The original of tliis book is in tine Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924027343304 Cornell University Library PE 225.M42 Conditional sentence in Angio-Saxon, 3 1924 027 343 304 THE CONDITIONAL SENTENCE IN ANGLO-SAXON A DISSERTATION PRESENTED TO THE BOARD OF UNIVERSITY STUDIES OP JOHNS HOPKINS UNIVERSITY BALTIMORE FOE THE DEGREE OF DOCTOR OF PHILOSOPHY BY FRANK JEWETT MATHER JR- MUMCH 1893. PRINTED BY DR. C. WOLF Ic SOHN. PREFACE. The following investigation is an attempt to write that chai)ter of Anglo-Saxon syntax which relates to the conditional sentence, by explaining so far as possible the nature, and by exhibiting all the varfeties, of that construction. With this end in view the writer has stated his con- clusions in connected foi'm with only examples enough to illustrate the different points. But an investigator is bound to present enough of his processes and material, fii-st to give the reader control of the work and second to facilitate further investigation in tlie same field. It is hoped that these ends have been attained by the somewhat minute division of the subject, and by the appen- ded statistics. Thus a body of material is offered sufficiently laige to relieve the future writer of an Anglo-Saxon syntax of the drudgery of collecting examples for this branch of his subject. In any such work the Latin originals must be taken constantly into account, if only to free one's self of the idea, which still has some currency, that all Anglo-Saxon prose is Latinized. The writer's collations which are as complete as possible have been presented in the comparatively few cases in which they appeared to have especial interest or value. The debased nature of most of the Latin which was translated into Anglo-Saxon makes it impossible in great measure to make the categories of classical Latin grammar the point of de- IV parture or ot contrast for the syntactical study of Anglo- Saxon prose, even if this procedure were desirable. Outside of a certain few texts it is quite impossible to prove any farreaching influence of Latin upon Anglo-Saxon. The many coincidences we should expect of course prove nothing. The language must be studied upon the basis of an unbiased survey of its own phenomena, not, as too often in the past, with the intention of making it square to the rules of Latin grammar. There are certain, frequent variations from the Latin which are instructive for the determination of the value of moods in Anglo-Saxon. These are cited in the course of the dissertation. From his experience with syntactical dissertations the writer believes that we shall better attain a scientific syntax of Anglo-Saxon, by the study of single constructions in all the important monuments, than by complete studies of the syntax of isolated texts, however conscientious and pains- taking. The many ambiguous forms in the Anglo-Saxon will invariably lead astray the investigator who cannot check his work by a great number of examples. The bibliography will show how far the writer has depended upon his own reading and to what extent he has availed himself of previous syntactical investigations. Such previous work is often silently supplemented by the writer's own reading. The obligations of a worker in the field of syntax to the standard works on that subject are so obvious as hardly to call for mention. ■ In the absence of any really authorita- tive work on Anglo-Saxon syntax or even of any ancient G-ermanic language as a whole, the writer has been driven to the more general portions of those grammars and syntac- tical studies of the classical languages, which contain sounder and broader work than has yet been produced in the newer field of Germanic syntax. Such a debt is wholly an indirect one and one that can hardly be expressed by page and para- graph ; for one cannot apply classical categories directly to Anglo-Saxon, the mood-system for instance dififei'S essenti- ally from that of Latin and of Greek; but the debt in none the less a real one. The writer is conscious of no far-i-eaching obligations to the authorities cited, except to Prof. Goodwin's article on the conditional sentence in Greek, and to the same writer's "Moods and Tenses", which did much to emancipate him from the conventional, view of the subjunctive, and to the article by Dr. Erbe which offered a classification of the subject which the writer has followed in principle if not in all de- tails. The statistical disseitations of Drs. Pleischhauer and Wohlfahrt are particularly valuable for an abundant material carefully and conveniently arranged. The writer is glad to acknowledge an obligation of a more personal nature to Dr. Bright of this university, to whom he owes, at least in part, the suggestion ot the subject of this dissertation, en- couragement and advice during its preparation, and the training which alone could give it any value. If the publication of this work marks a permanent ad- vance however small towards that desideratum a complete and thorough syntax of Anglo-Saxon, the writer will feel rewar- ded for the time and energy expended upon it. VI BIBLIOGRAPHY. The following works have been read entire. The references are to page and line unless otherwise specified. For the poetry the references are to the lines. All non-poetical biblical texts are cited by chapter and verse. With a few exceptions the references are not to the protasis but to the point necessary to establish the context of the sentence. O. E. T. The Oldest English Texts. E. E. T. S. No. 83. Ed. Henry Sweet. 1885. C. P. King Alfred's West-Saxon Version of Gregory's Pa- storal Care. E. E. T. S. Nos. 45 & 50. Ed. Henry Sweet. 1871-2. Or. King Alfred's Orosius. E. E. T. S. No. 79. Ed. Henry Sweet. 1883. Cons.'*) King Alfred's Anglo-Saxon Version of Boethius de Consolatione Philosophise etc. Ed. Samuel Pox. (Bohn's Antiquarian Library) London 1864. Beda. The Old English Version of Beda's Ecclesiastical History of the English People. E. E. T. S. No. 95. Ed. Thomas Miller. 1890—1. Chron.*'^) 'Parker Ms' ; Chron. E 'Laud ms'. Two of the Saxon Chronicles Parallel. Ed. John Earle, Oxford 1865. *) Since this antiquated edition must soon be superseded it has not seemed worth while to number the lines. The references are to the pages and to the general positiou on the page: h, 'high'; ni, 'middle' ; and 1, 'low'. Forms of especial interest are cited in addition by chapter and by paragraph, so tliat they may be found without great difficulty in a subsequent edition. **) Citations from the shorter paragraphs are given by the year, from long passages by page and position (h, m, 1). VII Mat, Mk., Lk., Jn., for the Oott. ms. R & L is appended to designate the Rushworth and Lindisfarne glosses re- spectively. The Gospels in Anglo-Saxon and Nor- thumbrian Versions. Eds. Skeat, Kemble, and Hard- wick (Cambridge University Press.) 185H— 78. Blick. The Blickling Homilies. E. E. T. S. 58, 63, 73 Ed. Richard Mori'is. 1874, 6, 80. Horn. The Homilies of the Anglo-Saxon Church. 2 vols. Ed. Benjamin Thorpe, for the Aelfric Soc. London 1844,6. Wulf. Wulfstan. Sammlung der' ihm zugeschriebenen Ho- milien. Ed. Arthur Napier. Berlin 1883. Rid. The Riddles of the Exeter Book in Grein's Bibliothek der Augelsachsischen Poesie. The following dissertations *) have been of assistance in the gathering of Material. Andr. Untersuchung uber die Syntax in dem Angelsachsischen Gedichte vom Andreas. A. Reussner. Leipz., Diss. 1889. Beo. Zur Syntax des Beowulf. E. Nader. Anglia VIH. Cr. Der Syntaktische Gebrauch des Verbums in dem angel- sachsischen Gedichte "Christ". B. Hertel. Leipzig, Diss. 1891. Dom. Die Syntaktischen Erscheinungen in Be Domes Daege. J. Hoser. Halle, 1889. Cad. Ex. Darstellung der Syntax in der sogenannten Cad- monischen Exodus. E. A. Kempf Halle, 1888. El. Darstellung der Syntax in Cynewulfs Elene. J. Schiir- . *) Eeferences to the poetry are to the line numbering of Grein's 'Bibliothek', with the exception of those for Beowulf which are to the revision of Heyne's 4 th edition Edited by Harrison and Sharp for the Library of Anglo-Saxon Poetry. 3rded. Boston 1888. All citations are made in the now universally accepted orthography. VIII mann (Kortings neupbilologische Studien 4. Heft). Paderborn 1884. Csed. Gen. Der Syntaktische Gebrauch des Verbums in dem Csedmon beigelegten angelsachsischen Gedicht von der Genesis. H. Seyfarth. Leipzig. Diss. 1891. Guth. Der Syntaktische Gebrauch des Verbums in dem angelsachsischen Gedicht vora heiligen Guthlac. Leip- zig, Diss. 1889. Aelfr. T., Gen., Ex., Lev., Num., Deut., Job. Die Syntax des Verbums in Aelfrics Ubersetzung des Heptateuch und des Buches Hiob, H. Wohlfahrt. Leipzig, Diss., 1885 (Aelfr. T. designates the introduction). Jul. Darstellung der Syntax in Cynewulfs Gedicht Juliana. C. P. B. Conradi. Leipzig. Diss., 1886. , Phoen. Uber den Syntaktischen Gebrauch des Verbums in in dem angelsachsischen Gedicht vom Phoenix. J. Plauer. Leipzig. Diss. 1891. Saints. Syntax des Verbums in Aelfric's Heiligen Leben. Kuhn. Leipzig. Diss. 1889. Studien zur Aeltricschen Syntax. B. Schrader. Gottingen, Diss. 1837. Syntax der Blickling Homilies. J. Plamme. Bonn. Diss., 1885. tjber den Gebrauch des Conjunctivs in Alfreds altenglischer Ubprsetzung von Gregors Oura Pastoralis. Wm. Pleisch- hauer. Erlangen, 1885. Darstellung der Syntax in den vier Echten Predigten des Angelsachsischen Erzbiscbofs Wulfstan. A. Mohrbutter. Munster. Diss., 1886. tJber den syntaktischen Gebrauch des Conjunctivs in den Cynewullischen Dichtungen Elene, Juliana und Crist. M. Prollius. Marburg. Diss., 1888. IX Other works consulted: Matzner, E. : Englische Grammatik. 2. Aufl. II&III. Berlin 1874. Koch : Historische Grammatik der Englischen Sprache. 2. Aufl. II. Gassel 1878. March. F. A.: A Comparative Grammar of the Anglo-Saxon Language. New-York 1871. Hotz G. : On the Use of the Subjunctive Mood in Anglo- Saxon and its further History in Old English. Zurich, Diss., 1882. Erdmann. Untersuchung iiber die syntax der sprache 01- frieds. Halle 1874. Kuliner, R. : Ausfiihrliche Grammatik der Grfecischen Sprache. 2. Aufl. Hannover 1869—72. — Ausfiihrliche Grammatik der Lateinischen Sprache. 2. Aufl. Hannover 1877—9. I Goodwin W.: Journal of Philology V 186—205 and Trans. Amer. Philol. Soc. 1873 p. 60—79. — Syntax of the Moods and Tenses of the Greek Verb. Boston 1890. Gildersleeve B. L. : 1) On si with the Future Indicative and lav with the Subjunctive in the Tragic Poets. Trans- actions of the American Philological Society 1876, p. 5 — 8, and 2) A Latin Giammar New- York 1885. Erbe M. : Uber die Conditionalsatze bei Wolfram von Eschen- bach. P. B. B. V, 1—50. Liittgens C. : tJber Bedeutung und Gebrauch der Hilfsverba in Pruhem Altenglischen. Scnlan and Willan. Kiel. Diss., 188S. TABLE OF CONTENTS. pp. Introduction and bibliography ...... Ill — IX SECTION I. The conditional construction in general and in Anglo-Saxon . 1 — 3 SECTION II. Mood and Tense. 1. Ambiguous forms . . 2. Subjunctive and Indicative .... 3. Theories of the subj. in A.-S. (Koch, March, Miitzner) 4. The Pres. Subj. in protasis. A. In the protasis to a command or wish B. In a protasis introduced liy buton 5. The indicative 6. The pret. subj. A. Unfulfilled . B. Ideal condition Note. Form of the weak pret in L. \V. S. 7. Tabular conspectus of moods and tenses in prot. and in apod. . . \>i— 'iO SECTION III. Formal conditions without conditional signiiicimce : 1. Indirect questions. 2. Concessive conditions. 3. Optative conditions. 4. Forms of address . 21—28 SECTION IV. General classification of conditional sentences 24 5 -0 6 -7 9- -12 13- -14 15 IG IfD 16- -18 XII CLASS I. The simple or logical condition. pp. A) In pres. or fut. time with pres. indie. 1. The apod, is a simple statement ... 25 Remarks. Jj'ormation of the prot. without gif: 1. Inversion. 2. imperative condition. 3. periphrastical condition ... . . . '2b — 27 2. The apod, is an exclamation or a question . . 27 — 29 B) The logical cond. in past time. Pret. indie, in Prot. 1. Apod, with pret. indie. . . 29 2. Apod, with pres. indie. . . .29 Eemarks inversion and periphrastical condition 30 CLASS II. Condit. with mandatory apodoses. Pres subj. in prot. . . . ... 30 A. The apod, is a command wish or exhortation. 1. Imperative in apod. .... 31 Note. Pres. indie, in protases of such conditions 31 2. Apod, with the suhj. (adhortative) . . 31 Note. Indie, in prot. of this condition. . 32 B) The apod, expresses a moral obligation . 32 Note. Indie, in prot. . . . . 33 CLASS III. The ideal condition. Pret. siibjs. in both clauses ... . 33 Appendix: The pres. subj. in prot., neither introduced by buton nor induced by a mandatory apod. 1. Formulas of address .... 34 ► 2. Guarded statement ... 35 3. Ideas of doubt etc. . . 36 4. Miscellaneous instances . . 37 — 38 CLASS IV. The unfulfilled Condition. A) In past time ..... ... 39 B) In pres. or fut. time ....... 40 Notes: 1. Inversion. 2. pser as the conditional conj. 3. In- dies, in the unfulfilled condition . . . . . .40 — 41 CLASS V. Conditions introduced by buton (nemne, nymde). A) The prot. is truly exceptive . . . . 41 B) States an indispensable condition . . 42 xm Remaike. 1. Buton with indie, etc. ..... 43 2. Expanded forms of the oc)ndition 44-45 SECTION V. The condition expressed by forms usually non conditional . 46 1. Relative conditions . ... 47 2. Temporal conditions . ... 48 a) Introduced by J)onne . . . 49 b) Didactic condition with |)onne 50 c) with eer . . . . . 50 d) With mid SECTIOi\ VII. The conditional sentence in indirect discourse and subordinate position . .... 62 A) Dependent upon a verb in present time . 63 Indirect discourse. a) Pres. indie, in both members 63 b) Pres. subj. in apod. 63 c) Pres, subj. in both clauses . . 64 XIV pp. The apod, is a bubject clause. With impersonal constructions. a) Indie, in both members ..... 64 b) The subj. in apod. ... , . 64 c) Subj. in both members .... 65 The apod, is a clause of purpose ; a consecutive clause 66 B) The conditional sentence dependent upon a verb in the pret. 1. Indirect discourse. a) Pret. subj. in both clauses .... 67 b) A few pret. indies 68 The apodosis is dependent upon verbs of commanding; is a subject clause; a clause of purpose ... .69 SECTION vni. Clause order and word order. Possible arrangements of mem- bers: The protasis, precedes, follows, is interpolated . 71 Order within the protasis. Direct order, when the prot. precedes . . .72 Transposed order. All but universal when the prot. follows . 73 The buton condition. Direct and transposed orders 74—75 Inversion 75 Constant principle tending to produce the transposed order 74 Ordei- within the apodosis, usually transposed, induced by J)onne or ne at the head of the clause ... 76 Remarks, on unusual orders of the prot 76 — 77 SEOTIOM IX. Peculiarities of the documents. In relation to the Latin ........ 78 nu and peer as conditional conjunctions in Rushwortli M itthew 79 Distribution and possible dialectical significance of nemne (nynide) 80 Conditional conjunctions and correlatives ... . 81 STATISTICAL APPENDIX .... . 82-88 Biographical sketch of the author (in conformity with the univeisitj' regulations). BIOGRAPHICAL SKETCH OP THE AITTHOE, The writer of this dissertation was born the son of Frank Jewett Mather and Caroline Graves Mather at Deep River Connecticut July 6* 1868. His childhood and youth were speut in his father's home at Morristown, New Jersey, where after attending several primary schools he prepared for college at the Morris Academy. In the Pall term of 1885 he entered Williams College; and was graduated from that institution with the degree of A. B. in July 1889. The following Pall he entered Johns Hopkins University and dniing the past three years has pursued graduate work with English as principal subject and German and History of Philosophy as first and second subordinate subjects respectively. He received the appointment ot University Scholar for the session of 1890 — 91 and that of Pellow for the session of 1891 — 92. He has taken courses under the following members of the faculty: Drs. Adams, Bright, Browne, Griffin, Lear- ned, Matzke, Todd, and Wood, and Mr. Lewis, to all of whom he owes something for instruction or for inspiration. Especially would he express his deep sense of obligation to Drs. Bright, Griffin and Wood with whom he has found the relation of student and teacher helpful in more than the formal sense. I. THE CONDITIONAL CONSTRUCTION IN GENERAL AND IN ANGLO-SAXON. The conditional construction is often regarded as a variety of the causal construction. Like the latter it consists of two members: a statement, the apodosis, which is con- tingent for its validity upon another statement, the protasis. Unlike the causal construction, the validity of the protasis is not, and cannot be, affirmed. The protasis contains only the assumed or suggested ground, the validity of which is to be weighed by the hearer. Thus the whole construction falls into the field of supposition; and when we say that a protasis is 'real' we mean only that it is of a nature to win assent from the hearer. The question of its validity is extra- giammatical and logical. The construction can state nothing more than the contingent relation existing between the idea of the apodosis and the idea of the protasis, offering the ideas themselves for judgement. We have seen that logically the apodosis is dependent upon the protasis, grammatically the relation is reveised. The apodosis is an independent clause while the protasis stands in a dependent and adverbial relation to the verb of the apodosis. The hypothetical nature of the protasis is usually marked by an introducing conjunction: in Anglo-Saxon generally by the conjunction gif. 1 Though the conditional construction can make no affir- mation as to the validity of the ideas presented, it can as to their nature; and we distinguish ordinarily three cate- gories of suppositions: logical, ideal, and unreal, and corre- sponding to them three great divisions ot the conditional sentence. 1. THE LOGICAL CONDITION. Tiie protasis is objectively real: that is it is to be tested by the standard of fact, and its validity, thus established or disproved. "All that the Logical condition asseits is the inexorable connection of the two members of the sentence" (Gildersleeve I, p. 6). 