\ \ PRINTED FROM THE HARVARD STUDIES IN CLASSICAL PHILOLOGY, VOL. XVIII, 1907 STYLISTIC TESTS AND THE CHRONOLOGY OF THE Mtf'i^ WORKS OF BOfi:THIUS v By Arthur Patch McKinlay WHOEVER undertakes to treat of Boethius finds himself in illustrious company. Potentates, churchmen, scholastics, and philosophers have busied themselves with this " last of the Romans." It would appear that but little remains to be said on such a well worn Bubject. Much less does it seem fitting in a beginner to essay that little. Yet, as the recent researches of Usener and Brandt and the acute suggestions of Rand have marked an epoch in Boethiana, one may hope to gain still further insight into the character and mode of thought of the author of the Consolatio. With this purpose in view, by the help of the so-called stylistic method, I intend to examine the writings of Boethius, in case it may be possible more accurately to place works the dates of which are not yet certain. To be explicit, I hope to show that the De Arithmetica and the De Miisica should be placed neither first nor together ; more definitely to place certain other works ; to throw light on the authenticity of the De Geometria and the De Fide Catholica, and incidentally to test the value of the so-called stylistic method in determining the relative chronology of an author's writings. For a definition of the meaning of stylistic method, and an illus- tration of its application, I may refer to the well-known work of Lutoslawski, entitled The origin and gi-owth of Plato'' s logic tenth an account of Plato's style and of the chronology of his wntings, 1897. Lutoslawski applies five hundred tests, comprising more than fifty-eight thousand instances. The tests are of various sorts, such as the relative frequency of hiatus and of synonymous expressions in works of different periods. Lutoslawski prefaces his conclusions with certain principles which he contends must hold good in all such investigations. As these principles practically coincide with my own, evolved independently, — for I purposely did not read the book till my investigation was finished — I cite the most important of them. 2G5612 124 Arthur Pate Ji McKinlay 1. The method is of Httle value in works of but few pages in length. 2. The method is of little value unless corroborated by considera- tions other than stylistic. 3. Synonyms are the best tests. 4. Of two works, the one which agrees in more criteria with a third work whose date is fixed, more nearly coincides with that work in time. Relying on these principles and tabulating his criteria, Lutoslawski shows that known early works of Plato, such as the Crito, have but few points in common with the last, the Laios, whereas the Sophistes, Politi- cus, and Philebiis agree with the Laivs in more than fifty per cent, of the tests used. Hence he concludes that the dialectical works come late in Plato's career. This much for the method. Anybody who has read Boethius with care will have observed two marked characteristics of his style. The first of these is the prevalence of various constructions, due, apparently, to the influence of his trans- lations from the Greek. As a large portion of Boethius's works consists of translations of Greek texts and commentaries thereon, one may naturally wonder to what extent Greek usage influences his style. Some light is thrown on the answer from his own words, CommentariP in Porphyrium, p. 71 a, in which he gives his theory of translation. " Secundus hie arreptae expositionis labor nostrae seriem translationis expediet, in qua quidem vereor ne subierim fidi interpretis culpam, cum verbum verbo exjjressum comparatumfjue reddiderim. Cuius incepti ratio est quod in his scriptis in quibus rerum cognitio quaeritur, non luculentae orationis lepos sed incorrui)ta Veritas c.\[)rimenda est." That our author carried out this purpose of a literal translation can be seen from even a cursory examination of his works; for traces of Greek usage are found in the commentaries and other works as well as in the translations. A few examples will suffice to illustrate my point. Qiiidcm- . . . autein and quuiein . . . vera in the sense of \i.