GRICE E CECCATO: PRÆTERITVM IMPERFECTVM PERFECTVM PLVSQVAMPERFECTVM

 THE SYNTAX OF THE IMPERFECT INDICATIVE  IN EARLY LATIN   By Arthur Leslie Wheeler   In his Studien und Kritiken zur lateinischen Syntax, I. Teil,  Mainz, 1904, Dr. Heinrich Blase has devoted considerable space to  my article, "The Imperfect Indicative in Early Latin" (American  Journal of Philology XXIV [1903], pp. 163 ff.). Since Blase  professes to present the substance of my article, except to the  'relatively few' German scholars who have access to the American  periodical, and since he makes a number of errors in mere citation  and statement, it becomes necessary for me in self-defense to make  some corrections. 1 But apart from these errors of detail, which  will be pointed out at the proper places, Blase disagrees with some  of the more important conclusions of my paper and it is with the  purpose of elucidating these views in the light of his criticism and  contributing something more, if possible, to a better understanding  of the problem that I offer the present discussion.   The functions of the imperfect indicative in early Latin may  be summarized as follows:   I. The Progressive 2 or True Imperfect, comprising several  types or varieties:   A. Simple Progressive.   1. dicebat = il he was saying."   1 That such corrections are justifiable is proved by the fact that K. Wimmerer, who  knows my article only through Blase's presentation, reproduces several of Blase's in-  correct statements. I regret the unavoidable delay in the publication of this paper  the less because it has enabled me to use Wimmerer's article, "Zum Indikativ im  Hauptsatze irrealenBedingungsperioden," Wiener StudienXXWll (1905, publ. Feb. 10,  1906), pp. 260 ff. The first four pages of his article are devoted to a general discussion  of Blase's critique of my views.   2 In this paper technical terms will be used as follows : progressive = German vor  sich gehendes (less exactly fortechreitendes) ; continuative or durative = wiaftrendes;  nature or kind of action=^Lfc<ionsarf; shifted = verschobenes ; descA\)tive= schilderndes;  reminiscent = erz&hlendes (see p. 365) ; relation (relative, etc.)= Beziehung, etc. Other  terms are, it is hoped, intelligible or will be defined as they occur.   [Classical Philology I October, 1906] 357     358 Arthur Leslie Wheeler   The nature of the action may be either progressive 1 or con-  tinuative (durative). The time is past, but the period covered by  the action of the tense may vary with the circumstances described  from an instantaneous point to any required length. The time  is contemporaneous with, usually more extensive than, the time  of some other act or state expressed or implied. When the tense-  action is continuative and extends into the immediate past or,  by inference, the present of the speaker, I would distinguish a  sub-class :   a) The Imperfect of the Immediate Past:  dicebat—"he was saying" or "he's been saying."  The action may or may not be interrupted by something in the  context. If interrupted, it ends sharply and we may term the  tense the "interrupted" type of this immediate past.   2. The Descriptive Imperfect (better, the imperfect used in  description) .   dicebat="he was saying" (in English often rendered by  "said").   This is in its purest form a simple progressive imperfect  employed in the vivid presentation of past actions or states.   3. The Reminiscent Imperfect (better, the imperfect used in  reminiscence).   dicebat=^ u he was saying" (as I remember, or as you will  remember).   In this usage the imperfect is a simple progressive implying  an appeal to the recollection of the speaker or hearer.   B. Customary Past Type.   dicebat="he used to say, would say, was in the habit of  saying, etc."   The nature of the action is the same as in A except that with  the aid of the context there is an implication that the act or state  recurred on more than one (usually many) occasions. These  recurrences are usually at some considerable distance in the past  and contrasted with the present, but cases of the immediate past  usage (Ala)) with customary coloring occur.   i Hoffmann Zeitpartikeln 2 , p. 185, characterizes excellently this feature of the im-  perfect : " die actio infecta, pendens, die Handlung in der Phase ihres Vollzuges, ein  Geschehenes im Verlaufe seines Geschehens, ein Vergangenes Sein noch wahrend  seines Bestehens."     Impebfect Indicative in Eably Latin 359   C. The Frequentative or Iterative Type.   dicebat = "he kept saying" (at intervals very close together).  This type is like B, except that it has no customary element and  the repetitions refer to one situation within comparatively narrow  limits of time.   The link connecting all these varieties with one another is the  progressive function. 1   II. The Aoristic Imperfect.   aibat = "he said" (equivalent to dixit, aoristic perfect).  The time is still past, but the progressive force is lost.   III. The Shifted Imperfect.  debebat = "he ought" (now).   The time is shifted to the present and the progressive force is  very much weakened, in some cases wholly lost, because of the  auxiliary character of the verbs involved.   For a more detailed treatment of the foregoing classes (except  the imperfect in reminiscence) I must refer to Am. Jour. Phil.  XXIV, pp. 163 ff. In what follows I shall select certain points  for discussion by way of elucidation and supplement to what was  said there.   the impebfect of the immediate past   The simplest progressive usage is well enough understood,  but the usage termed by me the imperfect of the immediate past  or interrupted imperfect 2 calls for some remarks. As a type of  this imperfect in its interrupted form cf. Plautus Cas. 178: nam  ego ibam ad te. — et hercle ego istuc ad te. Here the action is con-  ceived as continuing until interrupted by the meeting of the speak-  ers. The fact of the interruption does not, of course, inhere in the  tense but is inferred from the context. Indeed, the interruption  may not occur at all, as will be seen by comparing the second type,  e. g., Stick. 328 f. : ego quid me velles visebam. nam mequidem  harum miserebat. Here visebam is interrupted like ibam above,   1 The nature of the action seems to me the most distinctive feature of the tenses.  In this I differ radically from Cauer, who considers contemporaneousness the essential  feature of the imperfect, cf. Grammatical militans, 1903, pp. 93, 94, against Methner,  whose Untersuchungen zur lat. Tempus- und Moduslehre, Berlin, 1901, 1 have not seen.   2 B. Wimmerer Wien. Stud., 1905, p. 262, Anm. 2, calls attention to the fact that  this imperfect of the immediate past in its interrupted form is still common in Italian.     360 Arthur Leslie Wheeler   but the action of miserebat is conceived as continuing not only up  to the immediate past, but into and in the present of the speaker.  But again this continuance in the present is not inherent in the  tense; it is inferred from the context. The nature of the action  is in both these types still progressive, or more exactly, continua-  tive, but temporally stress is laid on that period of time immediately  preceding or even extending into the present. 1   In this usage the Romans possessed a somewhat inexact sub-  stitute for the English progressive perfect definite, e. g., mequidem  . . . . harumnusere6a/ = (practically) "I've been pitying,"a form  which, like the Latin, may be used in the proper context to indi-  cate that the pity still continues in the present. 2 On the other  hand, the English "I was pitying," superficially a more exact  rendering, does not so clearly indicate this continuance in the  present, though "I was going to your house, etc." is an exact  rendering of Cas. 178.   Blase himself has collected some exactly similar cases, 3 of which  he says:   Das Imperf. wird gelegentlich auch von Zustanden gebraucht die  zwar in der Gegenwart des Redenden noch fortdauern aber nur mit Bezie-  hung auf die Vergangenheit genannt worden: Plaut. As. 392 quid quae-  ritas? Demaenetum volebam .... Das Wollen dauert fort, aber hier is t  es nur in Beziehung auf die in Gedanken vorschwebende vorausgehende  Zeit bis zur Ankunft vor dem Hause gebraucht.   'Blase {Kritik, p. 6) misrepresents my statement concerning this usage. He cites  from my paper Stich. 328, apparently as given by me in illustration of both the pro-  gressive use in its simplest form and of this immediate past usage, although it was used  as an illustration of the immediate past usage only. Again he quotes me as believing  that in the immediate past usage the action takes place within exactly defined limits  ("genau bestimmten Granzen"). Here is atwofold error. My statement (Am. Jour.  Phil. XXIV, p. 168) is "fairly definite limits" and refers to the simple progressive  usage, not to the immediate past usage. Blase's critique confuses the two usages.   2 There are traces of a tendency on the part of the Romans to express these shades  of thought with greater exactness, e. g., by the combination of a present participle  with the copula, Plautus Capt. 925 : quae adhuc te carens dum hie fui sustentabam.  Here carens .... fui is exactly equivalent to the English "I've been lacking,"  whereas sustentabam is inexactly equivalent to "I've been supporting." But Latin  did not develop such expressions as carens .... fui into real tenses, and remained  content with the less exact imperfect, cf . also iam diu, etc., with the present. See Am.  Jour. Phil. XXIV, p. 185, and Blase Hist. Syntax, p. 256. A complete collection of  such cases would be interesting. I would add here Amph. 132 : cupiens est, Rud. 943 :  sum indigens, and cf. the verse-close ut tu sis sciens (Poen. 1038), etc.   "Hist. Syntax III, 1903, Tempora und Modi, p. 148, Aran. This book had not  reached me when my article in Am. Jour. Phil. XXIV was written.     Imperfect Indicative in Eably Latin 361   With the first part of this statement I fully agree, but is it true  that in As. 395 the imperfect is used "nur mit Beziehung auf die  Vergangenheit, etc." ? If, as Blase says, "das Wollen dauert fort,"  then we are forced to say that the imperfect is used not merely  with reference to the past, but with reference to the present. 1 The  speaker really has in mind both past and present, and uses the  imperfect to express this double temporal sense, the action con-  tinuing from the past into the present, because at the moment of  speaking the past is somewhat more prominent. The tense is,  therefore, as explained above, only an approximate expression of  the thought. Had the present been more prominent, other ele-  ments being equal, some expression like iam diu volo would have  been employed.   