2. THE IDEAL CONDITION. The ideas expressed in the two clauses are wholly ima ginary, and merely conceived for the sake of argument or illustration. There is no concern as tho the objective reality or probability of the ideas, but only as to the validity of the conditional relation between them and the aptness of the ideas for illustration. The condition often contains the ex- pression of something desired. 3. THE UNREAL OR UNFULFILLED CONDITION. The ideas are presented as contrary to fact or unfulfilled. In this condition the apodosis only is necessarily denied. (cf. Moods and Tenses p. 149.) This division is both logical and convenient for classical Latin; for the three conditions have respectively special grammatical forms: The indie, for 1; the pres. and perf. subj. for 2; and the imp. and plup. subj. for 3. JFor Anglo-Saxon the formal divisions do not correspond to the logical. While the great majority of cases under 1 take the indie, there are also large categories which have the pres. subj. in protasis, (class 2 and class 5 B in the sub- sequent classification), which appear properly to belong to the logical condition. Then in the main the ideal and the unreal conditions use precisely the same form, the pret. subj., in both clauses, and they are only to be distinguished by .context and by function. Accordingly a five fold division suggested by the uses of moods and tenses has been adopted, as much more convenient for study and reference. This division does not clash seriously with these logical categories, while it brings ont with appropriate emphasis the forms of the condition which are peculiar to Anglo-Saxon. The con- sideration of mood and tense in the succeeding section will give the ground-work for the subsequent classification. II. MOOD AND TENSE. 1. AMBIGUOUS FORMS. The investigator of the syntax of the Anglo Saxon verb meets at the outset with the difficulty, that the indie, and subj. forms in several cases coincide. A paradigm will show the increasing confusion of endings from Early, to Late West- Saxon. The later endings are enclosed in brackets, Pret. Strong. Fret. Weak. Indie. Subj. Indie. Subj. Sing. 1. healp hulpe Sing. 1. hatode hatorie » 2. hulpe hulpe » 2. hatodesthatode(est) » 3. healp hulpe » 3. hatode hatode Phi. ], 2, 3. hulpon hulpen(-on) Plu. I, 2, 3. hatodon hatoden(on) Thus in Late West-Saxon strong verbs distinguish mood in the pret, only in the 1st. and 3rd sing. ; weak verbs only in the 2nd sing, (a distinction which is lost in the later language cf. p. 16) Pret. pres. verbs naturally show the am- biguity respectively in the strong and in the weak pret. The confusion of pret. plu. endings begins even in the Orosius and Pastoral Care. It is evident that the functions of the moods of the pret. are best to be distinguished in the Early West-Saxon texts. The present tense with the single ambiguity of the 1st. sing, offers no difficulty of this sort. A second difficulty is tliat of determining the exact signi- ficance of those verbs, such as sculan and willan, which are moving towards an auxiliary and modal significance. I find instances where their modal significance is unquestionable, but in general agree with Lilttgens that they retain their independent value ever in Late WestSaxon. I shall proceed immediately to the consideration of the pres. subj.; for this is the only point at which there is likely to be serious diflerence of opinion ; and if I can clear it up, or at least state the reasons for my belief, the remaining cases may be treated with greater brevity. 2. SUBJUNCTIVE AND INDICATIVE. All grammarians are essentially agreed as tho the great line of division between the indie, and the subj ; and the distinction may be stated as follows : The indie, makes a statement with as little color as possible. The speaker acts simply as a reporter of his thought or observation and expresses no particular relation of his own to the thought beyond that of reporter — preserves as objective an attitude as possible. The subj. expresses the entrance of a more subjective element, a more intimate relation of the speaker to the thought: 1) He may express with various degrees of inten- sity a personal interest in the statement as favorable or unfavorable to, wished or demanded by, himself; in which case we have, an optative or adhortative use of the subj. or, 2) He may imply an intellectual judgement upon the instrinsic value of the statement, in its relation to fact, or to probability; or he may wish to imply a restriction or definition of the limits and application of the statement. All these uses may be included under the potential subj. In the optative use of the subj. he adds something to the absolute statement, in the potential use he reserves something from it. We must also recognize the fact that the subj. may- be used in many cases formally rather than logically to mark a subordinate relation. Such uses are possibly alway referable to original, logical constructions, but as we actually find them in the different languages it is often impossible to class them under any one logical principle, as optative or potential. The standard views of the pres. subj. in protasis in Anglo-Saxon must be stated, before proceeding to my own consideration of the subject. 3. THEORIES OP THE SUBJUNCTIVE IN ANGLO-SAXON. It is generally conceded for Latin and for Germanics that the subj. in a condition falls under the potential use. Trouble begins at this point, and chiefly in connection with the pres. subj. in protasis. Still practically our only autho- rities in the Anglo Saxon field agree in stating that the pres. subj. expresses the condition as only assumed cr hypothetical. Koch II. p. 48 under the head of the conditional sentence states: The subj. is used "m den Satzen, die eine nur an- genommene Bedingung enthalten". I take it that Professor March means the same thing in his statement Gram. p. 194: ■'The subjunctive is used in a protasis when proprosed as possible, the imperfect when assumed as unreal (so in Latin and Greek)" though I confess that neather the definition itself nor the parenthetical gloss is entirely clear to me. It is evident that a rule so expressed can only be interpreted in its con- text. Independent citation shows little but lack of definition. The context and the references to the grammars (Hadley for Greek) make it fair however to assume that Professor March means to state with Koch that the purely hypothetical or ideal condition takes the pres. subj. This view of Koch and of March has been accepted by the great majority of writers upon Anglo-Saxon syntax. My own study forces me to be- lieve that this view of the Anglo-Saxon subj. has been set forth by those whose eye was less on the object than on the Latin grammar. Matzner makes no statement for Anglo-Saxon. His statement for English generally is quite as general as and more diplomatically guarded than that of Professor March, a model of cautious if not of accurate statement. "Moods and Tenses p. 140 § 391 shows that it is not proper to speak of a condition as "als verwirMicht ausgesprochen". II 129. "Der Konditionalsats, insofern er cine Bedingimg enthdlt, die nicht zugleich als verwirMicht ausgesprochen, oder schlechthin als unverwirldiclit gedaclit uird, Idsst vielfach den Konjunctiv zu, obivohl er den Indikativ audi in diesen Fallen nicht aus- schliesst." It would be unfair to assume that this statement is intended to apply with strictness to Anglo Saxon. Such a rule is simply a dignified way of covering the almost ine- vitable retreat before the subj. in modern English. Expressing my dissatisfaction with the theory that the pres. subj. marks a purely hypothetical or ideal condition, or with the implication that its presence is fortuitious and irrational, I would state the outward circumstances of the great mass of conditional sentences with the pres. subj. in protasis, before venturing upon any path of my own theory. There is a certain bumble satisfaction even in an unintelligent knowledge of the facts. 4. THE PRESENT SUBJUNCTIVE IN PROTASIS. A. The pres. subj. is found I'egularly in the protasis of a condition the apodosis of which expresses a command, a wish, or a moral obligation. The indie, is admitted in such conditions usually under the influence of a Latin original. B. The pres. subj. is found practically without exception in protases introduced by buton (nemne or nymde), in the senses of 'unless', 'except', 'if not'. The following is an attempt at an explanation of these uses of the subj. confessedly at some points where I can detect nothing but a formal use, insufficient. A. THE PRESENT SUBJUNCTIVE IN THE PROTASIS TO A COMMAND OR WISH. The following would seem to be the explanation of the subj. in the protasis of a command or wish. The speaker introduces a strong subjective element into the sentence by the expression of his own desire or command. He thereby falls out of the role of mere reporter and ex- presses a particular interest in the relation. This element of will may extend through the whole sentence, and influence the verb of the protasis, which becomes subj, the proper mood for the expiession of will or wish. The subj. in such clauses is then rather adhortative, at least in origin, than potential or hypothetical. Hotz, (p. 55) has explained the phenomenon in a some- what different way; and I am inclined to think that the ex- planation is right so far as it goes. The subj. appears in 9 the protasis on account of the striving for "symmetry" *) of form in the two members of the condition. It is therefore an analogical levelling due to the influence exerted by the subj. and imperative forms in the apodosis. We agree essen- tially but I have tried to explain the reason for this ana- logical change and to show that it is not wholly mechanical and irrational. A citation from Prollius p. 19, will show that this view is no absolute novelty, and will give due credit to those who have previously held it. "Der dennoch hi rielen condition nl en. Nebensatzen sich jindende conj. prms. verdanld, wie schon Hotz a. a. 0. p. 55 § 17 und Fleischhauer § 28 gezeigt hdben, seine Entstehung nicht seiner absoluten Geltung, sondern vielmehr einer Beeinflussung des iihergeordneten Satzes". To Hotz belongs the credit of first stating this principle which in some form must eventually supersede those merely conventional expla- nations of the pres. subj. in protasis which still find their way into dissertations on Anglo-Saxon syntax. Fully to persuade the adherents of the old view of this condition I am bound to show that such conditions are neither necessarily nor usually purely hypothetical. The piinciple involved in this construction will best be elucidated, by the citation of examples based on the same Latin which appear in several texts or which with similar protases have as apo- doses respectively a simple statement and a command. I hold that there is no trace of "pure hypothesis" in such a sentence as: Fed donue min sceap, gif du me lufige *) The difficulty with this explanation as complete is that the apodosis of the huton condition almost invariably indie never inflnenres the verb of the protasis which i.s invariably subj, Hotz has in some unaccountable way left out the huton clauses entirely in his dissci-- tation, a most serious omission in an otherwise excellent piece of work. 10 C. p. 43,4. based on the Latin, Si diligis me, pasce oves meas, nor in this evidently independent translation, Gif du me lufige, Iseswa mine seep. Hom. I, 290, 30. This passage is a paraphrase from John, 21, 16. The Latin rightly expresses the condition by the indie; for the Saviour in making the command evidently neither intended to cast any doubt upon this love as something morely possible or conceivable, but knowing and accepting the love of Peter, he enforced by the thrice-given command we find in the Gospel, the obligation of service which that love brings with it. It seems to me that the passage hardly admits of any other exegesis. The condition then is not merely hypothetical, and the Anglo- Saxon has changed the mood in both instances either on account of the active interest and intent which lie in the whole thought, or from the demands of symmetry or from both reasons. Another case shows the merest possibility of the dubita- tive use of the subj , though the Latin gives no countenauce to such a view. Satan can hardly be accused of really doubting the divinity of Christ, or if doubting, of being undiplomatic enough to express the doubt. The only reason for a subj. in the Latin would be that the sentence is a mere taunt. Even the most radical exegetes will hold I think that the temptation was a real one. I cite four passages which are based on two Latin passages which have however the same moods and tenses. Assuming for brevity that this is the original of all., Si films Dei es, die ut lapides isti panes fiant, we have it translated closely, Gif Jju Godes sunu sy, cwed f)set fas stanas to hlafe gewurdon. Mat. 4, 3 and three times besides with subj. in protasis. Blick, 27,5; Lk. 4,3; Hom. I, 166,12; a change from the mood in Latin 11 which is hardly to be explained on any other theory than the one advanced above. Similary Si filius Dei es, mitte te deorsiim, though translated mechanically in Mat. 4,6; Gyf J)ii Godes sunu eart, asend J)e f)onne ny{)er, appears in four other cases with the pres., subj. in protasis Gif du Gode.s Sunu sy, feall nu aduu, Horn. I 166,12 and 170,1; Blick. 27,10; Lk. 4,9. March explains the following subj. as denoting a con- dition which is 'possible' Onsend Higelace gif mec hild nime, beaduscruda betst, |)8et mine breost wered. Beo. 452, cf 1480. Without denying the correctness of the explanation, I am unable to see why it should not apply equally well to an indie, condition six lines previous. Ac he me habban wile dreore tahne, gif mec dead nimed. Beo. 446. The subj. as before is to be explained by the presence of a mandatory apod. It would be absurd to say that one of these con- ditions i^: more real or probable than the other. The following is a mere taunt and is not decisive for our principle; still the Latin shows the indie, against two translations by the subj. Si filius Dei es, descende de cruce; Gyf |)U sy Godes Sunu, ga ny{)er of J)£ere rode Mat. 27,40 also Hom. II, 256,5. The apodosis may contain an expression of moral obli- gati6n or fitness, which amounts to a mild command. Ac gif we wilnigen (conamur) d;et he dtes wos geswicen, donne sculon we hie — laeran daat he ne wilnigen leas giel- pes C. P. 367,22; so 199,13. Gif hire donne se widsace, donne is cynn. dat him spiwe dset wif on dajt nebb. C. P. 45,2. But it hardly needs extended proof to show that such conditions are very seldom to be classed as conditions of pure hypothesis or mere possibility; the fact lies in the vei'y nature of the case. A command or an exhortation is usually 12 expressed not in view of a remote possibility but with reference to a known on definitely anticipated state of things. Another and ooe of the strongest reasons for doubting that the note of hypothesis is dominant in the pres. subj. in Anglo-Saxon is that, except in indirect discourse or subordinate position, it is found only in the adhortative use in apodosis, and in other independent clauses it practically never expresses ideas of doubt nor of mere hypothesis. Outside of formal uses, wish and command are the most important ideas con- veyed by the pres. subj. in Anglo-Saxon. These conditions with adhortative apodoses do not ne- cessarily take the subj. in protasis. The subjective relation to the protasis may be suppressed. Still excluding the Gos- pels which translate somewhat mechanically a large majority of such conditions have the subj. in apodosis. If we may trust to the antiquity and reasonably faith- ful transmission, so far as syntax is concerned, by Alfred of Aedelbirhts Laws, circa 610, the indie, originally stood in such conditions and the subj. was just beginning to intrude at that time. There are only fifteen subjs to sixty five in- • dies, in protasis in this code and part of these may admit of another explanation. The later laws generally show a somewhat irrational mixture of forms with a great pi'epon- derance of subjs; while a few have carried the subj. through consistently. As indicated by some of the examples other elements as those of doubt or indefinite futurity may in isolated instances be found in such sentences, but the ruling principle foi- the choice of mood is that of will or wish. The other elements, unless accompanied by a command would hardly require the pres. subj. 13 B. THE PEESENT SUBJUNCTIVE IN A PROTASIS INTRODUCED BY BUTON. The protasis introduced by buton (nemne or nymde), in the senses of 'unless', 'except', 'if not', practically without exception takes the subj. It is not easy to see the principle which lies behind this use in many cases. The apodosis takes the indie. We may distinguish two general heads. 1. The buton clause as a true exception. Buton introduces a condition that must be excepted in thought before the apodosis can be conceived of as fulfilled — suggests an event which may be interposed to prevent the fulfilment of the main idea. It thus involves caution in state- ment, a special suvey of the situation, and a reservation from the absolute statement. In most cases too it deals with contingencies in the indefinite future, and perhaps properly falls into the receptive category of "mere possibility", though not of "pure hypothesis". There would seem to be logical ground enough for the use of the subj. in this case, thougli I confess to an inability to formulate a wholly satisfactory reason for it , and to this extent perhaps the traditional view of the subj. is justified. At ail events the subj. in the pro- tasis whether formal or logical, is independent; neither in- fluenced by nor influencing the mood of the apodosis. Examples : Ac hi gegaderiad monigfeald dysig on J)3ere fortruwunga 7 on pam gilpe, butan heora hwelc eft to hyre gecyrre Cons. 6 h. Micel yfel ded se de leas writ, buton he hit gebete. Hom. II, 2,22. Hio wind wid da god de mon aer gedon hsefd, buton mon simle swincende & wyrcende sie god weorc od ende. C. P. 445,14, Ealle we sind gelic eet- 14 foraii (J ode. buton hwa ocfeine mid godum weoreum foideo. Honi. I, 260,24. In the second case, buton introduces a condition witliout wliicli the independent clause has no validity — states an indis- pensable condition The independent clause contains the ne- gation of an idea, which depends for its realization upon the condition introduced by buton. The formula is; This will not happen, unless this happen. This is the most common form of the buton condition. I am unable to detect much of the element in these con- ditions, which traditionally demands the subj.; they seem to me to be usually a simple statement of the relation without particular subjective coloring. The contingency however becomes more stringent, when the condition is expressly stated as indispensable. This use of the subj would seem on the whole to formal rather than logical and possibly an analogical ex- tension from the former class. At all events buton with the subj. is stereotyped in Anglo-Saxon. Examples: Ne he on horses hricge cuman wolde, nemne hwilc mare nyd absedde. Beda 160,16. Monnan ic ne bite, nymde he me bite. Rid. 66,5. We ne beod na fulfremede, buton we beon afandode. Hom. I, 170,14. f)onne forlyst he eall his aerran god, buton he hit gebete Cons. 1701. Ne bid nan faesten gode gecweme, buton se man bine sylfne fram leahtrum forhsebbe Wulf. 286,12. NOTE. Genuine cases of buton with the indie, are very rare and were probably regarded as ungrammatical. Probably hardly half a dozen instances could be cited from a microscopic examination of all of Anglo-Saxon, excluding a few slavish translations in interlinear glosses. A sure case in L. W. S. 15 and the only one I bave noted is. Por{)on liiom ssend god on micelne brogan, fyr and hungor, butou hio to godes ge- leafan ser gecyrran willa|). Wulf. 216,14. REMARKS. 