustine, who also had the advantage of being in ignorance of 
 the rule, did, in De Civ. Dei, 5, 12 ; neither would he dare to write 
 though the battle lasted till evening, nobody could catch sight of an 
 enemy's back, as Caesar did in B. G. i, 26. 
 
 V. 
 
 And now a brief last word about the history and the hopes of the 
 doctrine here professed, that the tenses of the subordinated sub- 
 junctives are expressive, not mechanically dictated by a preceding 
 verb ; that they mean the same thing, tell the same story, as 
 the tenses of the corresponding independent indicatives or inde- 
 pendent subjunctives. 
 
 In 1872 Lieven (Die Consecutio Temporum bei Cicero), laying 
 down the traditional rules for the Sequence, proved by examples 
 that consecutive, causal, concessive and relative sentences (not 
 final) are exempt from the law when following secondary tenses, 
 and that unreal conditional sentences are exempt from the law 
 when following primary tenses. Other apparent exceptions he 
 accounted for on the theory of " pregnant " uses of the main 
 tense. His dictum (" The tense chosen in the subordinate sentence 
 is that which would have to be chosen if the sentence were inde- 
 pendent ") would have been a complete statement of the matter, 
 if it had been intended to be thoroughgoing. In point of fact, 
 however, he limits it to the cases above mentioned. The way in 
 which he went astray is clear : he treats the subjunctive in the 
 main as a mere mood of subordination. As he glances back over 
 the growth of the language from the paratactic to the hypotactic 
 stage, he sees independent indicatives becoming subjunctives, and 
 retaining their tense ; and so far he sees quite rightly. But he 
 fails to see the great part which is played by the passing over of 
 independent subjunctive constructions into the dependent form. 
 And, in so doing, he not only misleads himself in regard to the 
 history of the subjunctive causal, concessive, and consecutive sen- 
 tences (all of which, as I hope to show in a later paper, go back 
 to independent subjunctives^, assuming them to be substantially 
 merely subordinated indicatives, but also draws his line of limita- 
 tion for the exceptions very far short of the true point, and leaves 
 the old rules in the main standing. In spite of this, however, the 
 method he applied ends logically in the destruction of the traditional 
 doctrine, though he himself failed to see its full sweep. 
 
48 
 
 Five years later, Martin Wetzel, In his doctorate-dissertation 
 (Goettingen), said in his preface that the force and meaning of each 
 tense was the same in the subjunctive as in the indicative, so that 
 the question why this and that tense was found to have been 
 employed did not turn upon a Law of Sequence, but upon the force 
 inherent in each ; and that, consequently, to speak accurately, 
 there was no such thing as a Sequence of Tenses. The statement 
 is in reality nothing more than Lieven's statement more effectively 
 put, but subject to the same errors and limitations.' The exami- 
 nation is confined to an analysis of the uses of the tenses in subjunc- 
 tives which are such by reason of being in the indirect discourse, 
 and to certain changes of tenses of other subjunctives in the 
 indirect discourse after a main verb of one and another tense ; and 
 does not take up the question of the force of the tenses in de- 
 pendent subjunctives corresponding to independent subjunctives.^ 
 And Wetzel's subsequent work should have carried him on, 
 through a wider survey of the field, to the doctrine that the tenses 
 of the subjunctives in dependent constructions convey the same 
 meaning as the tenses of the subjunctives or indicatives, which- 
 ever it may be, in the corresponding independent constructions. 
 But, as we shall shortly see, he did not attain to this doctrine. 
 
 In 1882 Ihm, in his Quaestiones Syntacticae de Elocutione 
 Tacitea comparato Caesaris Sallusti Vellei Usu Loquendi (Giessen), 
 finds the solution of the whole problem in the application of the 
 doctrine of Absolute and Relative Time — that doctrine, taught by 
 Hoffmann and supported by Llibbert, which has been so potent 
 in Germany and America, for good or for evil, since the appear- 
 ance in 1870 of the latter's Die Syntax von Quom. Ihm was 
 followed in 1884 by Lattmann and Miiller in their Kurzgefasste 
 Lateinische Grammatik, and in 1885 Wetzel, in his Beitrage zur 
 Lehre von der Consecutio Temporum, amends, and, as amended, 
 accepts the doctrine of Lattmann and Miiller. The same man, 
 then, who in 1877 said, at the end of his university career, the 
 best thing that had then been said on the subject, and was distinctly 
 
 ' Cf. also, from the preface, p. 6, the following: Ac deliberanti mihisaepenu- 
 mero in eo potissimum omnium errorum fons et causa posita esse visa est, quod 
 temporum consecutionem illud efficere, ut tempora in coniunctivo enuntia- 
 torum secundariorum non omni ex parte eandem vim retineant quam in indica- 
 tive habent, sed sola verbi regentis forma definiantur, plerique opinantur. 
 