Blase asserts (Kritik, p. 6) that my statement that the speaker  has in mind both beginning and end of the action is not capable  of proof. It is true, I think, that the speaker has usually no  definite point in mind at which the action began. He simply  indicates the action as beginning somewhere in the past and con-  tinuing in the present. But in the very numerous "interrupted"  cases he has in mind a sharply defined end of the action. 2 Blase's  criticism seems justified, then, only with reference to those cases  of which Stich. 328, .... harum miserebat is a type.   But Blase classifies cases of this usage under no less than three  different heads in his Tempora und Modi. In addition to the  case cited above, As. 392 volebam, which he interprets, as I have  tried to show, almost correctly, he cites (p. 146) Trim. 400: sed   'Of. also the use of nunc, etc., with some of the cases: Plautus Merc. 884; quo  nunc ibas? ibid. 197, Ter. Andr. 657 f. : iam censebam.   2 B. Wimmerer Wien. Stud., 1905, p. 262, says: "Sohalteich .... die Konsta-  tierung eines," imperfect of the immediate past or the interrupted imperfect, "fiir  einen glucklichen Gedanken," though he would not make a special type of this use. It  seems to me so common (about 200 cases) as to deserve the degree of special notice  which I have given it (Am. Jour. Phil. XXIV, pp. 168 f.). He adds in a note: "Hier  tut Blase m. E. Wheeler einigermassen unrecht, wenn er dessen Behauptung, dass der  Sprecher in diesen Fallen Anfang und Ende der Handlung tiberschaue, unerweislich  nennt. Wheeler kann dies mit Becht behaupten, wenn es sich um einen Gedanken  handelt, der einen beherrschte bis zu dem Augenblick, wo man ihn konstatiert,"  pointing out also that Blase would be justified only in criticizing the form of my ex-  pression so far as I wished to apply it to the cursive " Aktionsart" (i. e., those cases  where there is no interruption?).     362 Arthur Leslie Wheeler   aperiuntur aedes, quo ibam 1 as "erzahlendes" (p. 148), Merc. 885:  quo nunc ibas as "sogenannt. Oonatus." The function of the  tense is essentially the same in all these cases, the only variant  being the presence or absence of interruption which is inferred in  all cases from the context.   Since Blase classifies so many of these cases under the head of  the conative imperfect, a consideration of that usage seems here  in place.   A "conative" imperfect ought to mean an imperfect which  expresses attempted action, but since there is no trace, at least in  early Latin (cf. Am. Jour. Phil. XXIV, pp. 179, 180), of such  a function, the term is a bad one. 2 Why then retain it, as Blase  does, for those imperfects which express "den wahrenden, aber  nicht zu Ende, geftihrten Handlung?" These imperfects are  chiefly of the type which I have termed "interrupted," where the  context implies it, or imperfects of the "immediate past," where  there is no interruption. 3 In neither case is there anything more  than a simple variation of the progressive (here more exactly  continuative) imperfect.   But most of Blase's cases are not even of this idiomatic inter-  rupted or immediate past variety. They are simple progressives  in contexts which imply that the action was interrupted 4 or not   liftam occurs often in this use : True. 921, Cas. 178, 594, Merc. 885, Tri. 400, etc. ;  cf . Am. Jour. Phil. XXIV, pp. 168-70.   2 Blase Syntax, p. 148, recognizes the inexactness of the term by his expression,  "sogenannten Oonatus." In Greek its unfitness is well expressed by Mutzbauer  (cited by Blase Kritik, p. 10, and Delbriick, Vergl. Syntax II, p. 306): "Ungenau  werden solche Imperf ekta conatus bezeichnet, von einem Versuch liegt in der Form  nichts" (Grundlagender griech. Tempuslehre, p. 45) ; cf. now Wimmerer Wien. Stud.,  1905, p. 264 : " In der Form liegt allerdings von einem Versuche nichts."   ^Wimmerer Wien. Stud., 1905, pp. 263, 264, remarks that he does not see why Blase  appears to think that there is a difference between his conception of the imperfect  de conatu and mine. Blase says (Kritik, p. 11), after defining these imperfects as  above : " Die hier vertretene Anschauung scheint mehr auf die Imperf ekta zu passen,  die Wheeler," the interrupted imperfect " nennt." This is the case, so far as Blase  confines his citations to instances of the interrupted type. There is, then, no essential  difference in our interpretation of the function of the tense in these cases. Blase  clings, apparently against his will, to the old terminology to which everybody seems  to object, whereas I would group these cases under a new term which seems to me  more exact. But Blase does not, as it seems to me, group together all the cases that  belong together.   4 1 use interrupted here not of what has been termed the "interrupted" usage,  whose distinctive feature lies in the fact that the time is in the immediate past, but as     Impeepect Indicative in Early Latin 363   completed: Men. 564 pallam ad phrygionem deferebat (Peniculus  simply depicts Menaechmus as he had last seen him; cf. 469:  pallam ad phrygionem fert) ; Cic. Sulla 49 consulatus vobis  pariebatur (just like all the other imperfects in the passage —  progressive of the descriptive variety); id.Milo 9: interfectus  ab eo est, cui vim afferebat (simple progressive, the interruption  being expressed by interfectus est) ; id. Ligar. 24: veniebatis in  Africam (progressive, the interruption being implied in prohibiti 1  five lines below) ; Caesar B. G. v. 9. 6 : ipsi ex silvis rari propug-  nabant nostrosque intra munitiones ingredi prohibebant (but  prohibebant is exactly like propugnabant — both were interrupted  by the act expressed by ceperunt in the next sentence, and note  the verb-meaning); Sallust Jug. 27. 1: atrocitatem facti lenie-  bant. at ni, etc. ( progressive = they were in the act of mitigating,  but, etc.); ibid. 29. 3 redimebat (progressive); Livy xxi. 17. 7:  mittebatur (progressive); Florus 1. 10. 1: nam Porsenna ....  aderat et Tarquinios manu reducebat. hunc reppulit (progressive  in description — that the act did not succeed is shown by reppulit) ;  Curtius vi. 7. 11: alias .... effeminatum et muliebrieter timi-  dum appellans, nunc ingentia promittens .... versabat animo  tanto facinore procul abhorrentem (again graphic description:  there is here nothing in the immediate context to show that an  effect was or was not produced. In fact versare animum does  not mean necessarily to succeed in turning one's mind, but merely  to work on one's mind; cf. Livy i. 58. 3 : Tarquinius .... ver-  sare muliebrem animum in omnes partes, where versare sums up  the preceding infinitives, but no effect is produced. So in Cur-  tius, loc. cit. , versabat has the same kind of action as is indicated  by the participles appellans .... promittens, which are summed  up in versabat); Ammianus xvi. 12. 29: his et similibus notos  pariter et ignotos ad faciendum f ortiter accendebat ( again graphic  description, cf. ibid. xvi. 32: his exhortationibus adiuvabat).   referring to interruptions in the more distant past. Where the interruption belongs  to the immediate past I have so indicated in the following criticism.   1 Surely the hearer in such a case as this would not have connected even the idea  of " nicht zu Ende gefiihrten Handlung " with veniebatis until he heard prohibiti, i. e.,  the interruption belongs purely to the context and not the immediate context at that.  This is true of many other so-called conative imperfects.     364 Arthur Leslie Wheeler   Vergil Aen. i. 31: arcebat longe Latio, cf. errabant (graphic  description = what Juno "was doing" at the time, and only the  outcome of the story proves that she did not succeed). Ibid.  239: hoc equidem occasum Troiae tristisque ruinas solabar, fatis  contraria fata rependens; nunc eadem fortuna viros .... inse-  quitur (immediate past with customary coloring, cf. contrast in  nwnc = I have been in the habit of comforting .... but now,  etc. This is one of the transitional cases between the pure custo-  mary part and the pure immediate past; cf. Am. Jour. Phil.  XXIV, p. 186, where Plautus, Mud. 1123: dudum dimidiam  petebas partem, immo nunc peto; Men. 729: at mihi negabas  dudum surripuisse te, nunc ea<V>dem ante oculos, attines, are  cited. In both of these passages, though there is no customary  coloring, there is the same contrast between continuance in the  past and the present as in Vergil loc. cit. Blase would probably  term both of the Plautus passages "erzahlende"). Tacitus Ann.  i. 6. 3 trudebantur in paludem ni Caesar, etc. (a very common  form of graphic description in Tacitus = the soldiers were being  crowded into .... but (ni) . . . . i. e., the effect was partly  produced, but was prevented, cf. Sallust Jug. 27. 1 above).   In all these cases, then, I can see no essential alteration in the  function of the tense. The idea "der nicht zu Ende geftihrten  Handlung" is derived in each case wholly from the context and  there is no reason for making a special category of imperfects  which happen to occur in contexts of this kind. Moreover, the  meaning of the verb has often been overlooked, e. g., prohibebant  (Caesar B. G. loc. cit.) may easily, with but slight aid from the  context, express "die nicht zu Ende gefuhrte Handlung;" cf.  redimebat, mittebatur, versabat, etc.   Whether the idea of real attempted action ever became con-  nected functionally with the imperfect remains to be investigated.  Certainly this did not occur in early Latin, and I doubt whether  it ever occurred. Among the cases cited by Blase are two which  more closely approximate this idea than any others. These are  Sallust Jug. 29. 3 : sed Jugurtha primo tantummodo belli moram  redimebat, existumans sese aliquid interim Romae pretio aut gratia  effecturum; postea vero quam, etc.; cf. Florus i. 10. 1: reducebat.     Impebfect Indicative in Early Latin 365   It is hard for us to feel the progressive force as the more promi-  nent in such cases. We regard as more important the attempt  which is implied in the context, but the Romans preferred to rep-  resent the act graphically as in progress, leaving the idea that it  was not successful to be inferred. When a Roman wished clearly  to express attempt (real conatus), he chose a clear conative  expression, 1 e. g., conari with infinitive.   