1. In this example fr.ora Beo. 1559 buton seems to have merely the modem, adversative sense of 'but'. J)set wses w;epna cyst, buton hit was mare, donne renig mon Oder to beadu lace setberan meahte. 2. The Indie, may follow, buton dset, Cons. 4 m, buton donne C. P. 399,3, buton fordyde 0. P. 4'23, 22, and similar collocations. In most of these cases we have merely a contracted clause, which leaves the proper verb of buton to be supplied from the context. The verb in the indie, is properly not introduced by buton but by the following conjunction. 5. THE INDICATIVE. The indie, may now be considered briefly. It is the mood of simple, objective statement. It offers a condition or a hypothesis for judgment without implication or reservation. a) The pres. indie, is used to express both present and future relations. The verbs sculan and willan had usually their full independent force. The forms of been are prefered though not exclusively for expressing future relations of the substantive verb, biod geraonigfaldade (muUipUcabunturJ Vesp. Ps. 138,18. b) The pret. indie, represents the action as taking place in past time, absolute or relative. It is the simple, historical tense. Anglo-Saxon atready formed a perfect with habbe and the past. part, and a pluperfect, but they are rare and deci- dedly a subordinate element of style or of syntax. 16 6. THE PRETERITE SUBJUNCTIVE. A. UNFULLFILLED. The pret. subj. is preeminently the mood of the unreal or unfulfilled condition. The CQnstruction seems exactly parallel with that in Latin. All the time relations expressed by the plup. and imp. subjs. in Latin are expressed by this one form of the verb. Examples are unnecessary at this point (cf Sect. IV, class IV). B. IDEAL. The pret. subj. has also a i-estricted use as the form for the purely imaginary or ideal condition, wicb is expressed without reference to fact. Job cwaed, Eala gif mine synna and min yrmd, f)e ic dolige, wseron awegene on aura waegan, fonne waeren hi swserran gesewene donne sand corn on sse. Hom. II, 4.54,22. Aelfric's translation of Job. 12^ is a verbal repetition of this earlier rendeiing. Grif ge nu gesawen hwelce mus {)aet wsere hlaford ofer ojjre mys 7 sette him domas 7 nidde hie aefter gafole, hu WLinderlic wolde eow pset fincan. Cons. Cp. 16, §2, p. 52h. XOTE. The form of the weak preterite in Late West-Saxon. This note illustrates a fact noted in passing by Sweet (Anglo-Saxon Reader xcviii) : "in the later language where there is no distinction between indie, and subj. in the preterite of weak verbs". It may be a mere subtlety to inquire whether the subj. function has gone over to and been superadded upon the indie, form, or whether the indie, form 17 hat: levelled out the only distinctly siibj. form I'eniaining, so that we have essentially two moods distinguished by function but not by form yet since many of the pret. subj. forms still remain distinct and since even in ambiguous forms the difference in function is so plain, it seems worth while to maintain the latter view as a more reasenable interpretation of the fact. I am loath to give up the subj. pret. form of weak verbs and to admit that it has become wholly fused with the pi'et. indie, for we tind it fulfilling exactly the same functions as it did when the distinction of moods was clearest. Accordingly I think that an alternative -st ending should be added in the paradigm for the 2"" sing. pret. subj. of weak verbs in Late West-Saxon. Reference to the paradigm of weak verbs will show that the only forms of the pret. indie, and pret. opt. with different endings in Late West-Saxon were the second persons singular, as: |)U hsefdest beside J)U hsefde. We must recognize the fact that even this single difference disappears in the later mss. and that the indie. 2"^ sing, displaced the corresponding form in the subj,, and the moods became exactly alike in form and only distinguished by difference of function. The proof of this is that forms apparently indie, are found in unreal and ideal conditions at a time when such conditions in all unambiguous cases require the pret. subj., and when, except in this apparent case, they never admit a plainly indie, form. Gif f)U wistest hwset pe toweard is, J)onne weope du mid me Horn. I, 404,27. as proved by the context and by the form weope, is an unfulfilled condition. Wite J)U for sod, gif fiset f)ine ague welan wseron |)e pu msendest pset J)U forlure, ne mihtest J)U hi forleosan Cons. cp. 7, § 3, p. 20'" Audader adfirmem, si tua forent quae amissa 2 conquereris nulla modo perdidisses. So 'tnihtest' : Cons, cp. 36, § 2; 74h. fc wat gif \>\\ mi hsefde (Bod. nsef- dest) Mine anweald dines selves, donne iist'dest du hwset hwega on fie selfum dses J)e n;\ifre {)iiiuin willum alaetan wol- dest, ne seo wyrd {)e on genimnn ne mihte Cons. cp. 11, §2, 34 h. The later Bodleian nis. shows an additional bad reading with the ^-st form the original must have had hasfde in both clauses. Ic wolde cwefan f)8et J)u uf)wita wsere, gif J)U ge|)yldig waere 7 gesvvugian mihtest (Gott. meahte) Cons. cp. 18, § 3, 20 m. Intelexeram, inquit, si tacuisses. The superior Cotton ms. has preserved the original reading meahte. Eala gif \>a wistest — — Lk. 18,42. quia si cog- nouisses. The list could be enlarged and other cases are cited in the course of the dissertation, but these instances will suffice to establish the fact. 7. TABULAE CONSPECTUS. It may be well to give a brief conspectus of the normal use of moods and tenses in the conditional period. PROTASIS. a. The present indicative. The pres. indie, may be used to express any pres. or fut. relation that is not unreal or purely imaginary, except in the condition with buton. (Cf. Sect. IV, class I a). b. The preterite indicative. The pret. indie, is used: 1) To express a real or logical condition in past time (cf. Sect. IV, class Ie); 19 2) or rarely, following a pret. in indirect discourse (ct. Sect. VII, 2). c. The present subjunctive. The pres. subj. is used : l)In a protasis the apodosis of which is a command, or wish (cf. Sect. IV, class II); 2) In a clause introduced by buton (nemne, nymde) (cf. Sect. IV, class V); 3) Often in indirect discourse (cf. Sect. VII 1) ; 4) In a few limited and miscellaneous cases (cf. Sect. IV, class III, appendix). d. The preterite subjunctive. The pret. subj. is used: 1) [n a purely imaginary or ideal condition (cf. Sect. IV, class III); 2) In a condition that is unreal or unfulfilled (cf. Sect. IV, class IV): 3) In indirect discourse following a pret. (cf. Sect.VII, 2). APODOSIS. a. The present indicative. The pres. indie, is used in the apodoses of a, b 1, c 2. b. The preterite indicative. The pret. indie, is used in the apodosis of b. c. The present subjunctive. The pres. subj. adhor. (or the imp.) is used in the apodosis of c 1 , and rarely in that of a and b (cf. Sect. IV, class II, note). 20 Otherwise the pres. subj. appears in the apodosis only in indirect discourse or other subordinate construction. d. The preterite subjunctive. The pret. subj. is used in the apodosis of d. Formal conditions without conditional significance may well be considered and dismissed in the following, short section before proceeding to thegeneral classification. 111. FORMAL CONDITIONS VVITHODT CONDITIONAL SIGNIFICANCE. I. INDIRECT QUESTIONS. These form the largest category. Gif has the force of hwseder. The mood is usually the indie. f)U wast, gif hit is, swa we soJ)lice secgan hyrdan Beo. 272. Smeaga nu gehwa on his mode, gif das beboda and odre fillice habbad aenigne stede on his heortan, donne tocnaewd he hwseder he is from Gode Horn. II, 228,20. cwedende, gif is rihtic dfem menn forleta wif his. Mat. L. 19, 3. dicentes: Si licet homini dimittere uxorem suam. Except for the stereotyped form wsere which follows a pret. f)a befran Judas gif he hit wsere. Horn. II, 244,6 also II, 250,28. {)£et he J)one wisan wordum naegde, frean Ingwina, frsegn gif him wsere sefter neodladum niht getsese. Beo. 1318. 2. CONCESSIVE CONDITIONS. A formal condition may have an adversative sense and therefore belong properly to the concessive construction. The following sentence is confirmed as a concessive construction by the Latin. Gif we nauht dses ne dood de us mon mid goode leanian dyrfe, ne do we eac nan woh de 22 us mon fore tselaii durfe. C. P. 337, 20, et si digna misericordiae retrihutione non agimus, nulla tamen perversa perpetramus. and gif eowere synna wseren wolcnreade ser dan, hi beod scinende on snawes hwitnesse. Horn. II, 322,10. Also Horn. I, 350,2. 3. OPTATIVE CONDITIONS. The condition expressing a strenuous wish difficult or impossible of attainment by the pret. subj. is to my feeling usually purely formal ,and eala gif is to be translated merely 'would that'. The mechanical explanation based on the pro- bable origin of the construction, is thai we must assume ellipsis of an apodosis expressing some such idea as 'how good', or 'how much better it would be'. Bala gif ic moste dam eadigan Laurentiam geefenlsecan. flom. I, 432,10. Eala gif du wsere hund 1 Hund is sawulleas, and on belle ne drowad Hom. II, 308,13. The following instance shows that use noted previously of the pret. indie, form for the pret. subj. which is found in nearly all late texts. Eala gif du witan woldest J)aere halgan rode gerynu Hom. I, 588,17. 4. CONDITIONS OF COURTEOUS ADDRESS. Certain forms of address which are not very common in Anglo-Saxon should receive mention here. Grammatically they occur generally in complete conditional sentences, but logically they have no more conditional force than the modern, courteous formula 'if you please'. Prom the preface to the Pastoral Care. Pordy me dyncd betre, gif iow swa dyncd, det we, eac sumas bee 23 oil daet gedii)fle Wfudeii C. P. 7,6. Similarly a con- dition may be used as a rhetorical form. Gif hwa J)onne cwyd: Hwaet ete we on f)am seofodan geare, gif we ne sawad ne ne gaderiad ure waestmas? (21) Ac eow sylle mine bletsunga fiaet an gear bringd {ireora geara waestmas. Lev. 25,20. "(Gif hwa J)onne cwyd hat den Sinn von: "Es Jconnte wohl Einer sagen — — " Worauf dann: (21) ich aher gebe Eueh etc)" Wohlfahrt s. 20. Possibly several of the conditions treated under IV, 3 appendix 1 "Formulas of address" should have been included rather in this category. The fact that they all had the pres. subj. in protasis seemed to justify their present classification. IV. GENEEAL CLASSIFICATION OP CONDITIONAL SENTENCES. A practical classification of conditions under five heads is immediately suggested by the treatment of moods and tenses. First, must come the great class of simple or logical conditions with the indie, in both clauses. Second, the conditions which with a mandatory apodosis have regularly the subj. in protasis. This class is so im- portant for Anglo-Saxon that the erection of an especial category for it seems justified ; and its logical nature suggests that it follow the simple condition immediately. Third, the ideal condition which is regularly expressed by the pret. subj. in both clauses. As an appendix to this class the instances of the pres. subj. in protasis not included under class two may be appropriately collected. Fourth, the unfulfilled or unreal condition, which takes the pret. subj. in both clauses. Fifth, the exceptive condition, introduced by buton nemne, or nymd:e, which has the subj. in protasis, forms appropriately a special class, though logically it is related sometimes to class three, some times to class one. For the sake of clearness, indirect discourse, contraction 25 and ellipsis, and conditions expressed by forms usually non conditional have been kept ont of the geneial classitication and relegated to special sections for fuller consideration. CLASS 1. THE SIMPLE OR LOGICAL CONDITION. The speaker simply asserts the relation between two ideas, without any implication as to the intrinsic validity of the ideas themselves. The hearer is supposed to test the matter by the standard of fact. Such conditions are by far the most numerous in Anglo-Saxon. There is a tendency to include all pres. and fut. conditions under this form. The indie, is found in both protasis and apodosis. A) All such conditions in pres. and fut. time are ex- pressed by the pres. indie. l)With a simple statement as apodosis. Fut. condition: gif ic eft gefare (vicero) swelcne sige get Romanum, {)onne mseg ic sijf)f)an buton itlcon f)egne Creca lond secan (revertero) Or. 156,31. Swa msere ge beod swa swa englas, gif ge of J)am treowe etad. Hom. I, 176,20. Pres. condition : Gif se blinda done blindan laet, he feallad began on senne pytt. C. P. 29,7 — he sceal hyran feondan , gif he nele freondum. Wulf 177,10. Gif ge sodlice ne forgyfad mannum, ne eower Fseder forgyfd eow eowre synna. Mat. 6,15. Fordi is se man betere, gif he gode gedihd, J)onne eall eta nytenu sindon. Hom, I, 16,14. REMARKS. F0R.MATI0N OF THE PROTASIS AVITHOUT e hit ne stod na ser durh hit sy)f. Hom I, 160,5. Gewite seo sawul ut, ne mseg se muil clypian, J)eah de he gynige, ne eage geseon, |)eah de hit open sy: ne nan him ne ded nan ding, gif se lichama bid sawulleas. Hom. I, 160,5. The final indie, condition proves the logical nature of the two inverted conditions. Gif rice wif and earm acennad togiedere, gangon hi aweig; nast du hwseder bid I)ses rican wifan cild, hwfeder faes earman. Hom. I, 256,13. And hit bid to menigfeald , gewurde hit J)riddan side, and mid ealle misdon gewurde bit feordan side. Wulf 305,1. Followed immediately by the exactly parallel condition, and, gif hit oftur gewyd, nyte we, hu J)aet faran mseg. Wulf. 305,6. Such collocations of the subj. and indie, go far to prove that the subj. is formal and not logical in such in- verted clauses. A case with the indie, noted by Matzner: bid se torr fjyrel, ingong geopenad, J)onne ic perest him Jmrh eargfare in onsende, in breostsefan bitre geJ)oncas. Jul. 402. The following 'clipped' form is probably to be regarded as an indie, geseo we senigne mann f)e georne hine sylfue to urum godum bugan wylle. eall |>8et he
fes ealles nis monnum f)onne mare Isefed to bugianne, buton swelce an lytles cafertun Cons. op. 18, § 1, p. 62"". Huic quartae, si quantum maria paludesi|ue premunt, quantumque siti vasta regio distenditur oogitationesubtraxeris, vix augu- stissima inhabitaudi hominibus area relinquetur. cf. Cons, of. 32, § 2, p. 116 h. where there is no Latin original. 3. Periphrastica condition. The periphrasis on |)a gerad, may introduce a condition. The snbj. may be found in the clause which elucidates gerad. Swa ic wolde, la Mod foet ftu fore up to us, gif de lyste, on J)a gerad |)8et pu eft mid us pa, eorJ>an secan wille for godra manna {learfe Cons cp. 7, § 3, p. 221''. Ascende ei placet, sed ea lege, ne uti cum ludicri raei ratio poscet, descendere iniuriam putos. fonne is hit alefed , |)at he mot mete |)ycgan on j)a gerad, J)e he his selmessan sylle. Wulf. 285,3 repeated 1. 6. Also Or. 52,33; 122,3. Chron. E 1091, p. 229'- 2) The apodosis may be an Exclamation or a question, a) Which is rhetorical and expects a negative answer, or dissent from the conclusion. 28 Gif donue se sacerd bid uugerad dses lareowdomes, hwset forsteut doniie his gehlyd? C. P. 91,24. Hu mseg donne senig man ryhtwislice 7 gescead- wislice acsigan, gif he nan grot rihtwisnesse on him nseff)? Cons. 156 h. Gif ge da ane luflad J)e eow lufiad, hwilce mede hsebbe ge J)onne set Gode? Hom. II, 216,20. Hwilc fgeder wille syllan his cilde stan, gif hit hine hlafes bittV odde naeddran gif hit iisces bltt? etc. Hom. I, 250,7. God is sealt; gif hit awyrd, on fam J)e hit gesylt bid? Lk. 14,34. b) Which is rhetorical and expects assent from the hearer. The words hu ne, in North, and Merc, ah ne, in the sense of nonne often introduce the apo- dosis. Odde gif hwelc folc bid mid hungre geswenced, & hwa his hwaete gehyt & odhielt, hu ne wilt he doniie hiera deades? C. P. 377,8, and gif man odrum miltsad, hu mioele swidor wile God miltsian and arian mannum his handgeweorc. Hom. I, 68, 23. and gif bisceopas forgymad — (^77) wa heoni f)8ere swigean! Wulf. 176, 28+. c) The question may not suggest the answer directly. It is proposed so that the hearer may supply the proper conclusion to the condition, or with the pur- pose, of gaining information, or of forcing the hearer to admit the validity of the condition. Porfam gif hig on grenum treowe fses fiug dod, hweet dod hig on |)am drigean? Lk. 23,31. Gyf min brodor syngad wid me, mot ic him for gyfan od seofon sidas? Mat. 18,21. Gif David hyne on gaste Dryhten clypad, hu is he hissunu? Mat. 22,45. 