 2 The force of the tense of the subjunctive in the dependent deliberative 
 question, the original consecutive clause, the final clause, is not the same as 
 the force of any existing indicative construction. 
 
49 
 
 on the right road, has led himself into the cloudland of Absolute 
 and Relative Time. Meanwhile, however, Hermann Kluge, of 
 the Gymnasium at Cothen, published in 1883 a treatise of great 
 importance, Die Consecutio Temporum, deren Grundgesetz und 
 Erscheinungen im Lateinischen.^ 
 
 In this treatise Kluge, omnia ad se trahens, ignores the great 
 suggestiveness of Lieven's and Wetzel's partial proposition of the 
 years 1872 and 1877 ; ignores the very great contribution to a 
 proper psychological treatment of the general question and the 
 explanation of important details given by Otto Behaghel in 1878 in 
 the treatise already cited ; and ignores the very helpful statement 
 of the general nature of the indicative and subjunctive tenses 
 given by Haase in the second volume of the Vorlesungen iiber 
 lateinische Sprachwissenschaft (edited by Hermann Peter, 1880). 
 Further, he is astray, me iudice, in very many important details, 
 which I have space barely to enumerate in part, without dis- 
 cussion : The theory that the imperfect and pluperfect subjunctive 
 are originally tenseless, differing from the present and perfect 
 only in presenting a more remotely conceived idea ; the theory, 
 naturally connected with this, that the use of the imperfect and 
 pluperfect in wishes, conclusions, and conditions referring to the 
 speaker's present is not derived, but original — a view which would 
 find it hard to reckon with the indisputable origin of the imperfect 
 and pluperfect indicative referring to the same time in cases like 
 oporfuerat, oportebat^ and the analogy of the history of, e. g. 
 auxiliaries like the English would, should, might (preterites), and 
 
 'A word of personal explanation must at this point be granted me. The 
 doctrine of this paper I taught, somewhat timorously, as became a young 
 instructor, as early as the years 1877, 1878, and 1879, ^^^^ with emphasis since 
 the year 1880, when I was called to another university and to a position of 
 responsibility. It was my purpose to publish and advocate my doctrine at the 
 earliest possible moment, but in the press of duties I allowed the years lo slip 
 by, and was obliged to see the main tenet of my belief first printed in Kluge's 
 book. As will appear below, I regard Kluge's treatment to be in many respects 
 unsound; but the essential doctrine is true, that the tense of the subordinate 
 verb is the direct expression of the speaker's meaning. Anticipated, then, in 
 date of promulgation, and quite possibly even in actual length of years of 
 possession of these views, I avow myself a supporter of Kluge, and a preacher 
 of his faith. The question of priority of publication is, at the present point in 
 the development of human nature, of much interest to the individual, but it is 
 of little consequence to the world. What is of consequence is that sound doc- 
 trine should be reached as early as possible, and taught by as many men as 
 possible. 
 
50 
 
 the German wiirde, sollte, mochte, etc. ; the theory that the 
 imperfect indicative indicates duration of action (" Dauer ") ; the 
 astounding theory that in a sentence Hke Livy's in I, 3 : tantum 
 opes creverant . . . ut ne morte quidem Aeneae nee deinde inter 
 muliebrem tutelam . . . movere arma . . . u/li alii accolae ausi 
 sint, the perfect is used because the statement fills Livy, as he 
 tells the story, with such interest that the incident appears to him 
 not to be on the same plane with the other points of the narrative, 
 but to be, in a word, remarkable enough to be brought into con- 
 nection with the actual present of the writer ; by which Kluge 
 means, as clearly appears elsewhere, that such perfects as ausi 
 sint are logical perfects, perfects definite, utterly failing to recog- 
 nize, as many had done, years ago, that these perfects are simply 
 subjunctive aorists corresponding precisely to independent indica- 
 tive aorists ; ' the theory that the final clause is developed out of 
 the consecutive clause. He errs, moreover, in attributing meta- 
 physical rather than concrete origins to the various dependent 
 subjunctive constructions. He gives no proof, such as has been 
 attempted in this paper, of the unsoundness of the prevailing 
 doctrine, nor does he protect his theory from attack by raising and 
 meeting the apparent objections founded on the common use of the 
 imperfect subjunctive in result-clauses (he is quite wrong in the 
 matter, regarding the tense as always indicating the action as 
 " laufend "), and the use of the form -turns fuerit in subordinated 
 conclusions contrary to fact, etc., etc. Nevertheless, the doctrine 
 that in the subordinate sentence the speaker's meaning alone 
 determines the tense is here for the first time stated sharply and 
 as covering the whole ground ; and to have done this is a very 
 great service. 
 