THE IMPEBPECT IN DESCRIPTION AND BEMINISCENCE 2   In strict accuracy we ought not to speak of a "descriptive"  imperfect, but of the progressive imperfect in description. The  term "descriptive" imperfect would be justified only in case we  could distinguish from the simple progressives those cases in which  the tense is used purely for graphic presentation of actions which  might more naturally have been indicated by the perfect. Such  a distinction may often be drawn, especially after the development  of a consciously artistic style, but the separation would be worth  little since the progressive function is equally characteristic of  both. The tense was chosen for graphic purposes because its pro-  gressive function made it the most vivid of the past tenses.   The chief difference between Blase's treatment here and my  own will become evident from a consideration of his definition  (Hist. Syntax, p. 147) :   Aber seiner Hauptverwendung nach ist das Imperf. im latein. ein  Tempus der Schilderung geworden welches einmal im Nebensatz seine  Stelle hat zur Bezeichnung von Zustanden und Handlungen, die wahrend  anderer genannter Zustanden und Handlungen dauerten, und dann im  Hauptsatz bei Schilderungen von Zustanden, Sitten, Gebrauchen, welche  in Beziehung stehen zu irgead einer vorher oder nachher genannten  praeteritalen Handlung.   ! This whole question needs investigation. All the forms of expression of real  conatus should be collected and compared with the tenses as has been done for "cus-  tom" by Miss E. M. Perkins The Expression of Customary Past Action or State in  Early Latin, Bryn Mawr dissertation, 1904.   2 " Reminiscence, reminiscent" are here proposed as equivalents for the German  "Erz&hlung, erz&hlendes, etc.," since the English "narrative," whether noun or  adjective, does not, as may the German "Erz&hlung," etc., imply an appeal to the  memory or recollection. Blase points out (Kritik, p. 12) that I misunderstood the  Latin equivalents narratio, etc., as employed by Rodenbusch (De temporum usu  Plautino, Strassburg, 1888) who thus translates this peculiar German "Erzahlung"  into Latin. My error may seem pardonable under the circumstances.     366 Abthub Leslie Wheeler   This elevates the descriptive power of the imperfect to a higher  position than seems to me justified, unless one defines all cases  having the progressive function as descriptive which Blase evi-  dently does not do, for he makes separate categories of the  "erzahlendes" (reminiscent) function and, as has been seen, of  the conative, 1 in all of which he recognizes the nature of the action  as progressive.   Again it is to be noted that he speaks of the 'description of  customs,' etc., i. e., he does not regard the use of the imperfect to  indicate customary action as important enough even for a sub-  class, although he makes at least varieties of the reminiscent and  conative uses. I shall take up this point more fully below, 2 merely  remarking here that the cases usually termed customary are fully  as peculiar as those termed by Blase conative and far more  numerous, at least in early Latin.   1 would, then, understand as an imperfect used in description  one which is used in a descriptive passage to present any act or  state vividly to the hearer or reader. What Blase's conception is,  I can not discover. He appears to make a distinction (Kritik,  p. 7) between "Erzahlung" 3 (= here "narrative"?) and"Schilde-  rung" ( — description), e.g., in Plautus Bacch. 258-307, Capt.  497-515, Terence Andr. 48ff., 74-102 — passages which I had  cited as descriptive, 4 he sees "reine Erzahlung, keine Schilde-  rung." On the other hand, in Terence Phorm. 60-135, which I  had also cited, he sees "eine Erzahlung mit einzelnen Situations-  malereien." Without quibbling over our characterization of the   i "Conative" is used in this passage merely as representing Blase's classification.   2 With regard to Blase's peculiar distinction between imperfects in dependent and  independent clauses I would remark that in the study of probably two or three thousand  cases of the tense I have never been able to see any essential difference in function due  to the presence of a case in a dependent clause, cf . Am. Jour. Phil. XXIV, p. 166, n. 1.  And certainly customs, etc. ("Sitten, Gebrauchen") maybe described in a subordinate  clause as well as in an independent clause.   sif " Erzahlung " is here used by Blase in its technical sense as explained on p. 365,  note, my objections are strengthened, for there is certainly no special "appeal to  recollection" in the imperfects of these passages. One might as well say that the  descriptive presents and infinitives (so-called historical) in the Bacchides passage,  etc., are different from the same usages in, say, Livy, because here the speaker is  supposed to be telling of personal experiences, which is chronologically impossible in  Livy's case.   4 Some of the imperfects are primarily customary.     Imperfect Indicative in Early Latin 367   passages in question let us consider the main point, so far as it  can be discerned in Blase's discussion: that there is to him some  difference between the imperfects in the first group of passages  and those in the Phorm. 60-135. With his characterization of  the latter passage I agree, and I had classified the imperfects in  it as imperfects used in description ("Situationsmalereien"). 1  But what is the difference in the effect of imperfects in this pas-  sage and those in the Bacchides or those, to take a typical passage  from Blase's Tempora und Modi, in Caesar Bell. civ. i. 62. 3 ?  I give the essential parts of the three passages:   Phorm. 80 if. : hie Phaedria continuo quandam nactus est puellulam  .... hanc amare coepit . . . . ea serviebat lenoni .... neque quod  daretur quicquam .... restabat aliud nil nisi oculos pascere, .... nos  otiosi operant dabamus 2 .... in quo discebat ludo exadvorsum ilico  tonstrina erat quaedam, etc.   Bacch. 279 flf . : dum circumspecto, atque ego lembun conspicor ....  is erat communis cum hospite et praedonibus .... is ... . nostrae  navi insidias dabat. occepi ego opservare .... interea nostra navis  solvitur .... homines remigio sequi, navem extemplo statuimus ....   Caesar Bell. civ. i. 62. 3 (in which Blase expressly characterizes nun-  tiabatur, etc., reperiebat as " schildernde," cf . Syntax III, p. 147): Caesar  .... hue iam reduxerat rem, ut equites, etsi difficultate, .... fiebat,  possent tamen .... flumen transire, pedites vero ad transeundum  impediuntur. sed tamen eodem fere tempore pons in Hibero prope  effectus nuntiabatur, etc.   To me there is no difference between the imperfects in the  passages of the Phormio and Bellum civile, on the one hand, and  those of the Bacchides, Captivi, and Andria on the other. All  seem to me to be progressive imperfects in description, some are  also customary (see the collection) and have been classified  under that head as the more important element. Is it not better  to separate such cases as Phorm. 87 operant dabamus, 90 sole-  bamus from the progressive-descriptive types than to group all  together, 3 as is done by Blase?*   1 This term refers to the imperfects, I suppose, though Blase does not specify  exactly what he means.   2 Primarily customary.   3 Blase apparently takes a similar view of the frequentative imperfect; cf. Kritik,  p. 7 and see below.   4 In his Kritik, p. 7 Blase attempts to refute my assertion that the words of Quad-  rigarius are not exactly given by Gellius ix. 11 by pointing to the words of Gellius : ea res     368 Arthur Leslie Wheelek   The usage termed by Blase "erzahlendes," for which I have  proposed in English the term "reminiscent," seems to me to be  closely related to the so-called descriptive imperfect. Blase not  only considers this an important variety {Syn. Ill, pp. 145-47),  but is inclined to regard it as perhaps an original function. 1  According to his definition {Syn., loc. cit. after Delbriick) the  imperfect is thus used "wenn der Sprechende etwas aus seiner  personlichen Erinnerung mitteilt oder an die personliche Erinne-  rung des Angeredeten appelliert." Both the descriptive and  reminiscent uses, therefore, result from the use of the progressive  function to represent a past act vividly. The reminiscent effect  is due to the fact that in this usage the past acts are restricted to  those which concern the personal experience of the speaker or  hearer; it is a more intimate usage. As clear cases I cite from  Blase's list: Cicero Rep. iii. 43; ergo ubi tyrannus est, ibi non  vitiosam, ut heri dicebam, sed ut nunc ratio cogit, dicendum est  plane nullam esse rem publicam. Here Cicero clearly indicates  that he is repeating the substance of his own words of the day  before = " as I was saying yesterday, let me remind you." 2 So  Catullus 30. 7: eheu quid faciant, die, homines, cuive habeant  fidem ? certo tute iubebas animam tradere, inique, me ....  idem nunc retrains te, etc., where the poet reminds his friend (?)  of the latter's advice. In both cases the progressive force is  clear, and, as Blase says, the tenses stand in no clear temporal  relation to any preterite in their context. Now since the peculiar   .... sicpro/ecfoest in libris annalibus memorata. But profecto refers to the content,  not to the exact, words of the passage in the libri annates. And when Gellius gives a  word-for-word citation, he introduces it by more definite language, cf . ix. 13. 6 verba  Q. Olaudii .... adseripsi. In ix. 11 he is almost certainly paraphrasing, cf. haut  quisquam est. nobilium scriptorum, and in libris annalibus. This is the opinion of  Hertz, who prints this passage in ordinary type. The name of Quadrigarius is not  given, but Gellius was probably taking the substance of the account from him. I  have excluded this passage from the certain remains of early Latin.   iKritik, p. 15: "War die vorliterarische Periode des Lateinischen ahnlich der des  Alt-Indischen (vgl. Delbruck, p. 272) und des Alt-Griechischen (Brugmann Gr. Or. s ,  § 539. 2), so haben wir in den Resten des erzahlenden Gebrauchs ebenfalls eine uralte  Verwendung zu sehen;" cf. pp. 49 f.   2 The English imperfect is employed in the same way, e. g., " The facts are as fol-  lows, as I was saying yesterday," or in vulgar expressions like " Warn't I tellin' ye?"  