29 Gif ic sod secge, hwi nelle ge me geiyfan. Horn. II, 226,5. cf. Jd. 8,46. Such variatioirs of the apodosis will be found in other classes, but having once called attention to their main varieties, it will be unnecessary to carry this merely formal and obvious classification of pro- tases through beyond this large branch of the logical condition. B) The logical condition in past time. The pret. indie, is found in the protasis of the simple condition in past time. It is the form in direct narration. The apodosis may of course be in any time subsequent to that of the condition, that is with reference to the speaker in past, present or future time. The verb then of the apodosis may be in the pret. or in the pres. indie. As in the simple, pres. condition the apod, may be a simple statement, or its equivalent an exclamation or a question. 1. Apodosis with pret. indie. Gif hie f)onne sod ne ssedon, poime naeron najier gode ne pa, ne nu. Or. 86,10. Gif two men ofer iii coman rideiid to an tun, al pe |)(?)unscipe flugsen for heom. Chrou E 262,34 (Peterboro). Gif ic ryhtwis w;ps, ne ahof ic me no fordy. C. P. 253,7. Ond he pes biscop ricum monnum no for are ne for ege naefre foreswigian nolde, gif heo on hwon agylton. Beda 162,12. 2. Apodosis with pres. indie. Gif Crist for us eallum dead was, donne weordad ealle men deade. G. P. 43,9. Gif we god underfengon of Gode.';* handa, hwi ne sceole we eac yfel underfon ? Hom. II, 4.'')2,:>2. 30 Gif hi mid tenigan pingan Cristan gegladodan, on f)isum eanrian life, J)onne habbaf) hi reste frono {)£ere nontide J)8es seternesdaege of) J)aes moiiandseges lihtincge. Wulf. 219,35. and gif hi min ehton, f)onne ehtad hi eae eower. Horn. I, 556,2. REMARKS. 1. Inversion. An interesting and possibly isolated example of an idiom that still survives is: And nau man ne doiste slean oderne man, nsefde he nsefre swa mycel yfel gedon wid I>one oderiie. Chron. E, 222,7. The sense is rather concessive than conditional. 2. Periphrastical condi tion. The periphrasis on I)8et gerad is found with a subj. due to the grammatical subordination of the following clause. 7 siex hund gisla on his geweald underfeng, on {)8et gerad psei, hie him sil)j)an ece |)eowa8 wseren. Cr. 122,3. F. Or. 52,33. Her Eadmund oferhergode eal Cum- berland, and hit let to eal Malculme Scotta cyninge, on J)8et gerad "pset he wisre his midwyrhta segfier ge on sie ge on lande. Chron. 945. CLASS II. CONDITIONS WITH MANDATORY APODOSES. Conditions the apodoses of which express a 'command', 'exhortation', or 'moral obligation' have generally the pres. subj. in protasis. These conditions are most of them logical, but since Anglo-Saxon changed the mood in most cases it seems desirable to make a separate division for this characteristic construction. The indie, is freely admitted, but generally under Latin influence ; and the better texts show a prepon- clerance of the subj., abundantly large to establish it as the better idiom. 31 A. THE APOD. IS A COMMAND, WISH OR EXHORTATION. 1. Apodosis with the imperative. seo, gif J)U dyrre! Beo. 1379. Saga, gif J)ii cunne EI. 856. cf. Rid. 33,13. Gif dii wille din lac bringan (offers) to dam wiofude, & du donne ryhte ofdence hwsethwugu dse.s de din niehsta widerweardes gedon haebbe, forlget donne an din lac beforan daem weofude, and fer aerest sefter him etc. C. P. 349,9. Gif he Godes man aj, fylgad ge him Beda 100,23 from Si homo Dei est, sequimini ilium. Gif se lareow vvel tsece and yi'ele bysiiige, dod swa he tsecd, and na be dam J)e he bysnad. Horn. I, 242/22. cf. II, 48,34 and contrast Horn, II, 50,3, indie, conditions, which diifer only by the absence of a command in the apodosis NOTE. THE PRES. INDIC. IS FOUND IN THESE CONDITIONS, especially in the Gospels; very seldom in Alfredian texts; often in Aelfric. Eornostlice gyf pu bringst (offeres Fut.) I)ine lac to weofode and I>u J);er gel)encest I)am altare etc. Mat. 6,23 & 24. Horn. I, 54,19. Fseder gif pa wilt (vis) afyr J)ysue calic fram me. K and L have vvelle.) Lk. 22,42. Gif du wylt becuinan to dan ecan life, heald pas bebodu. Horn. II, 400,5. 2) Apodosis with the Subj. (ad hortative). ond gif heo beam hebbe, feo det beai-n to dem londum efter hire; 0. E. T. Charter 45,1 15. Gif his hwa sie lustfuU mare to witanne, sece him J)onne self J)8et Or. 100,25, Gif he |)onne eow eac forhogie 7 eow ne wille arisan, mid py eower ma is, sy he f)onne from eow forhogad. Beda 102,2. utan frefrian ahwsenede and heortan ormode, alysan gehaefte, gif us to {)am onhagige. Wulf. 209,5, Gif hwam twinige be dam gemsenelicum seriste, f)onne under-^tande he J)isne drihtenlican (■wyile. Horn. I 132,27. 32 The desire may be expi'essed by ap expression like ic bidde, instead of a hortatory subj. Gif hit nsenge f)inga to dsege beon msegge, ic halsio J)3ette ne sy long faec bitweonum. Beda 290,21. translates Si millatenus hodie fieri potest, obsecro, ne sit longum spatium in medio. Similarly. Gif hit eallinga f)us arseded seo (de- finitum est) 7 {)es cwide onwended beon ne mote, ic biddo 7 halsio J)3et ne sy mare fsec betweonum J)on fieos nehsta neaht an. Beda 290,23. NOTE. THE PRES. INDIC. IS FOUND RARELY IN THESE CONDITIONS. Gif du I)£es J)ing dest (facis), geswutela I)e sylfne iniddan-earde. Jn. 7,4. Gif hie donne giet dper tueonad (dubitant), gongen donne to dfem halgura gewrithm, frine dara hwret he don odde iKran spylen. C. P. 103, 9. Gif we untrume beod, uton beon gedyldige, swa swa ae eadigra Job us eoUum bysnode. Horn. II, 328,24. B. THE APODOSIS EXPRESSES A 'MOBAL OBLIGATION', or a suggestion of what is 'befitting' in view of the protasis. Such an expression is usually a disguised or a weakened form of the speakers command or wish. The apodosis con- tains words, like sculan (in the sense of 'ought' only), him is {)earf, is cynn, the inflected infinitive with eom or beom, etc. Ac gif we wilnigen (conamur) dset hie dses wos geswicen, Sonne sculon we hie l^ran dset be ne wilnigen leas gilpes. C. P. 367,22. gif bit fonne dead wurde butan fulluhte and hit on fam preoste gelang sy — — J)onne sceal he J)olian his hades. Wulf. 300,18. G't hit sie mannes gemet fset he ciricean halgan sceole, donne is hit ealles gerisnost jf)aet bit sy on dsem daege de se sige onweald wses ; Blick 205,23. Gif hire donne se widsaece, donne is cynn daet him spiwe dset wif on dset nebb. C. P. 45,2. Gif hwyle J)onne ofer fset 33 gefremman, |)onne seondon heo to bescyiienne {prinandi sunt) Cristes lichomau 7 blodes. Beda 74,16. NOTE. INDIC. IN PROTASIS. Such conditional sentences have naturally the indie, in protasis more frequently than those under A. Still a majority take the subj. Ke sceole we tellan, gif we hwajt lytles to gode gedod, ac we sceolon geriraan ure misdaeda. Horn. II, 430,6. Mine gebrodra, gif ge gode sind, f>onne sceole ge emlice \viJ>ercorenca manna yfelnysse forberan. Horn. I, 526,33. Class III. THE IDEAL CONDITION. The ideal or purely hypothetical condition is usually ex- pressed with the pr^t. subj. in both protasis and apodosis. The conditional ideas expressed are purely imaginary. They are not to be tested by their correspondence to fact, but by the validity of the relation established between the ideas and its aptness for illustration. Such conditions con- stitute a very small class in Anglo-Saxon. The very few conditions having the pres. subj. in pro- tasis, which are not explained under other heads will be appended here, as more nearly related to the ideal condition than to any other. Ideal condition. Eala psette dis moncyn wsere gesselig, gif heora mod (76) wsere swa riht 7 swa gesta|)elodl Cons. cp. 21, p. 741+. possibly unreal but translating! felix hominum genus, si uestros auimos amor quo caelum regitur regat. and deah he nu J)a J)reo hsefde, gif he naere hliseadig, donne, wsere him (122) deah sumes weord scipeswana. Cons. cp. 33, § 1, p. 1201+. Menig man is cristen geteald on sibbe, {)e wolde swide hrade widsacan Criste, gif him man bude pset man bead J)am 3 34 martymm. Horn. I, 212,25. Hwa wolde me aefre gelyfan, gif ic wolde gereccan J)8et cTornas getacnodan welati, Jonne J)ornas priciad, and da welan gelust fuUiad? Horn; II, 88,17. Gif nu eall {)ises middaneardes wele come to anum men, hu ne waeron |)onne ealle of)re men wsedlan butan him? Cons. cp. 13, p. 38m. At eadem si apud unum quanta est, ubique gentium congeratur, ceteros sui inopes fecerit. Gif da ealle awritene wseron, ic wene ne mihte pes middaueard ealle fa bee befon. John 21,25. quia si scri- bantur etc. With the apod, in adhortative form. Gif eow swa lice f)uhte, utan gangan on J)issuni carcerne and hine ut forleetan, Blick 247,1 ; in the more vivid form with pres. indie. Gif hwa alefed wsere, odde liraleas on J)issum life, he bid donne swa hit awriten is etc. Horn. I, 236,28. appendix: the pees. subj. in protasis. In instances neither introduced by buton, nor induced by a command or a wish in the apod. The agregate of such cases is small, and the average for a text is only three to four. It seems quite impossible to state a principle which will cover all the cases. Some are undoubtedly capricious variations. The small number of these cases is in evidence that the pres. subj. was not the case of mere possibility or assumption. Practically all the cases put under such rubrics in the dissertations belong under class IL. 1. Formulas of address. With inversion, probably subj. but possibly 'clipped' in- dies. Ac frine hie J)onne setter hu monegum wintrum sio sibb gewurde — — : f»onne is ^set gefter L wintra 7 feower hundum. Or. 182,16. also Ahsige Jjonne 182,19. Gif 35 |»u f)onne daet gemet habban wille 7 J)a nyd fiearfe witan wille, fonne is |)0et mete 7 drync 7 cla{)as * * * j)aet |)e is gecynde. Cons. cp. 14, § 1, p. 42 ii. In all these cases the proper apodoses are lacking. A similar condition with apodosis: Gif hwa smeaga hu dis ge- wurde, J)onne secge we, {)8et des martyr his lyf adreah on laewedum hade. Horn. 1, 452,24. Wilt du, wit unc abbidde ondrincan? cwsed ic: Ic wille 7 me leof is, gif du msege. Beda 392,32. The condition is merely a courteous form, and the sentence might properly be paraphrased and modernised thus : Do you wish us to order anything to drink ? I said. 'I should like it, if it is convenient'. To make this a subj. of 'doubt' or of 'mere possibility', would be to put a rank discourtesy into Bishop Wilfrid's mouth and that to a lady, the Abbess Hereburh, in doubting the resources of her reflectory. This subj. was probably a more courteous form of address than the indie, though there is hardly enough material to make the point perfectly clear. 2. A, modest, guarded, or tentative statement of a con- dition. Ic seolfa cude sumna brodar, done ic wolde dset ic nsefre cude, dses noman ic eac swylce gemunan mseg, gif dset owiht bryciae. Beda 442,9. cuius etiam nomen, si hoc aliquid prodesset, dicere possem. We willad nu mid sumere sceortre trahtunge f)Ees rtedinge oferyrnan, and geo- penian, gif heo hwsel digles on hyre htebbende sy. Horn. I, 388,29. The condition prepares the hearers minds for diffi- culties in the lesson, by suggesting only the possibility of their existence. Gif hwa elles secge, we sceotad to him. Horn. II, 306,19 The context from Thorpes translation i?. 3* 36 "Thus wrote Jerome, the wise expositor, concerning the holy rood, how it was found. It any one say otherwise we refer to him". Aelfric anticipates any possible disagreement, without wishing to assume its actual existence in his audience. So : We willad eow geswutelian nu cerest, gif eower hwylc nyte hwset mannes Beam sy, pset Crist sylf is raannes Beam. Hom. II, 106,38. A repetition of this sentence is in Wulf 288.4. The idea behind the following subj. is reverence, gif {)iu willa sie, wuldres aldor, J)get me wserlogan sweoi'dum aswebban, ic beo sona gearu to adreoganne, J)aet fu, dryhten niin, edelleasum deman wille. Andr. 70. So: Elene 788. 3. Ideas of doubt or of incomplete knowledge. Ac to hwon yldest pi\ J)3et jju raj)0st do, gif J)U God sy, J)fet man fas menn witnige & cwelle. Blick 138,1. Hom. I, 598.5. Gif du from J)issum msessepreoste gefulwad waare, J)onne ne eart du fulfremedlice no on riht gefuUwad. Beda 402,26. wsere because the bishop speaks only on the authority of what Herebald has just told, and casts the responsibility of the statement upon him. It is a sort of logical indirect discourse. A some what similar case, though the mood may be also iniluenced by the imperative: Gif |)is gebod wasre eow. geseald (as you say it was) fram eowerum Drihtne, asti- gad hider mid gefean on min scip. Blick 233,21. contrast the parallel passage in Andreas. 344 where we have indies., al- though the construction hardly varies otherwise except in the absence of a command. Gif ge syndon J)egnas paes pe frym ahof ofer middangeaid, ' swa ge me secgaj), 7 ge geheoldon, J)£et eow se halga bead, f)onne ic eow mid gefean feriau wille ofer brimstreamas. Similar with pret. subj. in protasis is Hom. II, 144,15. Possibly both instances should have been included under Class II. 37 4. Miscellaneons. A case where the illustration is in the indie, and the application in the siibj. There is possibly a glimmer of de- precation not of denial (Hotz 46) in the subj. conditions. They could certainly be as well expressed by the indie. The anticipation of the restriction expressed by eallunga may affect the mood of the first subj. condition, gif fara lima hwile of bif), donne ne bif) hit no full mon swa hit ter was, gif eac hwylc god man from gode gewite, donne ne bij) he J)e ma fuUice god, gif he eallunga from gode gewite Cons. ep. 37, § 3, p. 1901. Similar is, Hom. I, 292,23. A very rare use of the subj. to express ability is in- stanced in; J)onne ne reccap hi (the wood-fowl) J)ara metta, gif hi f)8es w uda benugon (may enjoy). Cons. cp. 25, p. 88m. I can see no reason for the subj. in this case; or if it is 'ideal' as seems possible, why the apodosis should have the logical form. Gif donne hwelc mon msege gesion ("poterit notarej da birhtu {)£es heofoulican leohtes mid hluttrum eagum his Modes, donne wile he cwej)an {)aet sio beorhtnes faere sunnan sciman sie J)£esternes to metanne wij) pa, ecan birhtu. Cons. cp. 34, § 8, p. 146 h. In the following example secge is probably subj., from msege. Hw£eJ)er du hit asweotolor ongiton maege, gif ic de sumne bisne get secge. Cons. cp. 34, § 4, 138 m. A condition in which the subject is indefinite is found with the subj. : Ac fajt he taeld to unalyfedlieere wiglunge, gif hwa da wyrta on him becnitte, buton he hi to dam dolge gelecge. Hom. I, 476, 4. 5. A few eases nearly related to. class II. The apodosis contains an expression of obligation or fitness. 38 Pulwian f)Oime f)set ceiiDende wiif od{)e f)a3t beam J)tet f)aer accened bi(t, gif heo syn J)reade (urgetur) mid frecenisse deades nfenige gemete is bewered. Beda 76,19. The apodosis is a negative expression of 'it is right' or 'flttiug'. We motan swyj)e wferlice on islce wisan us heal- dan, gif we sceolon us mid deofol gescyldan. Wulf 301,5. Motan plainly expresses obligation. Sceolon iii form can be indie. Gif hwam twinige be aeriste, ponne mseg he under- standan on pisura godspelle, J)8et jifer bid sod serist. Horn. I, 503,4. I incline to believe that the implied suggestion, (cf. the disguised command, Cted. Gen. 425), that the hearer learn, has influenced the mood, while I realize that it may be only a courteous and guarded foim of address as noted previously. In the instance, gif we gemete sin on mold wege, odde teor odde near fundne weordan, hset hi usic bindan and in bsel- wylme suslum swingan. Jul. 334, the subjs. may be caused by the verb of commanding hset. The elliptical construction, gif J)u wene, fast hit J)in bocland sy, J)fet |)U on eordast — hit J)onne wasron mine waeter, J)a f)e on heofenum wseron, f)onne ic mine gife eordwarum dselde. Wulf 260,2 admits of a double explanation. First that tlie condition to sy is properly concessive (so Cajd. Gen. 657 is to be explained), or second and more probable that an impei-ative is implied, as, 'If yon think it is 3'our land, remember that the waters were mine etc' Ex. 22 ^ owes its subj. to its surrounding of legal conditional sentences with mandatory apodoses. In the case: Hu wilt |)u nu andwyrdan jja;m woruld saeljium, gif he cwedan to de? etc. Cons. cp. 7, § 5, p. 221 the better Colt, ms gives the alternative reading cwedad. More space has been devoted to this small class than it deserves intrinsically. Still since the citations are complete for my own reading they shov/ how limited was the use of the 39 pres. subj. outside of the categoi'ies specified in classes II and V and I think dispose quite effectually of the 'mere hypothesis' theory as an explanation of the pres. subj. or a principle of classification. And it is noteworthy that the in- stances cited fall into small, special classes, leaving II and V the only classes that include any significant number of pres. subjs. Class IV. THE UNFULFILLED OR UNREAL CONDITION. The pret. subj. is found in both members of the sentence. The construction seems to be precisely like that in the Latin unreal condition, except that the Anglo-Saxon expressed by one tense the pret., the' time relations for which the Latin used two, the impei'. and plup. subjs. A. The condition is unfulfilled in past time. * * Scipio se consul 7 Ster frecenlice gewundod weard, 7 eac ofslagen WEcre, gif his sunn his ne gehulpe. Or. 18(J,23. Ne gefiolle be no on sutc opene scylde tet he his brodur ofslage, gif he u'r ne gesefstgode dastte his brodur lac wferou doncweoi'dlicor onfongne donne his C. P. 235,2. nisi enim Cain iiuidi.vset acceptam fralris Jiostiam, minime pervenisset ad extimjuendum vitam. Gif J)tet deofol hine ne gesawe on are gecynd, ne costode he hine. Blick 33,35 gif we hit forsuwian dorston, ne ssfde we hit eow Horn. 11, 324,23. fordi buton he drowode for us, ne mihte ure man cuman to godes rice. Hom, II, 40,29. Buton in this case has only the force of gif ne. So nefne : Grendel ac- wealde, swa he hyra ma wolde, nefne him witig god wyrd forstode and dfcs maunes mod. Beo. 1055. Fordon gif f)U fisses monnes fea in his synnum deades ne onfenge, ne burne 40 his wiite on pe. Beda 216,7. translates slavishly: Si enim huius uiri in peccatis suis mortni pecuniam non aecepisses, nee poena eius in te arderet. B. The condition is unfulfilled in present or in future time ; or the relation is a general one and consequently the time is not specified. fordsem, gif se weobed ufan hoi nsere, & deer wind to come, donne tostencte he da lac. C. P. 217,21. Si enim in altari fossa non esset, omne, quod in eo sacrificium reperiret, superveniens aura dispergeret. Gif he ongunne and anginn hsefde, buton tweon ne mihte he beon J31mihtig God. Horn. I, 276,18. Gif ealle menu on worulde rice wferon, fjonne nsefde seo mildheortnyss nasnue stede — — Horn, J I, 106,4. Nolde ic sweord beran, wsepen to wyrme, gif ic wiste hu wid dam seglsecam elles meahte gylpe wid gripan, swa ic gio wid Grendle dyde. Beo. 2578. gif du fonne swelces nanwuht nsefdest, |)onne ne f)orftest du de nanwuht ondrsedan. Cons. cp. 14, § 3, p. 461. with the subj. indie, form. NOTE. 1. Inversion. Inversion of thie protasis and ommission of tlie conjunction, is more common in these two cases tlian in any otlier. It is liardly found however with any form except w se r e , and then chiefly in late texts. Ahte ic minca handa goweaid , and moste ane tid ute weordan, wesan ane winterstunde, {>onne ic mid |>is werode. — Cajd. Gen. 367. Nseron swa manega martyras, naere seo myoele ehtnyss. Saints II, 328. Hom. I, 94,33. Ealle we scoldan forweordan ecan deade, nsere I)set Crist for us dead f)rowode. Wulf 111,6. Cons. cp. 35, § 2, p. 1561. 