 In no school grammar or manual published in Germany since 
 then, however, has this doctrine been taught, so far as my knowl- 
 edge goes — not, at any rate, in the grammars of Schottmiiller- 
 Putsche (1884), Ellendt-Seyfiert (edition of 1886), Kuhner's 
 Elementargrammatik (1884). Of the still recent grammars, etc., 
 published a little earlier, Josupeit's (1882) states to the full the old 
 doctrine in these words : " In the dependent subjunctive the con- 
 
 ^ This forcing of the meaning of the perfects in question is as shortsighted 
 as it is extraordinary ; for Kluge fails to see that, when he has tortured these 
 perfects into perfects definite, he still has to confront and account for the 
 fact that, as noticed on page 65 of this paper, they themselves are followed in 
 most cases by the imperfect and pluperfect. 
 
51 
 
 ception of time utterly vanishes ; that conception is given by the 
 governing verb ; nothing remains to the subjunctive except the 
 conception of the act as complete or still lasting with reference to 
 the governing verb " (§83). Feldmann (1882) says (§69, 3) that 
 " result-clauses are not subjected to the Sequence of Tenses." Gold- 
 bacher (1883) says that in all "innerlich" dependent subjunctives 
 the tense is under the influence of the tensejm the governing sen- 
 tence; these " innerlich " dependent subjunctives being those that 
 are expressed as in the mind of the subject of the governing sen- 
 tence, namely, final sentences, sentences after antequam, priusquam, 
 dum, donee, quoad, many relative sentences, questions and subor- 
 dinate verbs in the indirect discourse ; in result-clauses, however, 
 that tense is used which would have been used in an independent 
 construction, excepting that in pure result-clauses with ut the 
 imperfect usually stands after the perfect. Here is to be seen a 
 single plant sprung from the seed planted by Lieven in 1872. In 
 the grammars of Holzweissig (I have before me the edition of 
 1885) and Ellendt-Seyffert (1885 and 1886) a richer growth 
 appears, but nothing more than in Lieven's treatise ; for these 
 grammars teach that the rules of the Sequence of Tenses hold, 
 but only for " innerlich " dependent sentences, while consecutive, 
 causal, concessive, and non-final relative sentences are not subject 
 to th^ rule. In no school grammar in Germany, then, has the 
 true doctrine found a lodgment. Still there is great significance in 
 this distinct narrowing of the field of the operation of the Law. 
 Such things show a drift of opinion ; and that drift is clearly away 
 from, not in the direction of, faith in the Law of the Sequence. 
 
 Antoine, in his Syntaxe de la Langue Latine, 1886, has got no 
 farther on than Lieven. In the latest French grammar, Reinach's 
 Grammaire Latine, a dissatisfaction with the old way and an un- 
 readiness to break with it are shown at the same moment in the 
 statement that " the concord of tenses in Latin is subject to two 
 general rules, which are rather logical tendencies than laws of the 
 language: i. If the main verb refers to the present or the future, 
 and the dependent to the present or the past, the present or perfect 
 of the subjunctive is used in the dependent verb ; 2. If the main 
 verb refers to the past, the imperfect or pluperfect is used in the 
 dependent verb "; and the same jarring of views is seen in the 
 quoting of a sentence from Kluge and another from Ihm in the 
 immediate neighborhood of the statement that " the other irregu- 
 larities in the consecutio temporum are to be referred to the struggle 
 
52 
 
 of logic with grammar," a sentence not to be reconciled with the 
 true doctrine that the Latin tenses successfully tell their own story. 
 So, then, it appears that no school grammar has yet taught this 
 simple doctrine. In one notable case, however, has a refreshing, even 
 if too brief, treatment appeared, in what may be called a grammar 
 for specialists. In the grammar of Stolz and Schmalz, published 
 in 1885, before that of Reinach, the entire treatment of the 
 " sogenannte consecutio temporum " is confined, with a noble 
 disdain, to fifteen lines and two-thirds ; and although no proof is 
 given, and no light thrown upon the apparent difficulties, as, e. g.y 
 the use of the imperfect in result-clauses (a matter especially 
 suitable for explanation in a grammar of such aims), yet it is 
 expressly laid down, in the exact words of Kluge's treatise, that 
 " a mechanical dependence of the tenses of the subordinate sen- 
 tence upon those of the main sentence does not exist, and that the 
 choice of the tense in each sentence depends upon the conception 
 lying at the bottom of it." After such a note as this, struck by 
 what may be expected to prove an influential grammar, I have 
 entire faith in the success, at no remote time, of the true doctrine, 
 to the immense relief and profit of the Latin-studying mind. This 
 true doctrine cannot, however, be preached to the people in the 
 highways. It can reach them with ease and conviction only 
 through their sacred books, the school grammars. And I there- 
 fore address my protest to that body of actual or potential makers 
 of those sacred books, the members of the American Philological 
 Association. 
 
 William Gardner Hale. 
 
UNIVERSITY OF CALIFORNIA LIBRARY 
 BERKELEY 
 
 Return to desk from which borrowed. 
 This book is DUE on the last date stamped below. 
 
UNIVERSITY OF CALIFORNIA LIBRARY