Usually the time is denned by some adverb as by heri in Cicero. Notice, too, the  contrast between past and present as expressed in both passages by nunc.     Impebfect Indicative in Early Latin 369   appeal to recollection is the distinguishing feature of this remi-  niscent imperfect, it would seem proper to confine the usage to  those cases in which such an appeal is clear. Without discussing  doubtful cases I content myself with indicating those found in  Blase's lists which seem to me clearly not reminiscent. Plautus  Tri. 400: sed aperiuntur aedes quo ibam 1 (an immediate past of  the interrupted type). In the same category I would place  Cicero Att. i. 10. 2: quod ego etsi mea sponte ante faciebam, eo  nunc tamen et agam studiosius et contendam — -except that here  the action of faciebam is not interrupted, but is continued in the  present, cf. agam et contendam. Other immediate pasts are Ovid  Fasti i. 50: qui iam fastus erit, mane nefastus erat; ibid. 718:  si qua parum Komam terra timebat, amet; ibid. ii. 79: quern  modo caelatum stellis Delphina videbas, is fugiet visus nocte  sequente tuos (notice modo) ; ibid. 147: en etiam si quis Borean  horrere solebat, gaudeat; a zephyris mollior aura venit. Varro  R. r. iii. 2. 14: libertus eius, qui apparuit Varroni et me absente  patrono accipiebat, in annos singulos plus quinquagena milia e  villa capere dicebat. Here accipiebat seems simply progressive  and (also against Blase) contemporaneous with vidi just above.  dicebat is difficult and may, as Blase says, be reminiscent ; cf . the  exact details given by the speaker ; or did the phrase in annos  singulos influence the choice of the tense ? So in Cic. Off. i.  108 : erat in L. Crasso, .... multus lepos . . . . ; 109 : sunt his  alii multi multum dispares .... qui nihil ex occulto, nihil de  insidiis agendum putant . . . . ut Sullam et M. Crassum vide-  bamus, the imperfect seems to be progressive used in description.  In Ovid Fast. viii. 331: et pecus antiquus dicebat 'Agonio'  sermo, the imperfect seems to be customary; cf. antiquus and  Paulus s. v. Agonium: Agonium dies appellabatur quo rex  hostiam immolabat; hostiam enim antiqui agoniam vocabant.   But however much the interpretation of single cases may vary,  this is clear: the progressive force is discernible in all these cases.  It would be better, therefore, to content ourselves with this and  not to discover an additional appeal to recollection, unless such  force is perfectly clear, since the real imperfect function is not  altered whether the reminiscent force be present or absent.   lOf. p. 359.     370 Arthur Leslie Wheelee   One more remark needs to be made concerning the remini-  scent imperfect. This category has served as a convenient catch-  all for many cases of the imperfect which are difficult to classify  and especially for those in which it is difficult or impossible to  discern any progressive force, many of which I have classified as  aoristic. To classify these last cases as reminiscent is doubly  wrong ; first, because it usually involves a petitio principii, i. e. ,  an effort to discover imperfect function because the form is  imperfect; secondly, because the reminiscent coloring is con-  nected only with instances in which the imperfect (progressive)  function is clear. The shadowy appeal to memory does not exist  as a separate function. 1   THE IMPERFECT OF CUSTOMARY PAST ACTION   It has already been pointed out that Blase would not elevate  this variety of the progressive imperfect to the dignity of a sub-  class. The tense, however, occurs so often in the expression of  custom, habit, method, etc., that it seems to me worthy of sepa-  ration from other varieties of the progressive. In early Latin  I have counted about 450 instances in which the customary  coloring seems tome the most prominent element (see the table).   Blase (Kritik, p. 9) has objected to my statement ( Am. Jour.  Phil. XXIV, p. 176) that verbs whose meaning implies repe-  tition (vocito) or even custom (soleo) are especially well adapted  to the expression of the customary past function. He gives no  reason with regard to the first group, vocito, etc., where the mean-  ing is connected with the form. With regard to soleo, etc., he  says only that the reciprocal influence of verb-meaning and tense-  function appears "nicht nachweisbar, da doch der Verfasser  selbst ihr seltenes Vorkommen im Imperfekt natiirlich findet,  weil sie in jedem Tempus der Vergangenheit 'the customary  past function' ausdrucken." There appears here to be some mis-  understanding on Blase 's part and perhaps my statement was too  brief. I did not mean by reciprocal influence of verb-meaning  and tense-function that the tense borrows anything, as Blase  seems to understand me, from the meaning of the verb, but that  when a verb whose meaning implies repetition or custom occurs   i See p. 378 for further remarks.     Imperfect Indicative in Eaely Latin 371   in the imperfect tense, the expression of custom becomes especially  clear. The meaning of the verb and the function of the tense  are mutually helpful to the expression of the thought. 1 Verbs  like appello, voco, vocito, dico (="name") imply not merely a  single act of naming, but usually many acts at intervals. 2 There  are numerous instances of such verbs in the imperfect (see the  collection) and nothing seems to me to be clearer than that these  verbs are especially well adapted to the expression of custom — •  past, present, or future. If we compare Varro, M. r. i. 17. 2:  iique quos obaeratos nostri vocitarunt with id. L. L. v. 162: ubi  cenabant, cenaculum voeitabant, etc., we see that in the first case  the tense merely states, while the verb-meaning, together with  the context, gives the idea of custom or habit; in the second  (voeitabant) the verb- meaning is reinforced by the imperfect  tense — both aid in the expression of custom. This does not mean  that a Roman more often used the imperfect tense of such verbs  when he wished to express custom, but that when the imperfect  was used, a clearer expression of customary past action resulted. 3  As to soleo, consuesco, etc., the same principle holds, for cus-  tom and repetition are inseparably connected; but since these  verbs imply by their meaning the very function (custom) in  question, it is clear that the imperfect tense would occur more  rarely. When, however, the imperfect was used, there was, just  as in vocito, etc., a more emphatic expression of the customary  idea; cf. Phorm. 90: Tonstrina erat quaedam: hie solebamus  fere plerumque earn opperiri .... Here tense, verb, and particles  all lend their aid to the expression of the idea of custom or habit.  The same idea would have been expressed less clearly by hie fere  plerumque opperiebamur, or by hie fere plerumque soliti sumus  opperiri, or by hie opperiebamur. In the last form only does the   i Cf . Trans. Am. Philolog. Ass. XXX (1899), p. 19, where I first expressed this  view. That verbs like soleo "dominate the tense" (ibid.) I no longer believe; they  aid the tense, but it is impossible to say whether the tense or the verb-meaning is  more influential in the total effect. Cf. also Morris, Principles and Methods in  Syntax, 1901, p. 72.   2 If the intervals are very close together without the implication of custom, I  would classify as frequentative ; see below.   3 Am. Jour. Phil. XXIV, p. 177, n. 3, and the dissertation of Miss Perkins cited  above, p. 365.     372 Arthur Leslie Wheeler   tense-form become entirely dissevered from the influence of verb-  meaning and accompanying particles, and even here context is  operative. 1   THE IMPERFECT OF FREQUENTATIVE OR ITERATIVE ACTION   The progressive function inherent in all true imperfects renders  the tense well fitted to express repetition in the past. The repeated  acts may naturally occur at wider or narrower intervals, as the case  may require. All expressions of custom, for example, involve an  idea of repetition, but it is only to cases of the imperfect which indi-  cate an act as repeated insistently, usually at intervals very close  together, that I would give the title "frequentative" or "iterative,"  i. e., imperfects in which this element of repetition becomes more  prominent than any other. It seems to me that the existence of  a few such cases in early Latin is not fanciful. In Plautus'  Captivi, line 917 : aulas .... omnis confregit nisi quae modiales  erant: cocum percontabatur, possentne seriae fervescere, 2 a single  situation is described wherein the parasite repeatedly and insist-  ently asked, kept asking, whether, etc. There is something more  than mere progressive force, on the one hand, and there is no  idea of habit or custom, on the other. The primary element of  the tense is here repetition. When, therefore, Blase sees in As.  207 ff. repetition, he is right, for repetition in a general way is  present in all cases of the customary imperfect; but he is wrong  in viewing repetition as the more important element. The more  important element seems to me custom and in accordance with  this we ought to classify these cases as customary. 3   iln a review of Miss Perkins' dissertation Woch.f. kl. Phil., 1904, cols. 1277-80,  Blase has since admitted the truth of my assertion with regard to the influence of  verb-meaning: "Die Verbalbedeutung ist massgebend z. B.bei alien Verben, die  'nennen,' 'benennen,' bezeichnen, wie appellare dicere vocare, denn der Name  entsteht durch ein gewohnheitsmassiges Nennen. Damit ist der Grand gegeben (by  Miss Perkins) fur eine Behauptung, die ich .... bei Wheeler bezweifelt habe."   2 Blase (Kritik, p. 10) misses among my cases Rud. 540, which was nevertheless  cited, but escaped him because by a misprint the imperfect was not italicized. On the  same page he cites ten passages and says that I "hier uberall gewohnheitsmftssige  Handlungen erkenne." This is very inaccurate, unless "hier" refers to the last two  passages, As. 207 ff., Bacch. 424 — the only two of the list which I have classified as  customary. My classification of the other eight passages may be seen by referring to  the collection at the end of this paper.   