2. J)9er as the conditional conjunction. |)Eer is well established as a conditional conjunction in this class. I have found no examples for it in Aelfric. 41 hie will godes bearne nid ahofun, swa hie no soeoldon, fixr iiie leahtra fniraan larum ne hyrdon. El. 839 (cf. Prollius, p. 21). Dser we us selfum demden , donne ne demde us no God. C. P. 415,7. I>ii r yfel auht vva're, |)onne miiite hit God wyrcan. Cons cp. 35, § 5, p. 164 h. Sio acj'ld hine suide feorr of ealca haligro rime atuge, drer him oft da gesuino and da earfedu ne gehulpen. C. P. 37,9 : also 305,18; 355'''; 443" Cons,: cp 27 , § 3, p. lOOh ; cp. 35, § 5, p. 164h; cp. 39, § 12, p. 2341 {^xt by a misprint); cp. 40, § 5, 2401. 3. Indies in the unfulfilled condition. A condition universally recognized as contrary to fact may be expressed with the indie, thus heightening the impression of impos- sibility. Logically such a condition must be classed with the unfulfilled. Gif senig man mteg geriman J>a3re eordan dust , J)onne mieg he eac swilce geriman |)inne ofspring. Gen. 13, 16 (Wohlfahrt s. 20). Class V. CONDITIONS INTRODUCED BY BUTON. Clauses introduced by buton, nefne, or nymde, in the various senses of 'unless', 'except', 'if not' take tbesubj. We may distinguish two classes. The protasis is truly exceptive. A. It makes a reservation from the statement, or recog- nizes an element which may intrude to prevent the realization of the idea contained in the apodosis. The formula is: a thing will happen, except under certain conditions. donne singallice disse eordlican drohtunge gewuna wile to weor{)an, buton hine sio myndgung dara baligra gewita onbryrde C. P. 169,6. and geselj) eallne donne welau sefter dam anwealde, buton he hine mid la3san begitan mjege Cons, cp. 33, p. 2, p. 124 h. He gives all that wealth for power, unless he may get it for less'. Mycel yfel ded se de leas writ, buton he hit gerilite. Horn. I, 8,12 & II, 2,22. Ealle we sind gelice ajtforan Gode, 42 buton hwa oderne mid godiim weorcum foi'deo. Horn. I, 260,24. Buton Drihten da buih gehealde, on ydel waciad pa. hyrdas de hi healdad. Horn. II, 230,4. A curious use of nerane in the purely exceptive sense with the pret. subj. is; pa. wses geworden, J)3ette jjsere seol- fan neahte pa. brohton ban ute awunedon, nemne mou geteld ofer abrajdde. Beda 182,23. Then it came about, that same night the bones which had been brought remained out of doors (forisj, except (that) they spread an awning overhead. Sunnan-daga cypincge we forbeodad eac eornostlice and selc folcgemot, buton hit for mycelne nydf)earfe sy. Wulf 117,5. Gre sweltad deade, nymde ic dom wite, sodan swefnes, pSBS min sefa myndgad. Daniel 143. The condition is indispensable. B. An idea denied in the apodosis cannot be realized except through the realization of the idea expressed in the protasis. The apodosis may be formed by a negative sentence or by a rhetorical question requiring a negative answer. The latter form is rare. The formula is: A thing will not happen, except under a certain condition. dset scip wile hwilum stigan ongean done stream, ac hit ne maeg, buton da rowend hit teon, ac hit sceal fleotan raid dy streame C. P. 445,10. Eu sceal min cuman gsest to geoce, nemne ic gode sylle hyrsumne hyge? Guth. 339. Monnan ic ne bite, nymde he me bite. Hid. 66,5. We ne beod na ful- fremede buton we beon afandode. Hom. I, 170,14. ne bid nan fsesten gode gecweme, buton se man hine sylfue fram leahtrum forhsebbe. Wulf 286,12. With pret. subj. Ac hit Scipio nolde him aliefan wid nanum o{)rum J)inge buton hie him ealle hiera wsepeno age- afen 7 fa burg forleten etc. Or. 210,19, 43 KEMARKS. 1. A very few cases accur where the iiidic. follows buton etc. These were in all probability mistakes or solecisms. 1 have but one example for the prose. foifon hiom siend god on micelne brogan, fyr and hunger, buton hio to gi)de8 geleafan ser gecyrran willaf). Wulf 216,14. A case with nefne : ham cymed, gif he lial leofad, nefne him holm gestyred. Versus Gnomici (Ex) 106. Two cases from Genesis are probably only apparent exceptions. forhwon wast f)U wean, and wrihte sceome nymj)e J)U appel senne byrgdest of |)am wudu beam Jie io fie word um forbead? Csed. Gen. 880 (similar is 1401). 'Wherefore knowest thon woe — ^ except (that) thou tastedst the apple'? There is hardly any couditional force in the nyml>e clause. Nym{)e is not construed with byrgdest but with the entire clause. It is properly such a construction as is mentioned in the next paragraph and explained elsewhere. Prose would reiiuire nyml)e |):et, a collocatien instanced several times in the Beda. 2. Apparent cases of buton with the indie, where another con- junction or clause follows that conjunction immediately are oases of mixed or double construction and will be considered under the head of conlracted clauses. Odde hwfet is ure weordscipe — — buton diet we sint gesceapend lefter d;ere b'sene ures SciJ)|)endes? C. P. 249,16. 3. The few cases where buton is used in an unreal condition, as; fordi buton he drowode for us, ne mihte ure nan cunmn to Godes rice. Horn II, 20, have been classed with the unreal or unlulfilled condition. 4. Theie is scarce a trace of conditional force in this example. And nte.s senig word, J)iet Kuig man on hine funde, buton hit wiere eall sod, |)£et he siede. Wulf 205,19. where the buton clause is puperly rendered. 'There was no word — — which was not entirely true, of that which he said', or more closely 'no word, but was all true', which would seem to be a survival of the idiom. 5. The Latin fails us with regard to this construction ; for in most cases nisi conditions are freely paraphrased in Anglo-Saxon and buton conditions are formed from negative conditions or allied eonstruclicms in the Latin. 44 EXPANDED FORMS OP THE CONDITIONAL SENTBNCe. It is hardly necessary to state the obvious expansions of the con- ditional sentence from the type where each member is a simple clause. Either, or both, the protasis and apodosis may consists in a compound clause. Ac gif sio pynding -wierd onpennad, odJe sio wering wird tobrocen, donne toflewd hit eall, & ne wierd to nanre nytte, buton to fenne. C. P. 277,8. An apodosis may be restricted by more than one condition. In this case one is subordinate to the other either in time or importance. Both conditions have the indie. gif he yfle ded, Isesse wite he drowad on helle, gif he ana dider cymd, donne he do, gif he oderne mid him dider bringd. C. P. 33,1. gif seo donne bid eac fennegu, donne is wen dffit hio da odre wiers besmite gif hio hire ahrind. C. P. 75,23+. Gif 86 synfulla bid gebrocod for his unrihtwisnysse, {lonne gif he mid gedylde his Drihten herad, and his miltsunge bitt, he bid donne adwogen fram his synnum. Hom. I, 472,2. Gif he rihtwis bid, he hsefd J)onne maran gedincde furh his broo- nunge gif he gedyldig bid. Hom. I, 472,6. Gif de mau scotad to, J)u gescyltst de, gif J)u hit gesihst. Hom. II, 538, 10. f)urh da fandunge we Bceolon gedeon, gif we jefre widsacad deofle, and eallum hislarum; and gif we genealsecad urum Drithne mid geleafan, and lufu, and godum weorcum. Hom. I, 170,15. Two conditions one in past the other in present time with the same apodosis. Gif he nn for us arisan ne wolde, micle ma gif we him under|>eodde beod, he us eac for noht gehyged. Beda 102,18. Change of "moods to set off a condition from a subordinate one" (Hotz 58) is found rarely. Hotz has been led to emphasize this excep- tional construction from the examples in the Laws and in the Legal part of the Old Testament, which hardly correspond to the usage of literary prose. Possibly the following instance is an example of this change of mood to set oif conditions of different grades. Gif hwa gonge bilewitlice mid his fiiend to wuda treow to ceorfanne, & sie aos awient of deem hielfe, & swa ungewealdes ofslied his geferan, he donne sceal fleon. C. P. 165,24. but I prefer with Flelschhauer (p. 26) to interpret the indicatives as caused by a momentary dropping out of the construction into lively 45 description of the event in mind. The following example seems to show clearly the kind of change of luood desribed previously. S;e^e he, gif lie a3t leorninga Siete odJe elles hwiut dyde, gif senininga mare blad windos astah (adsurgeret), |)ict he sona insttcj>e Drihtnes mild- heortnesse gecegde 7 |)a miltse bied monna cynne. Beda 268,15. In the following instance the first condition has concessive force; this rendering is confirmed by the I)eah in apodosis. gif {>u him heodfeg wuht hearmes gespffice, he forgyfd hit J>eah, gif we him geongordom Ijestan willad. Cfed. Gen 657. Two conditions one with gif and the other with buton may be found with the same apodosis. Gif in de hrador ne gewitst fram Jacobe, and buton du wyrige Gristes naman, "pa scealt beon beheafod samod mid him. Hom. 11, 422,21. fonne gif he mid deofles weorcum hine sylfne bebint, donne nt mseg he mid his agenra mihte hine unbindan, buton se Jiilmihtiga God mid strangre handa his mild- heortnysse hine unbinde. Hom. I, 212,12. With subjs in both clauses, ac fset he tteld to unalyfedlicere wiglunge, gif hwa da wyrta on him becnitte, buton he hi to dam dolge gelecge. Horn. 1, 476,4. V. THE CONDITION EXPRESSED BY FORMS USUALLY NON -CONDITIONAL. The innumerable lines of association of ideas in language prevent the perfect crystallization of thought into fixed forms of expression. Grammatical categories mark merely centers of radiation, the fields of which overlap, unite or conflict in ever changing combinations. We expect to find and do find to the main idea of a construction appended or supperadded many minor functions, some of them apparently widely different from the original function. No construction affords a better example of the vicissitudes of form and of meaning than the conditional sentence in its many points of contact with other constructions. A condition is itself subject to, or may even consist in relations of time, manner etc. ; so we shall not be surprised to find temporal, and modal clauses etc. which retaining little or much of their primitive force have gone over to an essen- tially conditional meaning. Such secondary or double con- structions are highly interesting in the study of style. I have attempted to classify those that touch upon the conditional sentence in Anglo-Saxon under appropriate rubrics at this place. 47 1. RELATIVE CONDITION. "A relative with an indefinite antecedent gives a con- ditional force to the clause in which it stands, and is called a conditional i-elative" (Moods & Tenses p. 197). Such a construction is of course very common. The in- troducing relative is usually se de or swa hwa (hwylc) swa. A few instances of the translation of Latin conditions by a relative, and the reverse will show how interchangeable the constructions are, and that se de is equal to si quis. Se de biscephade gewilnad, god weorc he gewilnad. C. P. 53,7. Si quis episcopatum desiderat, honum opus desiderat. se dft God no ongit, no ongit God hine. C. P. 29,2. Si quis autem ignorat, ignorabitur. Se de Godes degn sie, ga hider to me etc. C. P. 383,1. Si quis Domini est, jungatur mihi. An instance of class 11 which properly takes the subj. se de godes cyrican, he (68) cwsed, rype odde reafige odde halig- nessa grid brece, k he forwyrde. Wulf 67,24+. Si quis ecclesiam dei demidaverit vel sandimonia uioluverit, anathema sit. cf. Rsede se de wille. Rid 60,16. and Se de me J)enigp, fylige he me. Hom. 160,3; Odsace nu . . se fe wille oJ)J)e se f)e dyrre |)aet angin iifere gistilled — Or. 206,6. A i-ather curious use of J)e representing gif is : Hu se Wisdom Iserde {)one, J)e he wolde waestmbereland sawan, fset he atuhge of £erest |)a J)ornas etc. Cons. Tit. cp. 23, p. XIV. The more usual form is represented by : Hu se Wisdom hine Iserde, gif he fsest hus timbrian wolde etc. Cons. Tit. cp. 12, p. XII, also f)e {si) C. P. 383,34. Swa swylc hyse-cild ne bid ymbsniden on J)am fyl- mene his flasces, his sawul losad. Hom. I, 94,12. de pe wille 48 fullice anweald agan, he sceal tilian serest J)set he haebbe an- weald his agenes nioodes. Cons. cp. 29, § 3, 106 m. Swa hwa swa wille bion disse weorolde freoiid to uii- gemetlice, he bid gehaten Godes feond. C. P. 421,34. 357,4. Swa hwa sprece, sprece he Godes worde swelce da woTd no his ne sien, ac Godes C. P. 371,15. Si quis loquitur, quasi sermones Dei. J)onecan pe seems to mean 'nevertheless if in the following'case : Twa ding mseg se weoiJ)Sci{)e 7se anweald gedon, gif he becymf) to J)am dysigan. He mseg hine gedon weoif)ne 7 andryse ofrum dysgum. Ac J)onecan J)e he {)one anweald forlset, odde se anweald hine, J)onne ne bi{) he nauj)er J)am dysegan ne weorf) ne andrysne. Cons. cp. 27, § 1, p. 94m. A few instances of translation of a Latin relative by gif. HwKt se donne unryhtlice talad, se de talad dset he sie unscyldig, gif he da good, ds us God to gemanan sealde, him synderlioe segnad. C. P. 335,11. — qui commune Dei munus sibi privatum vindicant. for dsem nan men ne bitt oderre dset he hine reere, gif he self nat da3t he afeallen bid; ne eac se, se his wunde sar ne gefret, ne wilnad he nanes Iseres. C. P. 441,9. departs from the Latin in the first instance and follows it in the second, quia nee levari appetit, qui et hoc ipsum, quia cecidit, neseit, et qui dolorem vulneris nan sentit, salutatis remedia non requirit. 2. TEMPORAL CONDITION. Temporal conjunctions nu, f)onne, ser are used to intro- duce conditional clauses. a) uu. Nu {si autem) he fara laessena rica reccend is, hu mic-le swidor wene we fset he ofer fa maran sie Cr. 58,25. Hwa mseg donne for dyllecum bion gehealden, buton miclum 49 geswince & miclum plio, nu se on dsem rice on swelce synne befioll, se de God self to f)8eui rice geceas? C. P. 393,10 Quis — — fiaJvahilur, ni ille hi hia culpa iiiteroenieiite tiir- hatus est, qui ad hmc fuerat Deo elegente preparatiis. cf. C. P. 393, 8,10; 383,16 and 377,21, the latter possibly rather causal than conditional. Ono nu is found in Beda. Probably nu is the conditional conjunction and we may accept Miller's statement, XXXI, "It is clear from the table that ono never means 'if, though Bede has been quoted in support of this sense" Ono nu nu faem mete ne bid clsene, f)am f)et mod ne bid claene, forhwon J)onne f)set wiif J)8et heo clsene mode of ge- cynd J)rowad, sceal hire in unclsenesse geteled beon? Beda 80,19. Si ergo ei cibus immundus non est — — cnr ei ill immunditiam reputetnr. So. Beda 78,J.4, 27; 84,6. Nu (si) alone. Beda 234,29. Eornostlice nu ge pe yfle synt, ciinnau gode sylena eowrum bearnum syllan, mycle ma eower Faeder syleS god Jiam de hyne biddad Mat. 7,11. Si ergo vos • nosti — dare. (R has nu nu J)onne.) Nu follows the subject, an unusual order. Ac we nu haefdou senigne dset untriogendes andgites swa swa englas habbad, J)onne mihte we ongiton {)a3t f)8et andgit bij) micle betere donne ure gesceadwisnesse Oons. cp. 41, §5, p. 254 h. Si igitur uti rationis participes sumus, ita — — iudicium — — habere possimus. I have not found a conditional nu in Aelfric or in Wulfstan. b) |)onne used of indefinite time may introduce a condition in the sense of 'if ever' or even with slighter temporal force 'in case that'. donne ahebbad da synfuUan swide np hira hornas, donne hi hi nsefre nyllad geeadmedan to dsem dset hie ongieten hira unrybtwisnesse, & da hreowsian. C. P. 425,23. Cornu quippe 50 delinquentes exaltant, si neqiiaquam se ad poenitentiam ex corjnitione suae iniqiufatis hmniliant. C. P. 265,4. Witodlice lie bid he J)8era sehta cleowa, {)Oiiiie he hi daslan ne maeg; ac he bid jisera nehta hlaford, f)oniie he him eallunga peowad. Horn. 1, 66,5. Hwiet fremnd him f)a3t J)set he hasfd, jjoiine he done (God) usefd J)e him da god forgeaf f)e he hsefd? Horn. 11, 410,22. These examples will suffice. Such temporal conditions aie common enough in all texts. It is difficult to distinguish sharply between purely temporal and temporal- conditional sentences in this case. c) fonne in a conditional sense may introduce a clause which expresses the concrete duty or phase of conduct that corresponds to a biblical type. This construction is tediously common in C. P. It was ratlier a favorite may of preaching, — and its end is not yet — tliis often far-fetched interpretation of all biblical circumstances as types. A page of such con- ditions may be found from C. P. 161,10. donne hie stondad up on dsem munte, donne dset flaesc nauht elles ne seed to dsem odrum buton tudor. C. P. 399,4. Eft, donne we his menniscnysse mid lule underfod, J)onne ete we J)£es lambes fet. Hom. II, 280,11. The following passage from Aelfric shows that gif and J)onne were interchangeable in this case. Buccan we offriad, odde ticcen, gif we ures lichaman galnysse oferswydad. Culfran we offriad, gif we sode bile- witnysse on uruni mode healdad. Turtlan we offriad, gif we on clsennysse wuniad |)eorfa hlafas we bringad Gode to lace, donne we buton yfelnysse (212) beorman on deorfnysse syfer- nysse and sod fsestnysse farad. Hom. II, 210,31+. d) Aer may be used to introduce a temporal condition in nearly the sense of buton. (cf. Fleischhauer 85). do3t mod dsette ne mteg gesion da flane ser hit sie ge- 51 wuudad, hit bedearf dset hit hsebbe simie on honda done scield Godes eges. C. P. 431,3. so 331,21; 445,2. e) Mid dy as a temporal condition, translating cum, may be noted here. Mid {ly f)onne seo kifa ne bid tudres to tilieune, ac se willa ma wealded in pum weorc fsere ge- mengnisse, f)onne habbad pA gesinhiwan dearfe be heora ge- mengdnesse, |)8et hi wepeu 7 hreowe don. Beda S'2,25; so 86,14. "Wid f)am: se deofol befran gif he wolde on hine gelyfan, wid J)am fe he gefremode his fulan galnysse. Saints III, 368. 3. CONDITIONAL SUBJECT CLAUSES. A subject clause, introducing a purely hypothetical or unreal idea, may be in a truly conditional relation to the verb of which it is the logical subject. The introducing conjunction is J)8et and the verb is always subj. There is often a comparative in the independent clause. Se de aenige dissa ierminga besuicd, him wsere betere dtet him wsere suma esulcweorn to dam suiran getiged, and sua aworpen to sses grunde. C. P. 31,2. In the Gospels: nyttre him waare J)8et an cweorn-stan sy gecnytt abutan his swiran 7 si on sse be- worpen donne he gedrefe senne of J)issam lytlingum. Lk. 17,2. utilUus est illi si lapis nialoris imponatur circa collum eius et projicalur in mare, quam etc. J)set him wsere ealra msest un- SEelJ), J)8et f)8et se fyrst wajre od domes deeg. Cons. cp. 38, § 4, p. 204 m infelicissimam vero, si esset. ceterna. ForJ)am sume menn wena|) |)8et f)set sie seo seleste ge- sself), |)set mon sie swa selig |)£et he manes J)inges maran ne J)urfe Cons. cp. 24, § 2, p. 82 h. Him wsere swa-deah betere |)8et he forburne fonne he tetburste. Hom. II, 480,8; 244,5. also C. P. 31,23; 33,29; 433,29; (cf. Pleischhauer 6—7). 4* 52 4) J)8er with the pret. siibj. has been classed with the unfulfilled condition. Fleischhauer {[). 27) regards the following construction with dser as wholly temporal. Its antithesis with the con- cessive clause, — Fleischhauer has stopped his citation at idelu, — certainly marks it as a temporal condition, similar to those with fonne. Sua eac ctser S.iet heafod bid unhal, eall da limu biod idelu, deah hie hal sien. C. P. 129,7. 5. CONCESSIVE CONDITION. feah in a conditional rather than a concessive force is instanced rarely. ga to J)inum huse 7 deah f»u on tun ga, nsenegum f)U hit ne sege. Mk. 7,26. Vade in domuni tuam, et si inuicum introiens, nemini dixeris. There is no true adversative sense in deah or el si, they simply have the force of 'if and mark a somewhat remote condition. The glosses B, & L trans- late by gif. Is J)set hwelc wundor, f)eah de he f)one dseg his deades odf)e ma f)one Drihtues dseg blide gesege, JDone he symle sorgende bad, od fset he cwome? Beda 268,6. non autem mirum, si diem mortis uel potius diem domini Icetus aspexit, quern semper, usquedum ueniret, sollicitus expectare curauit. feah de israhelitisc folc beo swa maenig fyld, swa swa sandcysel, f)e lid on ssestrande, |)8et fe {)onne to lafe bid, hit bid gehealden Wulfstan 198,21. si fuerit numerus filiorum Israhel, sicut arena maris, reliquie salve fient. Nis hit nan wundor deah hwa (si credatur) wene etc. Cons. cp. 39, § 2, p. 214h. 53 6. THE MODAL CONDITION. The modal conjunctions swa and swylce (swa swa) may introduce a conditional clause. a) swa or swa swa in the sense of 'provided that' is extremely rare in Anglo-Saxon. For examples of 'so' as con- ditional in Mid. and Mod. Eng. cf Matznerlll, 484. He has no examples for A. S. The following examples are all that I have been able to iind. Swa swa sio nafu bid simle swa gesund, hnseppen da felga on f)8et de hi hnseppen. Cons. cp. 39, § 7, p. 2221. Ic gelyfe to fe, min liifruma, faet J)U mildheort me nsefre wille an foilgeten, swa ic \>set gefremme, J)enden feorh leofad, min on moldan, \)set ic {)inum larum leofwendum lyt geswice Andr. l^t'O (Renssner p. 21). From Bosworth-ToUer I cite: He him det land forbead, swa he amiges brucan wolde. Charters (Thorpe) 202,10. Nim, swa hit de ne mislicyge. Apol. of Tyre (Thorpe) 20,12. To which I may add doubt- fully, he J)e mseg sod gecydan, onwrecan wyrda geryno, swa pa hine wordum frignest, ariht from orde od ende ford. El. 587. b) Swylce (swa swa and swa) in the sense of 'as if usually takes the subj. We iiave here it would seem a double, or compressed construction at once modal and conditional ; and swylce is to be resolved into 'as it would be if — . Such modal con- ditions are unreal or purely hypothetical. Swelce he openlice cusede. C. P. 11,12 (ac si aperte dicat) so 465, 20 and 35. Hu ne bid he donne swelce he saewe good, & him weaxe of dsera selc yfel? C. P. 341,6. Hwi raurcnast f)U wid min, svvilce J)U for minum dingum seo dines agnes benumen. Cons. cp. 7, § 3, p. 20 h. his freonde forspsec forstent him eal J)8et sylfe, swylce hit sylf sl)a3ce. Wulf 38,16. so 301,12. Also Wulf 53, 11 3, 54 8, 18, 25. swa swa. Eft heo, swa swa heo leohtlice gebyl- ged wsere, asfter f)on heo cwted. Beda 290,18. „quasi leuiter indignata. So Beda 122,15; 174,29. swa f)a wass he meted swa unseeded, swa he in f)a ilcan tid of J)issum leohte gelseded wtere. Beda '218,25. Swylc swa. J)is andwearde lif is swylc swa f)U set swaesendum sitte Beda 134,26. Ca. reads, swa gelic swa. In the case swelce cTu woldest da lean eallra f)inra godena weorca on JDisse worulde habban. Cons. cp. 7, § 3, p. 22 m. Woldest is probably the common subj. indie, form. To my feeling the modal idea is much stronger than the conditional in such sentences, and the construction is often merely a less vivid way of expressing a comparison. Still it approaches the condition near enough to be noted in this connection. 7. DATIVE ABSOLUTE. A Dative absolute is found only twice as a protasis. The construction is surely not native. J)a3t wif he onfeng fram hyra yldrum J)aere arednesse, (ea conditionej f)8et hio his leofnesse hsefde jjset heo J)onne f)eaw J)8es Cristenan geleofan healdan moste. Beda 58,13. The dative absolute is elucidated by a condition intro- duced by dser. da ge|)afedon he dsere arednesse 7 him wif sealdon, f)£ct dser seo wise on tweon cyme, J)£et hi donne ma of f)am wif cynne him cyning curan J)onne of pam wscpned- cyune. Beda 28,19. ea solum conditione — — ut etc. 8. PHRASE WITH BUTOIST. Buton as a preposition, 'without', with a relative is equal to a protasis for a conditional sentence. 55 Secead sibbe & god to eallum mannum buton daore ne mseg nan man God gesioii. C. P. 345,15. sine qua nemo videbit Deuin. Geleafa is ealra niEogena fyrmest; buton J)ara, ne mgeg nan man gode lician. Horn. I, 134,2. The buton phrase in each instance is equal to a clause: If a man be without these he may not etc. So Horn. II 526,6. A similai- use of a buton phrase as a condition is : fordon {)e nan wyrhta ne mseg gOd weorc wyrcean for Gode buton lufon & geleafau. Blick 111,14. Ne msBge ge nan ding don butan me. Horn. I, 310,1.S. Butan J)ani sehtum, gelice sind f)a dser biddad, and da de hi setbiddad. Horn. I, 256,4. 9. TWO SENTENCES AS MEMBERS OF A CONDITIONAL PERIOD. A sentence may state the condition for a following sen- tence and so form a protasis. This has been illustrated already for imperative sentences. Aerest hi sculon ongietan dtrt hi fleon dset dtrt hi luflad. donne magon hi sid iedlice ongietan da?t dset is to lufianne dset hi fer flugon. C. P. 441, 13. TI. CONTRACTED AND ELLIPTICAL CLAUSES. Anglo-Saxon as it has come down to us is rather an in- flexible language and little tolerant of contraction and ellipsis. Middle and Modern English offer a marked contrast in this respect. Outside of the exceptive clause with buton (nemne) we shall find little contraction and only the most obvious ellipses ; in most cases where the immediate context supplies the exact form of the missing member. CONTRACTION. A contracted clause is one that lacks a member, usually the verb, which must be supplied from the context or otherwise. CONTRACTION OF BUTON CLAUSES. Contraction of a protasis with buton (nemne) is very common, by the ommission of a verb. The verb of the apodosis is to be carried over to the protasis. Porfon ne sceal biscopa halgung on odre wisan wesan, nemne in gesomnuug 7 in gewitscipe freora odfe feower biscopa. Beda 72,16. Se feorda is: J)gette munecas ne leoren of stowe to oderre, ne of mynster to odi'um, nemne 57 |)Urb leafnesse his agnes abbudes Beda 278,13. 'unless he go by his own abbots leave', flwset gelyfj) se licharaa biitan |)urh fa sdwle? Blick 21,21. ne nsenig his ageii wiif forlste, nemne swa pset halig godspell laereS, for iutingan dyrne for- legenesse. Beda 280,1. Nullus coniugem propriam nisi, ut sanctum evangelium docet, fornicationis causa, relinquat. The Latin seems to show that the swa clause is parenthetical and that nemne is construed wich the phrase and not with the clause. ForJ)on we witon swif)e lytel fees J)e £er us waes, buton be geniynde Cons. cp. 42,2561. Hwaet is J)aet rice |)a?t he betgecd his Feeder, buton da halgan, meiin, etc. Horn. 1264,9. Buton (nemne) is followed immediately by a clause sub- ject, relative, temporal, causal, or modal, wich clauses have the mood appropriate to themselves invidually. We seem to have compressed into the one word buton tbe idea of a clause upon which the following clause is in different relations sub- ordinate. The complete, logical expression is "Unless it be, 'that' — , 'he who' — , 'when' — , 'because' — , 'as' — . As in the preceding case the verb for the buton clause is often to be supplied or carried over from the apod. We have in such cases then contracted, complex clauses. It is quite un necessary to suppose that these forms arose historically, within Anglo Saxou, from completely expressed complex clauses. The contraction may well enough have been coincident with the formation of such complex clauses. That these are cer- tainly logically contracted clauses, and that buton f)8et etc is not a mere formula introducing the whole clause is proved by the fact that such clauses have generally the indie, while buton requires the subj. Subject clause. Odde hwtct is ure weordscipe — — buton dset we siut gesceapene sefter dsere bisene ures Scippen- 58 des? C. P. 249,16. He wsere wierde ealre Roraana onwaldes for his monigfealdum dugiidum, buton JDaet he pa. wij) his hlaford won for odra monna lare. Or. 292,15. Relative clause Hwasdres d'omie dara yfela is betere ser to tilianne, buton swgedres swseder frecenlicre is? 0. P. 457,21. Non man ne astihd to heofonum, buton se de of heofonum astah. Horn. II 378,1. Temporal clause. & deah ne biod na gemengde, buton donne hi wilniad beam to gestrienne. G. P. 399,3. Causal clause. Porhwy bid se synfuUa onbryrd mid d^re breowsunga buton fordy de dsem synfuUan nauht ne helpad his godan gedohtas. C. P. 423,22. buton tordeem dset — C. P. 397,8. Clause of purpose with subj. ne gsed na mare to metinge buton J)8et |)u hit geseo and herige. Horn. I, 186,6. Very rarely buton and f)Eet are separated by the verb. Ac hwonon wurde J>u mid f>i.ssum woruld sorgum J)us 8wi|)e geswenced, .buton, ic wat, |)8et {)ii bsest dara waspna to hra{)e forgiten Cte ic pe aer sealde. Cons. cp. 3, § 1, p. 4m. Contraction of Clauses which normally require gif is rare. Even the carrying over of gif from a preceding clause, as instanced in the following example, is rare. For|)3em gif hit swa naere, J)onne nsere he J)8et |)ffit he gehaten is,' oJ)f)e asnig J)ing serwsere, oppe selt^wre, donne \Wtere J)£et betere donne he. Cons. cp. 34, § 2, 136 h. THE CONDITION EXPRESSING PURPOSE. A peculiar construction which expresses purpose with a conditional restriction is found in Anglo-Saxon. With regard to its survival in biblical English I can do no better than to quote Goodwin's 'Moods and Tenses' p. 180, Notel. "The English translation of certain conditional clauses in the New 59 Testament which have tliis peculiar construction preserves the sense of purpose or desire with the oiiginal form of pro- tasis. Thus, 'that they should seek the Lord, if haply they might feel ofler him and find him' Acts XVII, 27; and 'he came (to the tig tree) if haply he might find anything thereon' Mark XI, 13". -^Xdev eI S,pa n suf/^iasi sv am% Tlie follow- ing citations will show that the construction was already established in an early period of Anglo-Saxon. Cuomon hio to Ealdseaxna msegde, gif wen wsere {^tet he |)fer senige {)urh heora lare Criste begytan mihte. Heda 414,17. Ueuenint ad prouinciam A)itiqiioriiin Sicconwn. si forte aliqtios — — Chrislo adquirere ^wssent. Similarly with si forte. Beda 408,25; 366,27. da ahof hine up 7 ongan aweg gan, gif he hwaer {sicubi) senigne freond metan meahte, Beda 326,11. Similar is, 7 fast he wses Isedende in stowe halges gebedes, gif he dam untruman brodrum ful- tum sealde [tribueref) Beda 370,4, 326,11. gif is equi- valent to f)ylge.s : breac ealde healsnnge, gif hie hwylcne drycrseft htefden pset hie hine oferswidan 7 beswican sce- olden. Beda 58,21. ELLIPTICAL CONSTRUCTIONS. In these cases a complete clause must be supplied from the context to complete the sense of the conditional sentence. Ellipsis is piactically always of the apodosis ; for an apodosis is grammatically complete by itself and suggests no other member. This ellipsis is very rare. FORMULAS OF ADDRESS. Gif f)u J)Onne dset gemet habban wille 7 fa nydfeaife witan wille, fonne is fset mete 7 drync — — fset fe is 60 Gecynde. Cons. cp. 14, § 1, p. 42 h. some such apodosis as 'know that', 'I tell you', 'you may learn' or the like is to be supplied, upon which the apparent apodosis depends. The following instances, cited in another connection, and which appear to fall under this head are possibly rhetorical questions as in the Latin original rather than rhetorical conditions. The difference in meaning is slight, but if the clauses stand in the relation of question and answer it is unnecessary to assume ellipsis. The Latin favors this interpretation, the J)onne repeated through the clauses the other. A frine hie mon f)onne gefter hu monegum wintrum sio sibb gewurde — — : J)onne is f)3et sefter L wintra 7 feower hundum. Ahsige ponne eft hu longe sio sibb gestode; J)onne wses {)3et an gear. Cr. 182,16. Et hoc post quantum tern- poris? post annos quadrigentas et quadraginta. Quando? Anno uno. IXTRODDCED BY OddE OR ELLES. Odd"e or elles introducing a clause, which is not a mere alternative, are equivalent logically to a negative protasis formed fi'om the immediately preceding statement. They are often used to introduce a clause expressing tlie result of the noD-fulfulfllment of a command. Geoffra dine lac urum godum, oflde jju bist mid eallum dissum pinung-tolum getintregod. Horn. I, 424,21, 426,6. And tordi ic sprece de he me het, elles ic ne dorste on his andwerdnysse sprecan. Hom. I 462,31 ; II 310,6. C. P. 455, 12. The full expression of the condition which is always implied in these words is found in the following instance. Gif hit hwset elles bid, ic hi aedlice mid Godes willan eft to eow hweorfe. Beda 366,6. 61 MISCELLAXEOns INSTANCES. The proper apodoses are always easily supplied from the context. Synile ^e habbad J)earfau (to whom yon may do good), gif ge willaf) teala don, ac ge nabbaj) me symle. Blick 69,16. Gif J)U wilt OTigeotan & gef)enecan hu mycel hine be- leah faet he losode (you will see), pe leas he me yfel sacode wid God. Blick 189,'23. With an intoi'rupted construction. Huru gif he cwsede J)set he naere sumum odrum mannum gelic — ac he cwaed etc. Hom. 11,428,21. So C. P. 25,21. A PROTASIS IN PRESENT TIME WITH AN APODOSIS IN PAST TIME. Snch a conditional period is strictly illogical and we must always assume a mental change of the time of the apodosis; practically the foi'mation of a new apodosis descri- bing a continuous action in present time. J)set waes sif)|)an Crist geboren wses jjset we wseron of selcum f)eowdome aliesde 7 of selcum ege, (and are released) gif we him fulgongan vvillaf). Or. 214,22. So, J)isne anweald forgeaf Crist para apostolum and eallum bisceopum, (and grants it) gif hi hit on riht healdad. Hom. I, 232,33. So Hom. I, 22,25. VII. THE CONDITIONAL SENTENCE IN INDIRECT DIS- COURSE AND SUBORDINATE POSITION. An apodosis may consist of a subordinate clause with tlie verb in the tense and mood appropriate to the construc- tion. This may induce a change in the tense and mood of the verb of the protasis. When the apodosis is the object complement of a verb of saying or thinking, it is said to be in indirect discourse. Further a subject clause or the various sorts of adverbial clauses may form apodoses in a conditional sentence. Special studies of indirect discourse and of the adverbial clauses are much to be desired. At this point I can only classify such conditional sentences according to the nature of their apodoses, exhibit the uses of mood and of tense in such constructions, and present the material for several important texts. The temporal and causal constructions, since they induce no change of mood in the conditional sentence have been disregarded at this point and included under the general classification. It will be noticed that in dependent constructions in present time there is a tendency to isolate the conditional' clause, the verb of which often remains as in direct discourse, unaffected by the change of mood in the apodosis. 63 A. THE Al'UDn.SLS IS DKl'F.N'DEN'l' UPON A XKUB IN PRESENT TIME. 1) Indirect discourse. The apodosis is the object complement of a verb of saying or flunking, and is usually introduced by {)8et. a) Pres. indie, in both protasis and apodosis. This unal- tered form of the conditional sentence is frequent. Nu is eow gessed J)aet ge sind gebrodra, gif ge done bend healdad sodre bi-oderraedene untobrocenne. Horn. II, 318,4. Ic wene |)8et J)as word ne sind eow full cude, gif we hi openlicor eow ne onwreod. Horn. 1, 588,26 & r)2«,in. C. P. 145'; 377'; 397'"; 3R:V'^ Cons. cp. 7, § 2, p. 181. Blick 21,7; 89,15; 181,32. The apodosis may be introduced by hu, hwelc, or hwseder. dserbufan is geteald hwelc he beon sceal, gif he untajlwierde bid. C P. 53,10. C. P. 57,14 (hwaeder); 161,15 & 315,15 (bu). Cons. cp. 27, g 2, p. 96 h and cp. 30, § 1, p. 108m (hu). Horn. 11, 98,14; 282,28. With a buton clause as apodosis. Ac wife gehwa to gewissan, buton he his flsesclican lustus and gal- nysse gewanige, |)aet he ne hylt bis cristendom mid rihtum bigenge. Horn. I, 96,2. b) The verb of the apodosis is pres. subj. that of the protasis pres. indie. Horn. II, 344,33. We gelyfad f)8et he msege beon gehealden, gif he his synna mid wope behreowsiad, and be lareowa tsecnunge hi gebet. Horn. I, 292,23. 1,124,14; 268,34. II, 146,34; 316,27; 344,33; 423,30. C. P. 425,1 (hwseder). 