3 Blase (Kritik, pp. 9, 10) seems to imply that I have said that the frequentative  imperfect is commoner in later Latin. I have nowhere said this and my statement,     Imperfect Indicative in Early Latin 373   the aoristic imperfect  Excessive deference to the principle that a difference of form  implies a difference of meaning and the well-known tendency of  investigators to abhor an exception are chiefly responsible for the  unwillingness of some scholars to admit that the imperfect occurs  in Latin with no progressive force, i. e., as an aorist. While I  can not pretend to criticize this method as applied to Sanskrit and  Greek by Delbruck, 1 it seems to me that there are reasons against  its application, in the same degree at least, to Latin. The situa-  tion in early Latin differs essentially from that in Sanskrit and in  Greek. In the first place there is no 'great mass' 2 of cases of the  imperfect in which real progressive force is not discernible, and  the cases (about sixty) are restricted almost entirely to two verbs,  aibam and eram. This seems to indicate that the phenomenon  arose on Latin ground alone and has its explanation in some  peculiarity of the few verbs concerned. Again the greater wealth  of tenses in Sanskrit and Greek would lead us a priori to expect   Am. Jour. Phil. XXIV, pp. 186, 187, "Latin seems .... to have been unwilling to  take that step," implies the opposite belief. When I added (ibid., p. 187), " If the fre-  quentative imperfect in early Latin is still in its infancy, etc.," it was naturally not  implied that it ever passed out of its infancy ! The facts in later Latin are not known  because they are not collected. Wimmerer naturally repeats from Blase's Kritik both  these errors ( Wien. Stud., 1905, p. 263). He, too, is of the opinion that it is of no ad-  vantage to separate so-called iterative imperfects from those of customary nature:  " wenn doch in jedem Falle erst auf Grund des gewahlten Tempus aus dem Zusam-  menhange erkannt wird, dass es sich um eine Gewohnheit handelt." To this it must  be answered, first, that it is by no means always, and often not at all, on the basis of  the tense that we recognize the presence of customary action. Such action may be  expressed in many ways, the tense being but one element ; and, secondly, if the cases  interpreted by me as frequentative are really essentially different from any other  variety of the progressive, then they should be classified separately, at least until it can  be proved that they belong elsewhere.   1 It will suffice to quote two of Delbruck's statements. He says of the Greek tenses :  "Man muss sich eben mit der Erwagung begnugen, dass es einem Schriftsteller bald  gut schien, zu konstatieren, bald zu erzahlen, ohne dass wir uns seine Motive immer  klar machen konnten" (Vergl. Syn. II, p. 304, cf. pp. 302, 303). A saner. method is  evinced ibid., p. 304 : " Den Unterschied zwischen Perfekt und Imperfekt (of Sanskrit)  in den einzelnen Stellen nachzuweisen, sind wir nicht mehr im Stande." This is at least  safe agnosticism, biding its time until the lost distinctions shall be found. Blase is  in entire agreement even as regards Latin with the first statement of Delbrflck, cf .  Kritik, p. 12.   2 Delbruck (ibid., p. 304, of Greek) : "Aber .... bleibt doch auch eine grosse  Menge von Stellen ubrig, bei denen wir einen Grund fur die Wahl des Tempus nicht  ausfindig machen konnen."     374 Arthur Leslie Wheeler   in those languages a larger number of instances in which it is  hard to differentiate similar tenses, whereas the much narrower  tense-system of Latin exhibits a tendency to merge the functions  of similar tenses, cf. the perfect in -v- with the reduplicated per-  fect and the formally aoristic perfect in -s-. In accordance with  this preliterary development we should expect indications of the  same tendency in the literary period. The aoristic imperfect is,  I believe, an illustration of this tendency, resulting from the  merging of the functions of imperfect and preterite (aorist) in  certain verbs. The restricted range of the phenomenon and its  probable explanation (see below) would make it unlikely that we  are here dealing with a survival of an Indo-European confusion.   As illustrations of the aoristic usage I will cite : Plautus Poen.  1069 : nam mihi sobrina Ampsigura tua mater fuit (cf. fecit 1071),  pater tuos is erat frater patruelis meus. Here there seems to be  no difference between erat and fuit. Ibid. 900: et ille qui eas  vendebat dixit se furtivas vendere: ingenuos Carthagine aibat  esse, where aibat and dixit seem to be equivalent. For other cases  see the collection.   It is quite possible that others may be able to detect true im-  perfect force in some of the cases which I have classified as aoristic.  Blase, though not quite certain of his own classification, has con-  vinced me that I may have been wrong with regard to Varro H. r.  ii. 4. 11: in Hispania ulteriore in Lusitania .... sus cum esset  occisus, Atilius Hispaniensis minime mendax .... dicebat ....  L. Volumnio senatori missam esse offulam cum duobus costis,  etc. There are so many exact details here that we suspect  Scrofa of reminiscing. So possibly Varro ibid. iii. 17. 4 dice-  bat. 1 But though perhaps a dozen 2 cases might be taken from the  total of those which seem to me aoristic, enough remain to establish  this category on a firm basis.   The exact process by which the progressive function became  lost can not, of course, be proved. I have suggested (Am. Jour.  Phil. XXIV, pp. 188, 189) that it is a weakening due to the constant   'Blase is quite right (Kritik, p. 11) in classifying As. 208 aibas as customary. I  neglected to exclude this from four cases cited from Rodenbusch. It was classified on  my own slips as customary.   2 1 have indicated in the collection those which seem to me questionable.     Imperfect Indicative in Early Latin 375   use of certain verbs in ever-recurring similar contexts, until  in the case of aibam the originally graphic ' force was used out of  the form and aibam became a mere tag to indicate an indirect  discourse. 2 With eram the vagueness of the verb-meaning and  the frequency of its occurrence side by side with fui were the  chief influences. In contexts where there are many other imper-  fects all of a definite time, these usually colorless verbs naturally  take the prevailing color 3 of the context; cf. As. 208 aibas.   In his "Tempora und Modi" (Syn. Ill, p. 145) Blase expresses  his belief that an aoristic imperfect as accepted by Luebbert and  J. Schneider has been proven not to exist by E. Hoffmann (Zeit-  partikeln 2 , pp. 181 ff . ) . But neither Luebbert nor Schneider seems  absolutely to have believed in an aoristic usage. 4 Luebbert says  (Quom, pp. 156 ff.) that in Men. 1145 and 1136 ff. we find aoris-  tic perfect and the imperfect, etc. "promiscue gebraucht da der  Unterschied zwischen beiden gering war." "Grering" indicates  that there was to him some difference, even though it was slight.  Schneider's statements are not consistent. In his De temporum  apud priscos scriptores latinos usu quaestiones selectae, Glatz,  1888, pp. 14 ff., he says correctly that in many cases no difference  can be seen between aibat and dixit, and that "aibat aoristi  munere fungi," but he adds that the imperfect represents an act  as "infectam ideoque aliter intellegendam acsi perfectam." Hoff-  mann's supposed refutation is very weak. In the first place he   1 If originally reminiscent, the explanation is the same ; for the reminiscent usage  is due to the speaker's effort to represent a past act graphically.   2 Cf. Am. Jour. Phil. XXIV, p. 188, where it is stated that the indirect discourse is  always present or implied (rarely) with aibam. Occasionally the object is represented  by a pronoun. Bacch. 982: quid ait?, Capt. 676: ira vosmet aiebatis itaque, etc.   8 Cf. Blase (Kritik, p. 11): "wo aibam mitten zwischen Imperfekta der wieder-  holten oder gewohnheitsmassigen Handlung steht und unmdglich anders gef asst werden  kann."   4 But cf. O. Seyffert in Bursian's Jahresb. LXIII, p. 32: " Das Imperf. findet sich.  bekanntlich bei den Scenikern mehrfach in einem so geringen Bedeutungsunterschiede  vom Perf . und bisweilen unmittelbar neben demselben, dass man ohne wesentliche  Anderung des Sinnes und oft auch unbeschadet des Metrums (Rud. 543, Capt. 717) das  eine Tempus f iir das andere einsetzen kann. Es zeigt sich dies besonders bei den verba  dicendi; das Imperf. von aio vertritt ja geradezu das fehlende Perfect;" cf. ibid.  LjXXX, p. 336, where Seyffert repeats the statement that aibat, e. g., Ps. 1083, represents  the lost perfect of aio. In Am. Jour. Phil. XXIV I had overlooked this remarkable  anticipation of my own conclusions.     376 Arthur Leslie Wheeler   confuses different uses of the tense, asserting, for example, that in  Plautus Tri. 400: aedes quo ibam, etc., the imperfect is wholly  analogous to that in Tacitus Ann. ii. 34: simul curiam relinquebat.  commotus est Tiberius, etc. ; cf. iv. 43 sequebatur Vibius Crispus,  donee, etc., and that in the last two cases the imperfect jars on us  because such an action is not usually presented "in der Phase ihres  Vollzugs." Such an application of the tense may seem strange to  a German, but to one who speaks English, it is entirely natural  and could not for a moment be mistaken for anything but a  simple progressive imperfect. To refute such a usage as a supposed  aorist is to knock down a man of straw. The supposed analogy of  these cases to Tri. 400 does not bear on the point, but it may be  remarked that ibam is analogous only in the fact that its action is  progressive and interrupted, but it belongs to the immediate past  type. 1 Hoffmann then cites ten cases of aibat, six of which may  be taken aoristically, and asserts that the tense is in all used  "in voller Gesetzmassigkeit." This assertion rests on entirely  inadequate foundation. 2   the shifted imperfect  Blase seems right in restricting the 'shifted' imperfect to one  class (Kritik, pp. 13, 14) = an imperfect subjunctive with present  meaning; for, as he says, there is no real shifting if the  preterital sense remains. But when he adds 3 that "ein sicherer  derartiger Fall ist weder bei Plautus und Terenz, noch sonst im  Altlatein vorhanden," I can not agree. He accepts as cases  of shifting Varro, L. L. viii. 65: sic Graeci nostra senis casibus  .... dicere debebant, quod cum non faciunt, non est analogia,  and ix. 85: si esset denarii in recto casu .... tunc in patrico  denariorum dici oportebat, and ix. 23: si enim usquequaque esset  analogia, turn sequebatur, ut in his verbis quoque non esset, non,   1 See above, pp. 359 ff.   2 J. Ley Vergilianar. quaestion. specimen prius de temporum usu, Saarbriicken,  1877, apparently believes that eram and fui in Vergil are so nearly equivalent that  metrical convenience often decided between them ; cf . Blase Syn. Ill, p. 164 Anm. I  have not seen this dissertation, but the explanation is, on its face, insufficient.   