64 fordsem he wenjj J)8et he hie f)onne ealle hsebbe, gif he hsefd jiaet pant he swij)ast wiliiaf) to begitaiine. Cons. cp. 24, § 4, p. 86 h. A buton condition which as always has the subj. in protasis. Horn. I, 26,17. Coffs. cp. 41, § 2, p. 244m. c) The pres. subj. is found in both clauses. weiif), gif he donne lust begite 7 J)£et J)Urhtio, {)set he donne getihhod hsefd, j[)8et he f)onne hsebbe fiille gesselfa. Cons. cp. 34, § 2, p. 144 h. nu ne willad reccan, gif he dser suelc to cyme, hu he dseron libban scyle C. P. 73,22. Ac gesaga me hwylce mede J)U wille syllan fam men, gif hwiic sy etc. Beda 128,25. The apodosis is dependent upon a verb of promising wishing fearing etc. The verb of the apod, is subj., that of the prot. indie. — — hie magon ondiaedon dset him weorden da wyrt- truman fsercorfene on dys andweardan life, gif hie for hieia giemeliste nyllad beran da bleda godra weorca C. P. 339,19. The apodosis is a subject clause. With impersonal constructions. a) The indie, is found in both clauses. Oft eac gebyred dsera de him segder dissa ondraedad, gif he wid segder gehealdad, dset he befeallad on idel gielp. C. P. 457,18. After the substantive verb and an inflected infinitive, C. P. 201,15; 231,4; 233,16; 301,1. Horn. I, 48,33. 11, 278,1. b) The verb of the apodosis only is subj. Be dam is to smeagenne hu micclura se rihtwisa mid eadmodre heofunge God gegladige, gif se unriht- 65 wisa mid sodre dsedbote hine gegladian maeg. Horn. I, 342,13. C. P. 313,21. f.) The verbs of both clauses are in the pres. subj and eow diucj) feah J)set ge hisbban ece are, gif ge uisegen on eallre eorfene worulde geearnian f)3et ge habban godne hlisan setter eowrum dagum. Cons. 17, § 3, p. 66 h. Beda 134,20; 190,25. A condition of the form noted in Class IB, 2 appears is a subject clause with the pres. subj. in apodosis and the pret. subj. in protasis. Oder is fast hwa rice beo, gif his ^Idrau him sehta becwsedon, oder is, gif hwa J)urh gytsunge rice gewurde. Hom. I, 256,20. Class IB 1 with pres. subj. in apodosis. C. P. 57,22. Two instances of appositive or supplementary clauses as apodoses may be included at this point. With pres. subjs. in both clauses. J)reo hund biscopa — — ge- setton done canon, "pset nan raiesse preost — — on his wununge wif- liades inan rtebbe, buton hit sy his modor, odde sweastor etc. — — : ind gif he dearnunga oMe eawunge wifes bruce, Jiaet he his hades lolige. Hom. II, 94,29. Pres. subj. in apodosis pres. indie, in protasis. Micel mildheortnys pasa Metodan Drihtnes, J>a3t we beon gecigede 3wa gesajliglice ures Scyppendes frynd, gif we his hsese gefyllad etc. Hom. II, 316,21. Here should perhaps be classed the construction in which the apodosis appears as the complement of a passive verb which has already an apparent subject. The logical and grammatical relations of this construction can hardly be satisfactorily defined ; and I class it here because of its obvious nearness to the impersonal construction with a dative, in which the apodosis is the logical subject of the im- personal verb (cf. the citation under b). The construction appears in C. P. only after the inflected infinitive to manienne, and the subj. appears in both clauses of the conditional period. 66 Forasem hi sint to luanigenno, gif hie da halwendaii forhcefednesse gehabban ne ujiegen, & — — — — divt hie wilnigcn dxve hyde tta-s gesinscipes 0. P. 101,31; so ac.o.U; 275,1; 349,18. The apodosis Is a clause of purpose. Both clauses have the pres. subj. we |)8es sculon hycgan georne, jf)aet we on Adame, gif we sefre masgen, and on his eafrum swa some andan gebetan Csed Gen. 396. The verb of the apodosis is subj that of the protasis indie. Negative clause of purpose: dylaes him dy wirs sie, gif hie da trumnesse dsere Godes giefe him to unnyte hweorfad. C. P. 247,4. Both clauses have the subj. dylaes he iinde da duru betynede ongean hine, donne he cume, gif he cume idelende to, & da mid him ne brenge etc C. P. 379,20. Beda 76,25. In the following clause of purpose sceolde seems to have lost all sense of a past tense and to have merely modal force. f)set man Godes deow ne sceolde on him sylfum truwian, ne mid waepnum winnan wid woruldlicum cempum, gif he Cristes fotswadum filigan wile. Hom. II, 248,6. The apodosis is a consecutive clause. The verb of the apodosis pres. subj. that of the protasis is pres. indie. Se donne bid siwenige se de his &git bid to don beorht scinende, dset he msege ongietan sodfsestness*^, gif hit donue adistriad da flseschican weorc. C. P. 67,24. Cons. cp. 40, § 1, p. 201. Ideal and unreal conditions dependent upon a verb in pres. time, retain their usual form with the pret. subj. in both clauses. 67 Hwit't ge witon [xiet ge giet todtege wneroii Somnitum J)eowe, git' ge him ne aliigen iowra wedd 7 eowre af)a.s. Or. 122,11. C. P. 51,11; l-'57,2; 306,2; 311,13. Horn. 1, 216,6. \ B. THE CONDITIONAL SENTENCE DEPENDENT UPON A VERB IN THE PRETERITE. Weak verbs show almost invariably ambiguous forms. Unambiguous forms are almost without exception in the pret. subj Apparent violations of the rule of sequence of tenses have been treated in Section YI under the head of ellipsis. 1) Indirect discourse. a) Pret. subjs. are found regularly in both clauses, stede, gif he hine underbEsc besawe, J)3et he sceolde forlaetan fiset wif. Cons. cp. 35, § 6, p. 170h. ssede J)set hit psem cyninge Isesse edwit wane, gif J)SBm folce buton him fagiet misspeowe, swa him ser dyde. Or. 82,33 '&c. 194,10. C. P. 93,8; 355,18; 405,16. Beda 52,21 (buton); 56,20; 98,3; 102,21; 120,21; 134,4; 170,28 ; 374,25; 396,24; 416,17. dohte gif he hi ealle ofsloge, J)8et se an ne setburste fe he sohte. Horn. I, 82,12. Hom. Vol. I, 16,35; 62,2; 82,18; 108,7,8; 124,26; 178,7; 226,1. Vol. II: 178,23; 246,25; 252,7; 304,19; 478,28; 538,7; 576,1. Indirect questions, ascedon, gif he feohtan ne dorsten, h wider hie fleon wolden. Or. 54,1. Hom. 11, 242,16. Cases of virtual indirect discourse with no verb of say- ing expressed are found. Gif {)onne hywlc wif to dam unspedig wajre psei heo das ding begytan ne mihte, J)onne sceolde heo bringau twegen culfran briddas, odde twa turtlan. Hom. I, 140,2; also 134,16; 196,10. Vol. II, 58,33. 68 b) A few instances of pret. indies, are found. J)yder be cwicd piot man niiht geseglian on anum monde, gyf man on nilit wicode etc. Or. (Othere) 19,12. bodode, J)£et him wses Godes graina onsigende, gif hi to Gode bugan noldon. Horn. I, 246,16. An instance probably unique in which the verb of the protasis is pret. indie, while that of the apodosis is pret. subj. This difference of moods common in the pres. was avoided in the pret. ic dohte J)8et hit wsere la3ssa sedryt to gehyrenue, gif man da ane boc rset on anes geares ymbryne, and da odre on dam seftran geare. Horn. II, 2,10. In the following instance the construction changes from pret. subj. to pret. indie, in the clauses more remote from the governing verb, which retain the form of direct discourse. It is a good example of the tendency to isolate all clauses not immediately dependent upon the verb. Swa swa mon on eald spellum segf), J)8et an nsedre wsere de hsefde nigan heafdu 7 simle, gif mon anra hwilc ofsloh, fonne weoxon jjses siofon of psem anum heafde. Cons. cp. 39, §4, p. 216 m. Some what similarly. Beda 268,15. where the second condition has the pret. indie, (cf. Section IV final note). Pres. subjs. depend upon a verb in the pret. de we ser bufan cwsedon, dset he donne to foo, gif he niede sciele. C. P. 59,9. Mat. 22,24. The verb of the apodosis is pres. subj., that of the pro- tasis pres. indie. Moyses wrat, gif hwass brodor dead bid 7 Isefd his wif 7 naefd nan beam, f)a3t his brodor nime his wif etc. Mk. 12,19. A version of the same sentence in which both subj. and indie, appear in the compound protasis, wrat, gif hwass brodor byd dead 7 wif htebbe 7 se byd butan bear- num, peet his brodor nime his wif etc. Luke 20,28. 69 The apodosis is dependent upon verbs of commanding, promising, wishing, fearing etc. Fret. subs, in both clauses. This construction is nearly allied to indirect discourse. foif)an pe hie ondredon, gif hie hwiiiim ne wunnen, {)£et hie to rafe aslawoden 7 seeargodon. Or. 212,18. bebead gif hwa hit forsoce, psdt he sceolde beou forbaerned on hatum ofne. Horn. II, 18,25, Or. 266 10. C. P. 65,1. Horn. I, 124,5. Vol. II, 40,13. gehet J)£et he wolde Cristes deowdom geceosan, gif he him lif 7 sige forgefe on faem gewinne, etc. Beda 122,34. Beda 126,10; 234,31; 242,33; 306,24; 308,19; 328,19; 332,11 ; 344,26; 380,1; 390,19. An infinitive following a verb of commanding represents an apodosis, the verb of the apodosis is pret. subj. het him ceowan (512) mid scearpum todum his lidegan flngras, gif him alyfed wsere. Horn. II, 512,1. The apodosis is a subject clause. Pret. subj. in both clauses, wses cweden ftette, gif heo tram wifum clsene waeren, fset heo moston onfon 7 ficgan fa foresetenesse hlafas. Beda 84,14. C. P. 93,8; 459,22. Hit wtes swa gewunelic on ealdum dagum J)3et gif hwcera sum fserlic sar become, faet he his reaf totfere. Hom. II, 454,13. Vol. I, 6,34; 40,34; Vol. II, 100,32; 166,32. The following instance of a supplementary clause as a protasis may be a inserted. Pret. subjs. are found In both clauses. 7 geweard da sejiatos him betweonum, gif hie raou I)riddan 8i{>e oferwuune, J)8et mon ealle (Cartania) towurpe. Or. 210,5. The apodosis is a clause of purpose. swa J)£et he mehte EEg{)erne gersecan, gif hie aenigne feld secan wolden. Ohron. p. 90,10. Chroii. E 218,11. sealde J)8et betste hors |)3em biscope dast he hwsedre 70 on J)8em meahte fordas ofeiridan, o(lf)e glf hwylc ned gelumpe, |)8et he hi'sedlicor feran seolde. Beda 196,6. So. Beda 122,3,11; 264,32. G. P. 397,19. Horn. I, 80,7. An action in past time may have as its end a continuous purpose, which is properly expressed by the present tense following a preterite. An instance with the subj. in apod, and the indie, in prot. is. Se Godes Sunn wtes on his gesthuse genyrwed, jjset he us rume wununge on heofonan rice forgife, gif we his wille gehyrsumiad;. Horn. I, 36,1. The apodosis is a consecutive clause or a clause of result. He teohhode gif hw£et get-yngoden (-en from Gott ms) on J)8em frydorae, piet hi hit eft on dam freodome mid hreowsunge gebeton, 7 gif hiora hwilc swa heardheort wsere f)3et he nane hreowsunge ne dyde, f)8et he J)ouue hsefde rih-tlic wite. Gons. 41, § 3, p. 248m. Beda 228,18; 274,29; 316,21; 328,34. Horn. 1, 406,21. With the pret. indie, in apodosis pret. subj. in protasis. to don f)sette Ecgfrid se cyning him geheht ge lond ge micel feoh to gesyllenne, gif he da cwene gesponan 7 gelseran meahte, fset heo brucan his gesynscypes. Beda 316,21. VIII. CLAUSE ORDER AND WORD ORDER. The~ elaborate study of word position in Anglo-Saxon which Ml'. C. A. Smith of tiiis University has nearly pre- pared for publication, makes it superfluous for me to treat this subject otherwise than in broad outline. Logically and in the order of thought the condition pre- cedes the conditioned. While this is the order of the mem- bers of the conditional sentence in the majority of cases, Anglo-Saxon, like every flexible language, is free to adopt other arrangements of the clauses. The possible relations of the subordinate to the indejien- dent clause, as concerns position, are three. The criiiditional clause precedes. This is the most usual, arrangement. Git ic monnum cueme & licige, donne ne beo ic no Godes deow. G. P. 147, '20. Drihteu, gif f)U wilt, {)U raiht me geclansian. Hom. I, 120,11. The great majority of conds of class II with a command in the apod, have this order. The conditional clause follows. Only less common than the previous case, ac wit on niht sculon secge ofersittan, gif he gesecean, dear wig ofer Wffipen. Beo 683. I)is bid ure rice, gif we hit uu 72 geear-niad. Horn. I, 264,5.' The buton clause usually follows its apod. The conditional clause may stand between different mem- bers of the apodosis. Swa eac treowa, gif hi beodon fuUum monan geheawene, hi beod heardran. Horn. I, 102,22. swa fonne seo saul, gif heo ne bid mid Godes worde feded, gastlice hungre & J)urste he bid cwelraed. Blick 57,11. Ne nan man ne mot, gif he Cristes miltse habban wile, his aslmessan behatan to wylle ne to wydetreowe. Wulf. 303,16. (deofol) wile oferstaelan J)one engel, gifhemseg, mid {)8em yfelum weorcum. Wulf 233,9. f)y he na mseig na wandian, gif he riht def, na|)or ne for heauum ne for ricum. Wulf. 275,20. cf. 305,21. be fjgem f)U miht sweotole ongitan, gif J)u tenigne mon gesihst willnian J)£es he nseff), feet J)am bij) anweald wana. Cons. cp. 36, § 3, 176 h. This interposition of the prot, is rare except with paren- thetical, conditional formulas like gif he mseg cited above, or when a conjunction properly belonging to the apodosis is separated from it by the protasis (notably fordsem). ORDER WITHIN THE PROTASIS. The protasis introduced by gif has frequently the direct order. Subject — copula — complement, when it precedes the apodosis. Gif donne segder bid unwis, donne is to ge dencanne hwEet Crisd self cuEed on his Godspelle. C. P. 29,6. Gif we willaf) areccan ealle da gewitnysse \>e be Criste awritene sind, J)onne gsed fser swide micel hwil to. Horn. II, 18,10. Gif f)U sie Godes Sunu, cwef> pset pa stanas to hlafum geweorfan. Blick 27,5. 73 When the complement is formed by an infinitive with an object or modifiers, the infinitive is often put at the end of the sentence, and its object or modifiers stand between the copula and the complementary infinitive, Gif se sacerd ne maeg dam laewedum mannum larspel secgan, huru he sceal {)urh his lifes unscseddignysse, him wel bysuian. Horn. II, 532,16. And gif ge nellad f)issum gewritum gelefan, ic sende ofer eow wyrmas and fuhias wedende. Wulf. 231,11. The direct order is not excluded but extremely rare v.'hen the protasis follows the apodosis. Ac we sceolon awendan urne lee fram yfelre gesihde, urne hlyst fram yfelre sjjrsece — — gif we willad becuman to dam estum fees ecan gereordes. Hom. II. 374,3. We beod sodlice gesceode, gif we geefenlsecad mid urum faerelde fordfarena manna lif. Hom. II, 280,30. ForJ)am aelc craeft 7 selc auweald bif) sona forealdod 7 forswugod, gif he bi{) butan wisdome. Cons. cp. 17,60 m. The Transposed order, subj. — complement — copula, is common when the protasis precedes and all but universal when the protasis follows the apodosis. The protasis precedes the apodosis. Gif mon hine (jet) on fyr dej), |)onne fleoj) f)ser neddran onweg. Beda 26,16. Gif ge fset secan nyllaf), |)onne magon ge J)£er eardungstowe hsebban. Beda 28,14. Gif we hit forsuwian dorston, ne ssede we hit eow. Hom. II, 324,23. gif f)U Strang sy, syle wcestmas Jbinre eordan. Wulf. 260,6. Gif hwa dearfan forsihd, hn t£eld his Scyppend. Horn. II, 328,17. The transposed is the usual order for an unfulfilled protasis. 74 Ac f)a3r hi senif^e wuht agnes odde gecyndelices godes on heora hsefdon, f)onne hsefden hi f)set mid him, f)eah he f)3et rice forleten. Cons. cp. 27, § 3, lOOh. The protasis follows the apodosis. ne bif) se cwuca donne uyttra f)e se deada, gif him his yfel ne hreowj). Cons. cp. 36, § 6, 182 m. Swa ded min se heofonlice Feeder, gif ge of eowrum heoilum eowrum brodrum ne forgyfad. Mat, 18,32. Ge beod mine frynd , gif ge da J^ing dod J)e ic eow bebeode. Horn. II, 522,28. ne hrepa du. J)«s treowes waestm, for- J)an de |)U bist deadlic, gif fii f)ses treowes waestm geetst. Horn. I, 14,1. Relative and Temporal conds. usually have the trans- posed order, though the dii'ect order is not excluded. The protasis introduced by bnton (nemne, nymde) is found both in the direct and in the transposed order. I am unable 1o state the principles which lies behind these changes except so far as they are covered by the rule given below. Two sentences like these from the same page show that the Older is not a rigid one. Ac hwsedere nan man ne cymd to Godes rice, buton he sy afandod. Hom. I, 268,8. compare ac hwsedere se man naefd na micele gedincde, buton he afandod sy. Hom. I, 268,18. Direct order. 7 him ne forgifed fsette he mote in Godes bus gongau, nemne he sy wsetre adwegen. Beda 84,25. fordan de nan man ne mseg habban Godes rice, buton he haebbe das dreo ding. Hom. I, 250,13. buton hwa beo ge-edcenned of wsetere 7 of haligum gaste, ne niaeg he infaran on godes rice. Jn. 3,5. Transposed order. 7,5 Se de to me cymd, lie mreg he beon min leorniiig-ciiiht butou he his wif hatige. Hum. 1, 'M)y,,'.\[. ne bij^ hit no dy hra|)or swa, deah hi his wiinigen, buton hi [la iif ealle hab- ban. Cons. ep. 38, § 2, 1241. Butou Drihten da burh ge- healde, on ydel vvaciad pa, hyi'das de hi healdad. Hom. II, 230,4. For rhetorical emphasis the object may be thrown for- ward the clause remaining otherwise unchanged. J)isses ic me ondrsede swyde to gebidanne, buton me godes gife gemiltsige. Wulf. 248,13. la liw^t is se man on life, butoD hine god aelmihtig gehealde — ? Wulf. 101,19. Inversion: copula — subject — complement has already been noted as a method of forming a protasis without a con- ditional conjunction. Two citations will suffice to recall the case. Nseron hi gegaderode 7 geradode , swa wiperwearda gesceafta, donne ne wurdon he iiaefre ne geworhte ne ear gegaderode, and gif he hi (158) ne bunde mid hi.s unabidend- licuni racentum, ctonne toslupan hi ealle, etc. Cons. cp. 35, § 2, 1561. The first condition is a not particulary happy variation of; Mitndus hie eci' tarn (Uitersls covirariiscixe ptnii- hus in imam forniam mlmine conuenisset, utsi uiuis cssct qui tarn dluersa conuDujcret. Bade mihte f)es cwyde beon Itewe- dum mannum bediglod, ntere se gsestlice getacnung. Hom. I, 94,33. Through the kindness of Mr. C. A. Smith I am able to state a principle which lies behind and explains many of these changes from the direct sentence order. It is that the Anglo-Saxon has a constant tendency to place a pronominal object before its verb. Consequently all clauses with pro= nominal objects tend to fall out of the direct order, and most of them into the transposed order. 