S0f. his Syntax, p. 149: " Der Indikativ des Imperfekts hat erst seit Beginn der  klassischen Zeit eine allmahliche Verschiebung aus der Sphfire der Vergangenheit in  die der Gegenwart erfahren."     Imperfect Indicative in Eably Latin 377   cum esset usquequaque, ut est, non esse in verbis. If these are  real cases of shifting, how do the following differ ? Plautus Merc.  983 e : temperare istac aetate istis decet ted artibus ....  vacnom esse istac ted aetate his decebat noxiis. itidem at tem-  pus anni, aetate alia aliud factum convenit; Mil. 755: insanivisti  hercle (perf. def.): nam idem hoc hominibus sa/[a] era[n]t  decern; ibid. 911: bonus vatis poteras esse: nam quae sunt  futura dicis. 1 If the passages from Varro move in the present  (Blase Kritik, pp. 13, 14), the same is true here; cf. Auct. ad  Herenn. ii. 22. 34: satis eratjiv. 41. 53 infimae (infirmae?) erant. 2  That Varro L. L. viii. 74 oportebat stands "zwischen zwei Per-  fekten" (Blase) is accidental. 3   This peculiar shifting was explained by me Am. Jour. Phil.  XXIV, pp. 189, 190, as due to the unreal (contrary-to-fact) idea  present in the context or in the meaning of the verb (oportebat,  etc.) or in both ; cf. Blase (Syn. Ill, p. 149) who also calls attention  to the auxiliary character 4 of the verbs involved and thinks that  the shifting began with verbs of possibility and necessity which  seems a probable view.   In conclusion a few words are necessary with regard to some  general aspects of the subject and its method of treatment. The  original function or functions 5 of the imperfect can not, of course,  be certainly inferred from a syntactical investigation of material  which is relatively so late even with the aid of etymology and  comparative philology. My statement (loc. cit., p. 184) that the  progressive function was probably original was therefore intended   i Cf. Rud. 269 aequius erat, True. 511 poterat, Aul. 424. For the other eases  see collection.   2 But not iv. 16. 23, which I now see is not shifted.   8 And both are cases of debuerunt! In his Kritik, p. 13, Blase denies my assertion  (loc. cit., p. 181, n. 1), that the perfect indie, and the perfect infin. of these verbs are  shifted in Varro, cf . L. L. viii. 72-74 ; viii. 48 ; viii. 50 ; viii. 61, 66. I am glad to find my  view supported by Wimmerer Wien. Stud., 1905, p. 264 : " Denn da der Grund der Ver-  schiebung hier vor allem in der Bedeutung der Verba liegt, so kann konsequenterweise  ebenso gut ein debuit wie ein debebat verschoben werden."   «Cf. Am. Jour. Phil. XXIV, p. 190.   6 It is uncertain whether the original meaning of the tense was vague, admitting  several uses which gradually became narrowed to one (the progressive), or whether  there was one original meaning which split into several related uses. The facts seem  to point to the second alternative.     378 Arthur Leslie Wheeler   only as a probability based upon the existence of this force in  nearly all the cases and upon the generally accepted etymology of  the imperfect form. But nothing like proof was claimed for this  theory. Blase is inclined, following Delbrtick and Brugmann, to  regard the reminiscent usage also as an original one (cf. p. 26,  n. 2), but he rightly says that no statistics can prove which of  these two is earlier. If my view that the reminiscent usage is  rather an application of the progressive than itself a separate  function is correct, then the progressive is older. The existence  of the reminiscent imperfect in Sanskrit and Greek certainly  makes it very probable, as Blase says, that it existed in preliterary  Latin also. If this is so, I am inclined to refer it to the same  general origin as the so-called descriptive imperfect — to the effort  to present a past act (here a personal experience) vividly. 1   But the search for original meanings must ever remain within  the realm of theory; nor can we hope even theoretically to reach  any considerable degree of probability in the establishment of  such meanings without the most careful collection and classifica-  tion of the facts within the period of written speech. And this  should precede the appeal to etymology and comparative phi-  lology. What is actually found in any given language, not what  according to comparative philology ought to be found, should be  our first aim. Although I would not minimize the importance in  syntactical study of the comparative method, it seems to me prop-  erly applied only as a supplement, not as the controlling factor  to which all else is subordinated. Indeed, a premature appeal to  comparative philology may result in premature conclusions,  for an investigator whose head is filled with preconceived notions  drawn from Sanskrit and Greek is all too apt to imagine peculi-  arities in Latin phenomena which he would not have perceived  at all, had he approached by a Latin route alone; and such  peculiarities have little value unless they can be recognized as  Latin without foreign assistance. Once recognized they may,  and often do, receive much additional light from comparative  philology. While it is true, then, that scholars will differ with   •Cf. Am. Jour. Phil. XXIV, pp. 185, 190, where it was surmised that the descrip-  tive application of the tense was Indo-European.     Imperfect Indicative in Early Latin 379   regard to a few cases' in any given syntactical phenomenon and  the ultimate classification must not neglect the aid of comparative  philology, yet the chief basis of investigation is agreement among  scholars with regard to the great majority of such cases viewed  as purely Latin phenomena. If this agreement is lacking,  comparative philology can rarely bring reliability to the results.   COLLECTION   The statistical table shows that this investigation is based upon  a collection of 1,223 imperfects. It has been my aim to exclude  from consideration (and from the table) all passages of dubious  authorship, corrupt text, or insufficient context. About 170 cases  have thus been excluded, a seemingly large proportion, but it  must be remembered that much of the literature of the third and  second centuries before Christ is fragmentary and very often  there is not enough context to render classification at all certain.  In so large a body of text it is probable that some cases have  escaped my notice, but most of the ground has been examined at  least twice and such omissions can hardly be numerous or alter  essentially the results. I have subjected the material to a careful  revision and the table differs slightly from that published in  Am. Jour. Phil. XXIV, p. 183.   It would seem unnecessary nowadays for any syntactical  scholar to state that he lays no stress on statistics as such, but  when a reviewer 2 attributes to me the conviction that I have  proved this and that by just so many exact figures, it seems  proper for me to disclaim any such conviction. The fact that  exact figures do not in themselves mean anything does not,  however, excuse one from being as exact as possible.   iCf. Wimmerer Wien. Stud., 1905, p. 262: "die syntaktischen Einzeltatsachen sind  viel zu sehr umstritten . . . . als dass auf sie allein eine brauchbare Klassiflkation  und Erkl&rung der Arten eines einigermassen verzweigten syntaktischen Gebrauches  gesttizt werden kdnnte." With this I agree, except possibly as to what is a "brauch-  bare Klassiflkation," but when he says (p. 61), with reference to my inference that the  progressive function is original: "Den Begriff aber hat die vergleichende Sprach-  wissenschaft langst festgestellt," I would suggest that such a conclusion could not be  regarded as 'firmly established' except with several investigations like mine as chief  ies.   2 In Archiv.f. lat. Lex. und Gk. XIV, p. 289.     380 Abthuk Leslie Wheelee   The method of citation adopted in the collection will doubtless  seem to many inadequate. It is especially true, however, of the  classification of tense functions, that very often a large body of  context must be taken into consideration. For this reason very  many of the citations even in Blase's "Tempora und Modi" are  quite useless and misleading because of their brevity. It seemed  best, therefore, to cite as fully as possible in the body of the  article, but in the collection to cite only each form and the place  of its occurrence. Those who are interested in examining a given  usage in detail will in any case revert to the complete context, as  I know by experience.   I. Progressive Imperfect  A. Simple Types, including imperfects in description, reminiscence,  and the "immediate" past variety.  Plautus, ed. Goetz and Schoell, ed. minor, Lipsiae, 1892-96.  Amph. prol. 22 scibat; 199 pugnabant .... fugiebam; 251 com-   plectabantur; 383 aiebas; 385 sci[e]bam; 429erat; 597 credebam;   603 stabam; 711 solebas; 1027 censebas; 1067 confulgebant;   1095 rebamur; 1096 confulgebant. 14   As. 300 scibam; 315 mirabar; 385 censebam; 392 volebam; 395   volebas; 452 volebam; 486 volebas; 888 suppilabat; 889 suspi-   cabar .... eruciabam; 927 ingerebas .... eram; 931 dissua-   debam. 13   Aul. 178 praesagibat .... exibam; 179 abibam; 295 poterat; 376   erat; 424 aequom .... erat; 427 erat; 550 meditabar; 625   radebat .... croccibat; 667 censebam; 707 expectabam ....   abstrudebat; 754 scibas; 827 apparabas. 15   Bacch. 18 (frag, x) erat; 189 volebam; 282 erat; 286 dabat; 297   dabant; 342 censebam; 563 erat; 675 sumebas; 676 nescibas;   683 suspicabar; 788 orabat; 959 restabant; 983 auscultabat   .... loquebar. 14   Capt. 273 erat; 491 obambulabant; 504 eminebam; 561 aibat;   654 assimulabat; 407 audebas; 913 frendebat. 7   Cas. 178 ibam; 279 aiebat; 356 rebar; 432 trepidabant .... fes-   tinabat; 433 subsultabat; 532 erat; 578 praestolabar; 594 ibam;   674 volebam; 702 volebam; 882 erant; 913 erat .... erat   .... erat. 15   Cist. 153 poteram; 187 exponebat; 566 perducebam; 569 adiura-   bat; 607 ai[e]bas; 623 properabas; 721 rogabat; 723 quaeritabas;   759 quaeritabam. 9   Cure. 390 quaerebam; 541 credebam. 2     Imperfect Indicative in Early Latin 381   Epid. 48 amabat; 98 solebas; 138 desipiebam ; .... mittebam; 214  occurrebant; 215 captabant; 216 habebant; 218 ibant; 221 prae-  stolabatur; 238 dissimulabam ; 239 exaudibam .... fallebar;  241 ibat; 409 apparabat; 420 adsimulabam; 421 me faciebam.  482 deperibat; 587 vocabas; 603 dicebant; 612 aderat. 20   Men. 29 erant; 59 erat; 63 ibant; 195 amabas .... oportebat;  420 advorsabar; .... metuebam; 493 eram; 564 ferebat; 605  censebas; 633 negabas; 634 negabas .... ai[e]bas; 636 cense-  bas; 729 negabas; 773, 774 suspicabar; 936aiebat; 1042ai[e]bat;  1046 aiebant; 1052 ferebant; 1053 clamabas; 1072 censebam;  1116 cadebant; 1120 eramus; 1135 erat .... vocabat; 1136  censebat; 1145 vocabat. 28   Merc. 43 abibat; 45 rapiebat; 175 quaerebas; 190 abstrudebas;  191 eramus; 197 censebam; 212 credebat; 247 cruciabar; 360  habebam; 754 obsonabas; 815 censebam; 845 erat .... quae-  ritabam; 884 ibas; 981 ibat. 15   Miles 54 erant; 100 amabant; 111 amabat; 181 exibam ....  erat; 320 ai[e]bas; 463 dissimulabat; 507 osculabatur; 835 cale-  bat .... amburebat; 853 erat; 854 erat; 1135 exoptabam; 1323  eram .... eram; 1336 temptabam; 1140 erat; 1430 habebat. 18   Most. 210 quaerebas; 221 su<b>blandiebar; 257 erat; 787 erat;  806 aiebat; 961 faciebat. 6   Persa 59 poterat; 171 censebam; 257 somniabam .... opinabar;  .... censebam; 262 erant; 301 cupiebam; 415 censebam; 477  credebam; 493 occultabam; 626 pavebam; 686 metuebas. 12   Poen. 391 dicebas; 458 sat erat; 485 accidebant; 509 scibam; 525  properabas; 748 dicebant; 899 vendebat; 1178 aderat; 1179  complebat; 1180 erat; 1231 volebam; 1391 expectabam. 12   Pseud. 286 amabas; 421 subolebat; 422 dissimulabam; 492 nole-  bam; 499 scibam; 500 scibas; 501 mussitabas .... scibam;  502 aderat .... aberat; 503 erat .... era<n>t; 677 habebam;  698 arbitrabare; 718 ferebat; 719 accersebat; 799 conducebas  .... erat; 800 sedebas .... eras; 912 circumspectabam ....  metuebam; 957 censebam; 1314 negabas. 24   Kud. 49 erat; 52 erant; 58 erat; 222 oblectabam; 307 exibat; 324  suspicabar; 378 scibatis; 379 amabat; 452 censebam; 519 age-  bam; 542 aiebas; 543 postulabas; 600 quibat; 841 erat; 846  sedebant; 956a faciebat; 9566 fiebat; 1080 aiebas; 1123 pete-  bas; 1186 credebam; 1251 monstrabant; 1252 ibant; 1253 erat;  1308 erat. 24   Stich. 130placebat; 244praedicabas; 328 visebam; 329 miserebat;  365 superabat; 390 negabam; 540 erant; 542 erant; 543 erat; 545  erant; 559 postulabat. 11     382 Arthur Leslie Wheeler   Trin. 195 volebam; 212 aiebant;.400 ibam; 657 scibam ....  quibam; 901 erat .... gerebat; 910 vorsabatur; 927 latitabat;  976 eras; 1092 agebat; 1100 effodiebam. 12   True. 164 vivebas; 186 cupiebat; 198 lavabat; 201 celebat metue-  batque; 332 dicebam; 333 revocabas; 648 debebat; 719 eras;  733 dabas; 748 volebas; 757 aibas; 813 erat .... valebat ....  petebat; 921 ibat. 16   Vid. 71 miserebat; 98 piscabar. 2   Fragmenta fabb. cert. 86 sororiabant; 87 fraterculabant. 2   Plautus, IA, Total 291  Terence, ed. Dziatzko, 1884.   Ad. 78 agebam; 91 amabat; 151 taedebat; 152 sperabam; 153  gaudebam; 234 eras; 274 pudebat; 307 instabat; 332 iurabat;  333 dicebat; 461 quaerebam; 561 aibas; 567 audebam; 642  mirabar; 693 credebas; 809 tollebas; 810 putabas; 821 ibam;  901 eras. 19   And. 54 prohibebant; 59 studebat; 60 gaudebam; 62 erat; 63  erat; 74 agebat; 80 amabant; 86 erat; 88 amabant; 90 gaude-  bam; 92 putabam; 96 placebat; 107 amabant .... aderat; 108  curabat; 110 cogitabam; 113 putabam; 118 aderant; 122 erat;  175 mirabar; 176 verebar; 435 expectabam; 490 imperabat;  533 quaerebam; 534 aibant; 545 dabam; 580 ibam; 656 adpar-  abantur; 657 postulabat; 792 poterat; exit, suppositic. I expec-  tabam. 31   Eun. 86 eras; 87 stabas .... ibas; 97 erat; 112 dicebat; 113  scibat .... erat; 114 addebat; 118 credebant; 119 habebam;  122 eras; 155 nescibam; 310 congerebam; 323 stomachabar; 338  volebam; 345 erat; 372 dicebas; 378 iocabar; 423 erat; 432 ade-  rant; 433'metuebant; 514 erat; 533 orabant; 569 erat; 574 cupi-  ebam; 584 inerat; 587 gaudebat; 606 simulabar; 620 faciebat  .... cupiebat; 621 erat; 681 erat; 727 adcubabam; 736 erat  .... nescibam; 743 expectabam; 841 erant; 928 amabant; 1000  quaerebat; 1004 scibam; 1013paenitebat; 1065 quaerebam; 1089  ignorabat. 43   Heaut. 127 faciebant; 200 erat; 201 erat; 256 volebam; 260 can ta-  bat; 293 nebat; 294 erat .... texebat; 308 scibam; 366 tracta-  bat; 445 erat .... erant; 536 oportebat; 629 erat; 758 opta-  bam; 781 dicebam; 785credebam; 844 quaerebam; 907 videbat;  924 aiebas; 960 aiebas; 966 erat. 22   Hec. pro. II. 16 scibam; 91 eram; 94 licebat; 115 amabat; 162  erat; 172 redibat; 178 conveniebat; 230 erant; 283 eram; 322  poteram; 340 eras; 374 dabat; 375 monebat .... poterat; 422  expectabam; 455 agebam; 498 orabam; 538 negabas; 561 aderam;  581 rebar; 651 optabamus; 713 credebam; 806 pudebat. 23     Imperfect Indicative in Eaely Latin 383   Phorm. 36 erat; 51 conabar; 69 erat .... supererat; 83 servi-   ebat; 85 restabat; 88 discebat; 89 erat; 97 erat? 99aderat; 105   aderat; 109 amabat; 118 cupiebat .... metuebat; 298 duce   bat; 299 deerat; 355 agebam; 365 habebat; 468 erant; 472 quae-   rebam; 480 aibat; 490 mirabar; 529 scibat; 570 manebat; 573   commorabare; 582 scibam; 595gaudebat .... laudabat ....   quaerebat; 596 gratias agebat; 614 agebam; 642 insanibat; 652   ven<i>bat; 654 opus erat; 759 volebam .... volebam; 760daba-   mus operam; 797 sat erat; 858 aderas .... aderam; 900 iba-   mus; 902 ibatis; 929 dabat; 945 eras; 1012 erant; 1013 erat;   1023 erat. 47   Terence, I A, Total 185  Cato ed. Jordan, Lipsiae, 1860.   p. 36. 2 sedebant .... lacessebamur. Total 2   Dramatic and epic fragments.  Naevius. Bell, pun., ed. Mueller, 1884.   5 immolabat; 7 exibant; 12 exibant; 65 inerant.   tabular, fragmenta, ed. Ribbeck 3 , 1897-98.   I p. 16 IV habebat .... erat; p. 322 II proveniebant.   II p. 30 VII faciebant .... tintinnabant. 9  Ennius, ed. Vahlen 2 , 1903.   Annal. 28 premebat; 41 videbar; 43 stabilibat; 82 certabant;  87 expectabat; 87 tenebat; 138 mandebat; 139 condebat; 147  volabat; 190 sonabat; 202 solebat; 216 erat; 307 vivebant; 307  agitabant; 309 explebant .... replebant; 343 aspectabat; 408  sollicitabant; 459 parabant; 497 fremebat; 555 cernebant. 21  Scenica. 15 eiciebantur; 123 erat; 127 inibat; 251 petebant; 324  scibas.   Saturar. 65 adstabat.   Varia. 45 videbar; 64 ibant. 8   Pacuvius, ed. Ribbeck 3 1, p. 65 XVI conabar. 1   Accius, ed. Ribbeck 3 , p. 162 V ostentabat; p. 162 VII scibam;   p. 165 VI expectabat; p. 205 X erat; p. 210 XII commiserebam   .... miserebar; p. 213 XX educabant; p. 251 XIII mollibat. 8   Incert. p. 273 V ecsacrificabat; p. 282 XXXII hortabar; p. 285   XLV scibam; p. 304 CI expetebant. 4   Turpilius, ed. Ribbeck 3 II, p. 101 II nescibam; p. 107 V sperebam;   p. 120 X videbar. 3   Titinius, ed. Ribbeck 3 II, p. 168 II aibat. 1   Afranius, ed. Ribbeck 3 II, p. 215 VI hortabatur; p. 217 XII sup-   ponebas. 2   Pomponius, ed. Ribbeck 3 II, p. 303 II cubabat. 1   Incert., ed. Ribbeck 3 II, p. 137 XXIV ferebat simulabat. 2   Dramatic and Epic Fragments, IA, Total 60     384 Arthur Leslie Wheeler   Historicorum fragm., ed. Peter, 1883.   p. 70. 9 nesciebant; 72. 23 erant; 72. 27 cymbalissabat; 72. 27 can-  tabat; 73. 37 mirabantur .... reddebat; 83. 27 apparebat ....  habebat .... sedebant; 94. 13 erat; 110. 7 habebat; 136. 5  erant; 137. 8 concedebat; 137. 8 praecellebat; 137. 10 b antista-  bat; 138. 10 audebat; 138. 11 licebat; 141. 29 erant; 142. 37 erant;  143. 46 captabat; 145. 57 erat .... erat .... sciebant ....  apparebat; 149. 81 mirabantur; 150. 85 sauciabantur .... opus  erat .... defendebant; 178. 8 erat .... tegebat; 178. 9 pot-  erat; 179. 23 indigebat; 184. 79 sciebat; 184. 86 erat.   I A, Total 34  Orator, fragm., ed Meyer, Turici, 1842.  p. 192 narrabat .... poteram; p. 231 existimabam .... arbitra-  bar .... stabant .... erant; 236 ferebantur .... lavabantur.   I A, Total 8  Lucilius, ed. Marx, 1904.  393 stabat; 394 obiciebat; 479 erat; 531 serebat; 534 ibat; 1108  gemebat; 1142 ibat (not in Mueller's ed.); 1174 volebat; 1175  ducebant; 1187 haerebat; 1207 premebat.   I A, Total 11  Auctor ad Herennium, ed. C. L. Kayser, 1854. G. Friederich's text in  C. F. W. Mueller's Cicero, Vol. I, has been compared throughout.  1. 1. 1 intelligebamus .... attinebant .... videbantur; 1. 10.  16 postulabat; 1.12. 21 erat; 1. 13. 23 defendebant .... erant;  2. 1. 2 existimabamus .... ostendebatur; 2. 2. 2 videbatur; 2.  5. 8 faciebat; 2. 19. 28 volebat .... metuebat .... videbat  .... sperabat .... verebatur .... hortabatur .... remove-  bat; 2. 21. 33 erant .... habebat; 3. 1. 1 pertinebant ....  erant .... videbantur; 3. 15. 26 demonstrabatur; 4. 9. 13 pote-  rant .... videbant; 4. 12. 18 inpendebant; 4. 13. 19 ingenio-  sus erat, doctus erat, .... amicus erat; 4. 14. 20 erat; 4. 15. 22  removebas .... abalienabas; 4. 16. 23 damnabant .... ini-  quom erat; 4. 18. 25 erant .... poterant; 4. 19. 26 proderas  .... laedebas .... proderas .... laedebas .... consule-  bas; 4. 20. 27 oppetebat .... comparabat; 4. 24. 33 putabas;  4. 24. 34 habebamus .... habebam .... erat .... obside-  bamur .... videbar; 4. 33. 44 adsequebatur .... profluebat  .... erat; 4. 33. 45 pulsabat .... ducebat; 4. 34. 46 videban-  tur; 4. 37. 49 erat .... oppugnabat; 4. 41. 53 veniebat ....  occidebatur; 4. 49. 62 inibat; 4. 55. 68 faciebat.   I A, Total 62  Corpus Inscr. Lat., Vol. I.  201. 6 animum .... indoucebamus .... scibamus .... arbi-  trabamur.   I A, Total 3     Imperfect Indicative in Early Latin 385   Varro, De lingua Lat., ed. Spengel, 1885.   5. 9 videbatur; 5. lOOerat; 5. 128erat; 5. 147 pertinebat; 7. 39erat;  7. 73 erant; 8. 20 erant; 8. 59 erant. 8   De re rust., ed Keil, 1889,   1. 2. 25 ignorabat .... despiciebat; 1. 13. 6 habebat; 2. 11. 12  ibam; 3.2. lstudebamus; 3. 2. 2sedebat; 3. 13. 2erat .... dice-  bat .... erat .... cenabamus; 3. 5. 18 dicebatur; 3. 16. 3 erat;  3. 17. 1 sciebamus; 3. 17. 9 ardebat. 14   Sat. Menipp., ed. Kiese, 1865, p. 198, 1. 1 regnabat; p. 223, 1. 9  findebat. 2   I A, Total 24   Grand Total, I A, 680   B. Imperfect of Customary Action.   Plautus   As. 142 habebas; 143 oblectabas; 207 arridebant .... veniebam;   208 ai[e]bas; 210 eratis .... erant; 211 adhaerebatis; 212 faci-   ebatis .... nolebam; 213 fugiebatis .... audebatis; 341 sub-   vectabant. 13   Aul. 114 salutabant; 499 erant. 2   Bacch. 421 erat .... eras; 424 accersebatur; 425perhibebantur;   429 exercebant ; 430 extendebant ; 438 capiebat ; 439 desinebat. 8   Capt. 244 imperitabam; 474 erat; 482 solebam. 3   Cist. 19 dabat .... infuscabat; 162 habitabat. 3   Epid. 135 amabam. 