76 ORDER WITHIN THE APODOSIS. The apodosis has the inverted order copula — subject — complement when ne or {)onne stand — at the head of the clause, when the verb is imperative or subj. (adhortative), or when the complement is thrown forward for rhetorical emphasis. The most frequent cases are : With J)onne as a correlative, far the most frequent form. Ac gif he Jonne daere styringe ne widstent, donne gescient he da godan weorc de he oft ser on stillum mod durhteah. C. P. 215,15. donue bid he idel, gif he on dsem sumra bidt ingonges in heofonrice. C. P. 285,15. Gif du das word mid weoicum gefylst, donne do ic swa du me tihst. Hom. I, 422,16. Gif du f)is nelt, donne scealt du — — mislice wita dro- wian. Hom. 1, 544,3. Ne stands at the head of the clause. Ne cwtede he no sua, gif he no on geate dset him wses dees wana, etc. C. P. 311,19. ne cymj) he nsefre to oferrura? ende, buton he ha3bbe swa scearp and get swa J)8et fyr. Cons. cp. 39, § 4, p. 216 e. Gif hi forseod Moyses se and dsere witegena bodunga, nellad hi gelyfan, |)eah hwa of deade arise. Hom. I, 334,20. The frequent use of correlative J)onne, and of ne at the head of the apodosis of unfulfilled and buton conditions, makes the inverted apodosis by far the most common. The direct order is of course frequently admitted, the transposed seldom unless the apodosis is itself a subordinate clause. REMABKS. Imperative and adhortative clauses with the subject expressed usually have the inverted order. Gif he Godes man sy, fylgadge him Beda 100,23. The direct order is also found in mandatory clauses. It 77 is very rare except in Aelfric. Gif hwa synful sy, he andette, and iiajlass herige. Horn. I, 448,4; also 124,34; 584,1. |>u leofa bntau me, gif {>u mcege. Wulf 259,5. This is simply another version of: leofa, gif itu maga, buton me. Horn. II, 104,6. 2. Two examples which show very well that inversion depends upon the presence of adverbial elements at the head of the clause are: Ac gif he gelyfdon J)one Hselend, fonne sitt he wid fone weig. Gif he nele biddan f);es ecan leohtes, he sit donne blind be dam we- ge unbiddende. Horn. I, 156,2. These sentences exemplify very well the conditions under which J)onne produces the inverted or admits the direct order. So with both ne and donne: Gif he abryd on dsere eht- nysse, hi ne bid donne geleafa, ac bid hiwung. Hom. I, 250,20. 3. A rhetorical throwing forward of a portion of the complement causes inversion of the rest of the clause. Gif se Halga Gast ne Iserd JjiL's mannes mod wid innan, on idel beod |)8es bydeles word widutan geclypode. Hom. I, 320,25. The same rhetorical tendency is shown by the following example, though the adhortative clause would probably in any event have the inverted order, and ealswa Niniuete, gif man J)8et gerfede, gold and glencga alecge man f)a hwile. Wulf 170,7. The rule is not rigid as is shown liy the following example of a infrequent word order. Gif du Gode sode da^dbote geoifrast, his sodan miltsunge {)u begytst. Hom. II, 418,11. IX. PECULIARITIES OF THE DIFFERENT DOCUMENTS AND CONDITIONAL CONJUNCTIONS. It is obvious that the conditional construction perhaps more than any other is affected with regard to its frequency and variety, by the nature of a document oi' the chai-acter of the Latin original upon which it is based. Thus we are prepared to find that Orosius shows veiy few conditions since it is a sti'aight-forwai'd paraphrase of a historical narrative. So we -find it natui'al that the annalistic style of the.Chronicle should use few conditions, and these mostly in the historical or preterite tense. The abundance of anecdote in Beda is sufficient cause for the prevalence of indirect discourse in that text. On the other hand the homiletic style with its continual arguments ad hominen, and warnings and illustra- tions shows comparatively little indirect discourse and an abundance of conditions most of them in present time. The Alfredian Boethius and the Homilies of Aelfric are unequalled in the richness and diversity of form whicli they give to this construction. The 'ideal' condition shows a much wider scope ; and the clause and word order is somewhat freer than in the other texts. There is certainly more feeling for style in these texts than in the Cura 79 Pastoralis which is a simple, idiuniatic rendering, and the Beda which is a some what slavish translation, of its original. With respect to this construction the Gospels follow the Latin more slavishly than any other text. They follow the moods of the original and even the word order as closely as possible. The feeling of reverence for the letter of the scrip- tnies which affects all translations from them is perfeclly natural and evident. A single instance will illustrate the difference between ihe Gospels and other texts The follow- ing passage has the indie, following the Latin, three times in the Gospels: and gyf hwa to inc hwst cwyd (dixcnt), secgad |)aet diihten hsefd his neode 7 he hine sona hider laet Mk. 11,3; cwyd (dlocerit) mt. 21,3; ahsad (inlerrogancrit) Lk. 19,31. An independent rendering shows the subj. in -protasis, the proper form with a numdatory apodosis: & gif inc (71) hwa J)fus \vif)cwe|)e, secgaj) J)a3t Drihten f)ieo ah J)earfe, rafe hie mon f)onne forltetej) to me. Blick 71,1. The Aelfiedian texts, poetry (though the material is small) and the Blickling Homilies seldom admit an indie, in the protasis of Class II, Aelfric, quite exceptionally lor a text presumably fiee from Latinizing tendency, uses many indies, in this case. The miscellaneous collection ot VVulf- stan's Homilies, shows again almost exclusively subjs. Rushworth Matthew shows a veiy interesting difference from all the other texts in the fact that conditional f)a3r and nu, which are only sporadic elsewhere, are regular conditional conjunctions in it. nu uuJ)oune 6,30; nu 10,25, nunu 22,45. I)ier 11,19; 12,7; 23,30; 24,22; 24,24; 24,43. The theory suggests itself that these may have been forms rather colloquial than literary. The fact that they 80 creep in a few times only in each document, and their use by the undoubtedly provincial scribe of Rushworth Matthew, give a little support to this view. nymde, nemne for buton. In another connection 1 may call attention to the dialec- tical significance of nemne, nymde. And may suggest that they are certainly Anglian and probably Mercian. Alfred, excluding Beda, Aelfric and Wulfstan do not use the form. It is represented in West-Saxon only by three instances in the Blickling Homilies and three in glossaries, and a few in the Paris Psalter. Beda gives nemne the preference over buton. This and the fact that Beda prefers swa swa for swylce in the sense of 'as if are only a foretaste of the many syntactical differ- ences that Dr. Miller will be abbe to adduce against the Alfredian authorship of Beda. The Vespasian Psalter- and Hymns and Rushworth Matthew prefer nemne and nymdie. The forms are also found in the Lindesfarne Gaspels, under the influence of the Eushworth gloss. Nemne, nymde is found beside buton throughout the whole body of the poetry, though not in the songs from the Chronicle and the Battle of Maldon which are undoubledly West-Saxon. Beowulf and the Caedmon cycle show it abundantly; it is found in the Riddles, and in the poems of the Cynewulf cycle. It is hard to see any considerable progress or develop- ment of the conditional construction during Anglo-Saxon times. Aelfric and Wulfstan show certainly an advance over the Pastoral Care and the Blickling homilies ; but Boethius shows already all the forms and varieties of the later texts. There was undoubtedly some advance toward the use of 81 auxiliaries and compound tense.s, but no very marked one. The extension of tbe-st ending to the piet, opt 2 ud sing, in weak verbs and the consequent levelling out of the distinc- tion between moods of the pret. of weak verbs is the only change of note that I have been able to discover. CONDITIONAL OON.IDNCTIONS AND CORRELATIVES. This matter will have been so thoroughly exhibited in the examples, that a brief and dogmatic recapitulati(jn will suffice at this point. For exceptional forms the reader is referred to section V. The regular conditional conjunction is gif The form gief is rarely found ; in Late West-Saxon gyf is common. Gef is occasionally instanced, especially in Wulfstan (Napier) homily 44, a text which seems to show a transitional voca- lism with strong Kentish features. J)8er is established as an occasional conjunction for the unfulfilled condition, and Rushworth Matthew makes nu a regular conditional conjunction. Buton (butan, in LW.S. bute, North, buta) also nefne, ueiTine (nymne) and nymde (nemde) intioduce the protasis of Class V. {)onne appears probably in rather over half the apodoses as a correlative to gif or buton. It is frequently found in both protasis and apodosis, in which case it could have had little more than the force of a particle, with the slightest tempoi'al significance. STATISTICAL APPENDIX. The purpose of this portion of the work is to make sufficient material accessible, to check readily the statements made in the text, or to command the field for purposes of further study. It is no aim of the writer to present all the material which he has collected. In view of the inevitable suspicion which must fall upon selected statistics, the writer has decided to give the complete material for the Pastoral Care, the Orosius, and the Homilies of Aelfric, thus giving control of the most important monuments of Early, and of Late West-Saxon respectively. The material from other texts, complete in each instance, is offered where it has particular value or interest. It seemed necessary only with Classes 11 and III to support the classi- fication by examples so for as possible from the whole Anglo- Saxon field. Statistics which have been made complete in the text are not repeated here. Such are: the pres. subs, not in- cluded under classes II and V which have been treated in Section IV appendix to class III ; and conditions in indirect discourse etc. for which the material from the Pastoral Care, Orosius, and the Homilies of Aelfric has been made com- plete in Section VII, also most of the matters treated in small type. In general matters rnei-ely accessory to the construction and treated somewhat minvitely in the Sections from V in- clusive have not been repeated. It is hoped that the large body of material from Ael- fric's Homilies, by all means the most important monument of Anglo-Saxon syntax and prose style, will be particularly acceptable. CLASS I. SIMPLE CONDITION. A. WITH THE PRES. INMC. IN BOTH CLAUSES. Pastoral Care. & ge don, swae we swide eade magon, mid Godes ful- turne, gif we da stilnesse habbad, daet[te] eall sio giogud — sien to liornunga odfreste. 79, 296,?, ;32:!r,, 33i, 351!-, 45io,ia,ao, 47io,i.!, 49i, 53i3, 633, (iOu, 20,22, 7] 8,15.19, 79m,«, 8l8, 83i;i. 93io,i«, 977,., III4, llSe, 123i5,i7, 1252o, 1297, 133i, 137.^.1:!, 147-.0, l57i5, lH7i3, I7I25, 1933, 1979, 201,4. 211i4,2o, 215i.-,. 2I7ii. 22O1, 222i.i2, 22522, 22922, 231i, 233io, 265,. 2713. 273,7, 277h, 28]j, 28323, 285i5, 295i2, 303,, 30720, 3277, 333,3,...,, 33.o,i, 34l9,u,i7, 351i4, 361io, 36724, 377i7, 38331, 385.2. 397o. 4072r,,2«, 411i4. 42l2,5, 423:iL.,35. 425n,s, 433,,27. 4372,, 44l9,i7, 445s,2i,2s. 447io,i9, 449,2, 451 24, 4592:-,, 4695. Orosius: 7 gyf {»ar man an ban flnded unforbserued, he hit sceolan miclum gebetan 2I12, 64i4, 12O7, I563i, 214?. Homilies of Aelfric. Vol. I. Gehwa mseg {)e eadelicor da toweardan costunge acuman, durh Godes fultum, gif he bid durh bodice lare getrymned. 43,7, 624,27, 14i.i2, I614, I82, 5227,29,33,31. 544,24, 563,17, 7028, 72l7,31, 86:13, 965,7,26,33, 98l, 10222, IIO3, 11422,23, 6* 84 1185,11,18, 120iT,2... 1242, 138ri, 1402i,,;)3, 14L>is,i5, 148i-2, 1563,M,«, IfiOii, 16231, 164», Ifi6-J4, 16821.. 170i5, I7620, 178l>i, 2047,i»,l';), 210:i5, 2l2i7, 222... 224i.'. 232;',;;,:i.-,. 234i2, 238i7, 240i..,i:i, 24831, 250i 1,211. 2522s, 254:., 256ir,.-j2. 2602,7, 2645, 2(i63,2c,:ii, 26819, 27232,34, 274k. 276?. 2862-1.21;. 306i,9,io, 31233, 3202:.. 3342... 33823, 340i7, 3509, 372i,-.. 378i2. 390io, 4O818, 4103,-28, 4123,6,26,34. 4l6]s. 420^,32, 422i3,io, 42828, 43826. 45639, 46022,34, 464v,. 46821. 4722.6,16. 48O22, 488,32, 494s, 49622,30, 498iH,24, 0O625, 5282U, 55625,27, 56223, 57026, 5743i.32,3.i. 576io, 586io,i9, 58835, .5902,6,26, 5922s. 5943,i8,2i, 598,3. Vol. 11: 128,13, I810, 2227, 28,1., 30i(.. SOs, 6O1, 623o, 6427, 6625, 74,13, 76lO, 1025,15,18,21,30, 1047. IO617, 13820, 170c, 174i3, 186e, 2OO4, 2086,10.31, 21027,2:.. 2I620, 2189,16,27, 2229, 2248, 2283, 230i7, 234i2, 24O29, 266i, 2706,9 278,3, 2802i,3o, 28222, 3O817, 314ii-.2s,2:., 3167,16,33, 3I813, 320i4, 322ii, 324i,5, 32816,17,27, 33O21, 3544, 356i2.i4, 362i.33. 3743. 3903i, 39225. 3942,10, 396i,27, 40626, 4O87, 410»,24,27, 418ii, 4304,6,i2, 4362i, 44225, 444i7,3o, 46430, 4666, 4?0i, 516,2.;. 522s,2s, 5283,ii.i9. 5323,11,13,16, 534:3,13,23,26, 536io, 54435, 55420, 56229, 5646, 574ii. 576,25, 58224, 58828,30, 59029, 592io,2i, 602ii,i6,i9. Apodosis is a question or exclamation. > Pastoral Care. 5924, 63i,5, 9124, 1.333, 281i3, 3338,io, 3414, .37116, 377:5,8, 405ii, 421i9. Homilies of Aelfric, Vol. 1: 26i7, 6823, 2507, s. 2522i. 2565, 302,:34, 3962. Vol. II: 2265,i3. 26821, 3285. 368n. 40S:w. 42O13, 4.3O17, 4322,14, 46225, 536i3, 582ii. B. PRET. INDIC. IN PROTASIS, PRET. INDIC. OR PRES. INDIC. IN APODOSIS. Pastoral Care. Pret. indie, in both clauses: 53s. 253?. 3893i. Verb of the apod, is pn^s. iudie. : Gif Crist for us eallum dead was, donne weordad ealle men rleade. 4?):i, 57i.i,l':), 217?, 40134. Beda. Pret. iiidic. in both clauses. Iij2t,u'. Pres. indie, in apod. 2ii, 8Sii.,2i. 102i8) Homilies of Aelfric. Pret. indie, in both clauses. Vol. I: gif he hwon hnappoile, dterrihte hine drehton nihtlice gedwiraor. 8617. 124.s. lHh::. 3o2i,i., 404:i, 534^, oGOio. Vol. II: 130:!, 1407,i7, IBOis. 242i«, 392io, 514i4. Pres. indie, in apod. Vol. I: fAn-., 112iii, 2()82!i, 442.;. iK^G>. Vol. iL: 8O1S, 2842fi, 41834. 4r)2.r2. 47(i2i. CLASS rr. THE CONDITION WITH A MANDATORY APODdSIS. THE VERB OF THU PROTASIS IS PRES. SUBJ. Pastoral Care. 43-1,. 2.1.1.1.-.. 127i (es/). 13!.;. l'>%. 197-,. 3234 (^minislrat). 34722 (hahetis), 349:i (offers), '.)^->7j ic.rpi-lUhir), 4 [riheilit]. 37 li.-. (Joqiiifiir) 383i (es/), 437'i (dispieiuni). Avu->\; (ris). 402 (is cynn), 199i:j (sceal), 22 (is dear!) 3()722 (sculon). Orosius. gif his hwa sie lustful! mare to witanne, sece him ponne self f)set IOO2.-,. 214io,i.3. Boethius. 14h(««s), \%\\ (iKrlwrrescis), 30 h, 1201, 1981, 2321 riR. Beda. 36i(; (est), i-j. (vis), 7224 (confint/nf), 26 (est), 1002a (est), 33, 1022, 278is. 2803, 400>i (vultis). Inflected infinitive: is cynn 452, 72.-., 74i. Sceal: SGir.. Weorde is 1367. ie biddo: 61, 2902i.2;i. Beowulf: 452, sec, gif J)U dyrre 1379. 1480. Genesis: 23! i, 2498, 2655, 2825. Satan: CSO. Andreas 417, 557. Elene: 540. 620, 722 (apod. 783); 856. Riddles: 33i3, 402». 86 Oldest English Texts (Charters) : No. 3411. 11,12,1., No. 38117,9, No. 4111.14,24,40, No. 4511. 12,15,16,25,26. The apodosis is a com- mand less vividly expressed, is min willa 414,?, we bebeodad 48i7, ann ic (I grant) 4I35, 4545. Gospels. Matthew : 43, 529,20,39, IG24, 2639,42, 2740,42,43, 26g3 (ic halsige). Mark 323, lie. Luke 43,9 2267, 2335,37. John 1226. Blickling Homilies. 275,io, 2925, 439, 71i, 109i8, 153i9, 177i, 17925, I8I14, 183i6,i7, 22525, 233i8,34, 24l2o, 243i8. Apod, with sceal etc. : 47 25,30, 492, 20523,a5. Homilies of Aelfric. Vol. I: ac gif he wille f)set him God milde s^, |)onne hlyste h6 g6des rsedes. 54i7, 56i7, 96i, 13227, 1484, 16612,17, I6821, ]70i,i8, I8O10, 226a, 24222,26820, 372i3, 37625, 4I623, 4243, 4484, 450i8, 460u, 46422, 4807, 5O623, 5I64, 584i, 59O19. Vol. II: 4834, 1042,6, 174i7, 184ii, 2I624, 2565,io, 264?, 280s, 290.30,31, 38828, 39228, 488i8, 594i7. Apod, with sceal etc. Vol. I: 89 (bidde' ic), 16621; 4822G (bid us I'sedlicor) Aelfric's Translation of the Hepta- teuch and of Job. (from Wohlfabrt) Gen. ISs, I83, 19i2, 32i7, 47,6. Ex. 124,10, 2l3,4, 224,6,7,9,12,26, 234,.'i, 2934. Lbv. lio, 37,12, 2O2, 24i9, 2526. Num. 3io, 13i. Dent. 1424, 15?, 21i8, 2222, 2324,25. Since the homilies ascribed to Wulfstan are to a considerable extent a repetition of texts already treated, it has not seemed necessary to offer statistics from them. Suffice it to say that like the early texts they show almost without exception the subj. in protasis in this case. The indie, in protasis (apodosis a command). 87 Pastoral Care. lOSo (dubUant), 199:,, 4032. 407is (is cyn). Beda 6425, 1304. Boethius cp. 7, p. 26 h; cp. 13, p. 38 b; cp. 33, § 1, p. 1201. Chron. E 3323. Gospels. Matthew: 4e, 623, Ssi, 1428, 17*, 18s,:i.i:-„i(i,i7, 197,21 21.i. 2423,26. Mark.: 834, 97,^2,35,1:1,15, 113,25, 1321. Luke: 92:1. lu.i, 159, 173,4 1931, 22i2. 233!.. John.: 74, 83:1, 10:i7,:i«. 14i5. 157, 188,23. Ill practically all of these instances the verb of the Latin condition is indie. Homilies of Aelfric. Vol.1: 54w. 62?, 1243i, 420.;, 464i3, 512.SO, .55(11. Vol. II: 8829,30, 104i, 246l'i. 27(117, 294:i, 3I812, 3282.1. 39234, 4006, 414i3, 4843i, 486.;, .o34s,ii, .fi92:.(i,34. Apod, sceal etc. Vol. I: 140i3, I(i2j7, 174:)3, 274io,rj, 5169,14, 5545,7. Vol. II: 221, 202s, 544.27. Aelfric's Heptateuch and Job. Introduction 143o. Ex. 21 22, 227,s. Lev. I14, 422, 7i7,25, 209. 2520. Num. 14i3. Deut. 19is, 253. Jud. 159. Job 82. CLASS m. IDEAL CdNOITlON. Boethius. Cp. 13, p. 38 m; cp. 14, § 3, p. 461; cp. KJ, § 2, p. 52 h; cp. 27, § 2, p. 961 (2 conds); cp. 32, § 2, p. 116m; cp. 33, § 1, p. 1201; cp. 34. §4, p. 138 m; cp. 3r., § 4, p. 1601; cp. 36, § 2. p. 174 h; cp. 38, ij 6, p. 208 ra: cp. 41, i; 2. p. 2441. Inverted protasis, cp. 27, S 3, p. 981. Ca3d. Gen. 407, 831. Andreas 478. John 21,25 (scribantur) Blick 247 1. Homilies of Aelfric, Vol. I: 4O27, 1125, 21225, Vol.11: 88x7, 454^2. Job. 126. Introduction to Old and New Test. 15,42. 88 CLASS IV. UNFULFILLED CONDITION. Pastoral Care. ISTa, 2II21, 2352, 25oi.,. 2H.5:j. 3032, 3092, 311i9,23, 3133, 3318, 359i, 385.M, 39132, 42729, 435i,3. Orosius. 1043, I8623, 2I85. Homilies of Aelfric. Vol. I: 4a, I2.12, I823,. 262c, 52i, 8228, 9433, 1247,18, 130,5, 14226, 146i4, 1507,, 164i6, ITOai, 17420, 184i9 (buton), 214:i2, 2208,i8, 224i9, 226io, 23627, 24884, 276i8, 292i, .320i, 32824, 3309, 33234 (buton), 33623, 40427,33, 4087, 48428,29, 5781,11,12, 59O4. Vol. II: 627, 4O29 (buton), 8O26, IO64, 144i3, 19223 15622, 234io, 248;;, 3OO27, 3202, 32423, 3442B, 358i.5, 3804, 412io, 5747. CLASS V. THE CONDITION WITH BUTON. A. PURELY EXCEPTIVE. Pastoral Care. 43^, 1696, 1917, 222i4, 349i6, 445i4. Homilies of Aelfric. Vol. I: 812, 146, 1645, 26O24, 464i, 590ii. Vol. H: 222, 22424, 2304, 3247, 3362i, 34023, 41634, 42221, 46222, 52828, 59223. B. INDISPENSABLE CONDITION. Pastoral Care. 157i7, 163i9, 207i, 2173, 25 lu, 345i9, 44I7, 445io. Orosius. 58x3. Homilies of Aelfric. Vol. I: 144, I623, 2629, 94io,3o, 122i9, 12832, 14433, 14831, 170i2,i4, 1866,10, 222i2, 250i3, 25420, 26612, 2688,18, 280,4, 30831,35, 41232, 46222, 47429, 512,i9, 598i4. Vol. II: 1032,35, 122, 74i5, IOO27, 2IO26, 222i3, 26625, 314i7, 31826, 32O20, 32221, 33622,29, 340i2, 4U2=. 4329, 47022, 582i2, 59228, 6028, 6046,i2. u. ■M^ F*^-- /; y.:'ii¥i m. it' ['%. %*ii l^j^\ ''* :t -rc^ .fu ■..>m i^i^-ml >'- •