1   Men. 20 dabat; 484 dicebam; 715praedicabant; 716 faciebat; 717   ingerebat; 1118 eratis; 1119 eratis; 1122 eratis .... erat; 1123   vocabant; 1131 erat. 11   Merc. 217 credebat. 1   Miles 15 erat; 61 rogitabant; 99 erat; 848 erat; 849 imperabat   .... promebam; 850 sisteba<h>t; 852cassaba<n>t; 855 a com -   plebatur; 856 bacc<h>abatur .... cassabant. 11   Most. 150 erat; 153 victitabam; 154 eram; 155 expetebant; 731   erat. 5   Persa 649 amabant; 824 faciebat; 826 faciebat. 3   Poen. 478 praesternebant; 481 indebant; 486 necabam. 3   Pseud. 1171 eram; 1180 ibat .... ibat; 1181 conveniebatur. 4   Rud. 389 habebat .... habebat; 745 erant; 1226 memorabam. 4   Stich. 185 utebantur. 1   Triu. 503 erat; 504 dicebat. 2   True. 81 memorabat; 162 habebam; 217 habebat; 381 sordeba-   mus; 393 habebat; 596 erat. 6   Pragmenta fabb. cert. 24 erat; 26 monebat .... erat. 3   I B, Total 84     386 Arthur Leslie Wheeler   Terence   Adel. 345 erat. 1   And. 38 servibas; 83 observabam; 84 rogitabam; 87 dicebant; 90   quaerebam .... comperiebam; 107 habitabat; 109 conla-   crumabat. 8   Eun. 398 agebat sc. gratias; 405 volebat; 407 abducebat. 3   Heaut. 102 accusabam; 110 operam dabam; 988 indulgebant   .... dabant. 4   Hec. 60 iurabat; 157 ibat; 294 habebam; 426 impellebant; 804   accedebam; 805 negabant. 6   Phorm. 87 operam dabamus; 90 solebamus; 363 erat; 364 con   tinebat; 366 narrabat; 790 capiebant. 6   I B, Total 28   Cato, De agr., ed. Keil, 1895, and fragmenta, ed. Jordan, 1860.  1. 2 laudabant .... laudabant; 1. 3 existimabatur ....  laudabatur.   Jordan, p. 37. 20 capiebam; p. 39. 8veniebant .... deverte-  bantur; 64. 2 dabant; 82. 10putabant(?); 82. 11 habebatur ....  laudabatur; 83.1 mos erat .... erat; 83. 2emebant; 83. 3 erat  .... studebat .... adplicabat; 83. 4 vocabatur.   I B, Total 18   Dramatic and epic.  Ennius, Ann. 214 canebant; 371 ponebat.   Scenica 355 suppetebat. 3   Incert. Ribbeck 3 1, p. 287 I aspectabant .... obvertebant. 2   Turpilius, Ribbeck 3 II, p. 101 V flabat .... erat. 2   I B, Total 7   Historicor. fragg.  p. 64, 114 unguitabant' .... unctitabant; 1 66. 128 temptabam  .... spectabam .... donabam .... laudabam; 83. 27 faci-  ebat; 109. 1 demonstrabant; 110. 6 proficiscebatur .... seque-  bantur; 123. 13 utebatur; 141. 31 vocabantur; 202. 9 claudebant  .... educebant .... continebant .... cogebant ....  insuebant.   I B, Total 16   I B, Total 2   I B, Total 1     Orators, ed. Meyer, p. 222 vocabant; 355 solebas.  Lucilius, ed. Marx 1236 solebat.     1 Perhaps different versions of the same passage ; cf . Peter. I count them as one  case.     Imperfect Indicative in Early Latin 387   Auctor ad Herenn., ed. Kayser.   4. 6. 9 videbat .... poterat; 4. 7. lOerant .... poterant; 4. 16.  23 putabant .... existimabatur .... putabant .... opserva-  bant; 4. 22. 31 concedebant; 4. 53. 66 erat; 4. 54. 67 solebat.   I B, Total 11  CIL. I. 1011. 17 florebat.   I B, Total 1  Varro, De ling. Lat., ed. Spengel.   5. 3 dicebant .... dicebant .... significabant; 5. 24 dicebant;  5. 25 obruebantur .... putescebant; 5. 33 progrediebantur; 5.  34 agebant .... agebat .... poterat; 5. 35 agebant ....  vehebant .... ibant; 5. 36 coalescebant .... capiebant ....  colebant .... possidebant; 5. 37 videbatur; 5. 43 erat ....  advehebantur .... escendebant; 5. 55 dicebat; 5. 66 dicebat  .... putabat; 5. 68 dicebant; 5. 79 dicebant; 5. 81 mittebantur;  5. 82 dicebatur; 5. 83 dicebat; 5. 84 erant .... habebant; 5. 86  praeerant .... fiebat .... mittebantur; 5. 89 fiebat ....  mittebant .... pugnabant .... deponebantur .... subside-  bant; 5. 90 praesidebant; 5. 91 fiebant .... adoptabant; 5. 95  perpascebant .... consistebat; 5. 96 dicebant .... parabantur;  5. 98 dicebant; 5. 101 dicebat; 5. 105 faciebant .... servabant  condebant; 5. 106 coquebatur .... fundebant; 5. 107 faciebant  .... vocabant; 5. 108 edebant .... ferebat .... decoque-  bant; 5. 116 faciebant .... habebant .... opponebatur; 5. 117  fiebant; 5. 118 appellabant .... erat .... ponebant; 5. 119  infundebant .... figebantur; 5. 120 ponebant .... ponebant;  5. 121 nominabatur; 5. 122 erant; 5. 123 habebat; 5. 124 dabant  .... sumebant; 5. 125 erat .... vocabatur .... ponebatur:  5. 126 erat .... vocabatur .... habebant .... solebat; 5. 127  apponebatur .... bibebant .... coquebant; 5. 128 arcebantur;  5. 129 ministrabat; 5. 130 vellebant; 5. 132 utebantur .... iacie-  bant; 5. 139 corruebant; 5. 141 muniebant .... exaggerabant  portabatur .... sepiebant; 5. 142 relinquebant; 5. 143 conde-  bant .... circumagebant .... faciebant .... vocabant ....  fiebat .... erat; 5. 146 erat; 5. 154 aiebat; 5. 155 coibant; 5. 156  vehebantur; 5. 160 adibant; 5. 161 relinquebatur .... dicebatur  .... impluebat .... compluebat; 5. 162 volebant .... cuba-  bant .... cenabant .... vocitabant .... cenabant; 5. 164  exigebant; 5. 166 legebant .... ponebant .... dicebant ....  involvebant .... erant .... dicebant; 5. 167 calcabant ....  insternebant .... appellabant .... operibantur; 5. 168 scande-  bant; 5. 169 dicebatur .... erat; 5. 173 valebant; 5. 174 vole-  bant .... erat; 5. 177 dicebant; 5. 180 petebat .... inficiabatur     388 Arthur Leslie Wheeler   deponebant .... auferebat .... redibat; 5. 181 exigebatur;   5. 182 dicebant .... erant .... ponebant .... stipabant  .... componebant .... pendebant; 5. 183 accedebat; 6. 4  dicebant .... inspiciebantur .... dicebant; 6. 7 dicebat; 6. 8  videbatur .... dicebantur; 6. 10 putabant; 6. 11 persolvebantur;   6. 14 erat; 6. 16 fiebant; 6. 21 dicebat; 6. 22 circumibant; 6. 28  conveniebant; 6. 47 dicebant; 6. 54 consumebatur; 6. 59 vitabant;  6. 60 ponebant; 6. 66 legebantur; 6. 70 spondebatur .... appella-  batur; 6. 71 dicebant; 6. 74 promittebat .... consuetude* erat;  6. 80 dicebant .... dicebant; 6. 89 acciebat .... videbatur;   6. 95 intererat .... fiebant; 7. 26 dicebant; 7. 36 appellabant;   7. 39 putabant; 7.40 relucebant; 7. 41 legebantur .... poterant;  7. 44 dicebantur .... fiebat; 7. 52 erant .... habebant . . . .  conducebantur; 7. 56 ascribebantur; 7. 57 habebant; 7. 58 com-  mittebant; 7. 63 dicebat; 7. 73 animadvertebantur; 7. 74 arabant;  7, 84 dicebant; 7. 91 dicebant; 7. 93 erat .... vocabatur; 8. 10  erat; 8. 17 erant; 9. 54 erat; 9. 56 dicebantur .... erat; 9. 60  notabant; 9. 68 erant; 9. 59 utebantur; 9. 76 dicebatur; 9. 83  pendebat .... dicebant; 9. 87 valebat; 9. 100 dicebatur . . . .  constabat .... dicebatur; 10. 70 dicebant. 212  De re rust., ed. Keil, Lipsiae, 1883.   1. 2. 1 solebant; 1. 2. 7 dicebat; 1. 2. 9 poterat .... effodiebat  .... appellabant; 1. 7. 2 faciebant; 1. 8. 6 vocabant; 1. 10. 2  pendebat; .... dicebantur; 1. 13. 6 faciebant .... erant . . . .  laudabatur providebant; 1. 29. 3 dabant .... dicebant;   1. 41. 1 inserebantur; 1.59.1 vocabant; 2.1. 1 praeponebant . . . .  putabant; 2. 1. 6 erat; 2. 2. 3 appellabant .... reiciebant; 2. 2. 9  hibernabant .... aestivabant; 2. 5. 3 vocabat; 2. 7. 1 solebat;   2. 8. 3 dicebant; 2. 11. 5 dicebant; 3. 1. 3 habitabant .... scie-  bant; 3. 1. 4 alebantur .... redigebant; 3. 1. 5 credebant; 3. 1. 7  habebant .... serebant .... pascebant; 3. 2. 6 habebat; 3. 2. 7  ostendebas; 3. 2. 14 accipiebat .... dicebat; 3. 2. 17 dicebat   3. 3. 2 dicebant; 3. 3. 6 erat .... pascebantur .... erat .... erat  3. 3. 7 habebant; 3. 3. 8 erat; 3. 6. 6 laudabant .... aiebat  3. 9. 19 dicebant .... vocabant .... dicebantur; 3.10.2iubebat  3. 12. 6 putabat .... appellabant; 3. 13. 2 appellabant; 3. 17. 3  capiebat .... dabat .... consumebat; 3. 17. 6 erat .... habebat  .... adgerebant; 3. 17. 7 coiciebat; 3.17.8 erat .... laborabat  .... aiebat .... despiciebat. 68  Sat. Menipp., ed. Eiese, 1865.   P. 126, 1.9 erat; 139. 10 radebat; 140.4 vehebantur; 141.1 sol vebat;  169. 8 loquebantur; 181. 2 solebat; 186. 5 suscitabat; 216. 1  habebant; 225. 10 habitabant. 9   I B, Total 289     Imperfect Indicative in Early Latin 389   C. Imperfect of Frequentative Action.   Plautus, Asin. 938 dicebam; Capt. 917 percontabatur; Epid. 59 mit-  tebat; 131 missiculabas; Merc. 631 promittebas; Miles 1410  dicebat; Persa 20 visitabam; 433 negabas; Kud. 540 promittebas;  True. 506 poscebat. 10   Ennius, Ann. 50 tendebam .... vocabam. 2   Historicor. fragg., p. 138, 11 expoliabantur. 1   I C, Total 13   II. Aoristio Imperfect   Plautus, Amph. 807 aibas; 1009 erat; As. 442 aibat; Bacch. 268 aibat;  Capt. 676 aiebatis(?); Cist. 143 ai[e]bat; 585 ai[e]bat; Cure. 488 aiebat;  582aiebat; Epid. 254 aiebat; 597agnoscebas; Men.532aiebas(?); 1141  aiebat; Merc. 45 poterat; 635 ai[e]bant; 637 aiebat; 638aiebant; 765  aiebat; 766 aiebat; 804 aiebant; Miles 66 ai[e]bant; 1107 aiebat;  1430 erat; 1431, erat; Most. 1002 aiebant; 1027 aiebat; 1028 aiebat;  Poen. 464 aibat; 900 aibat; 1069 erat; Ps. 650 aiebat; 1083 aibat,  1118 aibat; Eud. 307 aibat; 502 erat; 1130 aiebas(?); Stich. 391  aibat; Tri. 428 aibas; 874 aibat; 944 aibant; 956 aibat; 986 aiebas;  1140 aibat. 43   Terence, Adel. 494 erat; 716 erat; 717 aibat; Andr. 930 aiebat; 932  aibat; Eun. 700 scibas; 701 dicebat; Heaut. 203 erat; Hec. 238  aibant; Phorm. 572 aibant; 768 sat erat. 11   Historicor. fragg. 187. 126 poterat. 1   Varro, De r. r . 2. 4. 11 dicebat(?); 3. 17. 4 dicebas(?). 2   Auctor ad Herenn. 2. 1. 1 poterat; 4. 9. 13 erat. 2   II, Total 59   III. Shifted Imperfect   Plautus, Merc. 983 6decebat; Miles 755 sat era[n]t; 911poteras; Rud.  269 aequius erat; True. 511 poterat. 5   Terence, Heaut. 785 poterat. 1   Lucilius, 204 (Marx) sat erat. 1   Varro, De 1. L. 8. 47 oportebat; 8. 65 debebant; 8. 74 oportebat; 9. 23  sequebatur; 9. 85 oportebat. 5   Auctor ad Herenn. 2. 22. 34 satis erat; 4. 41. 53 infimae erant. 2   III, Total 14     390     Arthur Leslie Wheeler         I. PEOOBESSIVE (TeUB) ImPEK-            FECT            Total        II.   Aobistic    III.            Shifted        A. Simple    B, Cast.    G. Fre-            Prog.    Past    quent.        Plautus    433    291    84    10    43    5    Terence    225    185    28      11    1    Cato 1    20    2    18          Dramatic 2 and Epic    69    60    7    2        Historians 3    52    34    16    1    i      Orators*    10    8    2          Lucilius    13    11    1        1    Auctor ad Herenn.    77    62    11      2    2    CIL., Vol. I    4    3    1          Varro    320    24    289      2    5    Total    1,223    680    457    13    59    14     1 Except historical works the citations from which are included among the  historians.   2 Laberius and later writers not included.  3 Nepos and later historians not included.   4 Hortensius and later fragments not included.