\v . . . ' I cite the works of Boethius as follows: the two editions wepl 'Eptx7]velai, ed. C. Meiser; works on the quadriviuni, ed. G. P'riedlein; Consoln/io Philosophiae and Opiiscula Sacra, ed. Peiper; remaining works, ed. M(igne). In citing the works of Aristotle, I refer to the Tauchnitz edition for the first part of the Organon : to Waitz for the Prior a and Posterior a Analytica. ' As shown by E. K. Rand, Der dein Boethius zugesehriebene Triiktat De Fide Catkolica, Jahrhii.lier fur Klassische Philologie, XXVI, Supplementband, p. 42S ft. Stylistic Tests nnii Chronoloi^y of the JVor/vS of l^octhius 125 %k ajipear very often in the translations and in all the works that follow the DiaUnri in Porphyriiim. This usage, to a less degree, is found even in classical authors. See Cicero's Toxica, 51, 65 (?), 95 (?), and especially 60 : atque illud qiiiJem genus causarum, quod habet vim efficiendi necessariam, errorem adferre non fere solet; hoc auton sine quo non efficitur saepe conturbat. Another marked Graecism in Boethius is the use of qiioniam ^, quia, Y is the usual construction, whether with verbs implying a comparison or after comparative adjectives and adverbs. Of the former the following is a good example, Ilept Ep/x.7^V£tas, 10, 8 : ravra yap iKCLVwv Sta^epet TO) fXT) Ka66\ov eivai. Cf; Meiser, p. 14, 21 f. : haec enim ab illis diffe- runt eo quod non universaliter sunt. For the construction after adjec- See also the columns under quidem . . . atitern and quidem . . . zrro in my table on p. 138 below. ' There has been much discussion as to the origin of this use of quoniam, quia, quod in indirect discourse. For a review of the subject see Schmalz, in Berliner Philologische Wochenschrift, 1905, p. 557. Some argue that the usage comes in from the sertno plebeitis, others from the Greek. Probably Greek influence merely accelerated the adoption of the construction; for the idea was inherent in the language. To illustrate, there is so little difference between the infinitive after commemcro and the construction with quod, that it would have been surprising if the Latin writers had not been ready to make a free use of the latter. * Concerning the origin of this construction, Roensch, Itala und Vul^ata, p. 452, thinks that it crept in through the Christian writers from the Hebrew idiom. How- ever this may be, the discussion, in the preceding note, of the quoniam, quia, quod construction is applicable here also; for as the Latin ablative of comparison contained the idea of separation, it would have been strange, if, with the increase in the use of prepositions, a {ab') had not come to be used in constructions after comparisons. 126 Arthur Patch McKinlay tives see Porphyry, Isagoge, 3, 16 : *0 yap av6p(oiro<; irXeov €;(ei tov ^(uov TO XoyLKov Koi TO Ovrp-ov. Cf. M., p. 125c: Homo enim ab animali plus habet rationale et mortale. This use is frequent in the commentaries and other works. The influence of translation is further shown by the extraordinary frequency of idcirco quoniam, quia, quod ; idea quo7iiat?i, quia, quod and the like. How Boethius treated the Greek equivalents will appear from the following illustrations : 8ta to with infinitive {Wtpi 'Ep/xT/vcia?, 1,%z^ idcirco quoniam Meis., 7, 31), or idcirco quia {^Com. in Por- phyrium, 3, 20 ^ M., p. 129 c), or ideo quod (^Sophistici Elejichi, 5, 6 :=:M., p. ioi2d), or eo quod (Aristotle, Topica, 8, 12, 8 1= M., p. 1007 a), or propterca quod (Porphyry, Isagoge, 15,11= M., p. 155 d) ; TO) with infinitive (Aristotle, Categoriae, 6, 10 ^ idcirco quod M., p. 246 b), or eo quod (Aristotle, Topica, i, i, 9 = M., p. 911 b), or hoc quod {Categoriae, 6, 1 1 = M., p. 247 a) ; 810 =r eo quod {Topica, 4, 5, 7 ^ M., p. 950 c), or propter quod {ibid. 6, 9, 9 = M., p. 982 c) ; SioTTcp {ibid. 8, 12, \(i ^ eo quod M., p. 1008 b) ; SioTt {ibid, i, i, 10 ■=:eo quodM., p. 911 b). These collocations are so frequent that we even find liru turned by eo quod {Topica, 8, 10, 8:= M., p. 1005 b), or idcirco quoniam (Porphyry, Isagoge, 2, 19 rrr M., p. 100 c). I have collected all such collocations in the works of Boethius and find that, to omit translations, their sum approximates one thousand, a number which sufficiently attests their abundance. Naturally I do not hold that Boethius originated these expressions. I wish to show, however, that their frequency is due to his Greek studies.^ One of the most interesting of Boethius's Graecisms is his treatment of the definite article. At first he sometimes omitted it, as in Aristotle, Categoriae, 8, 19 to yap vyunvuv '^oiKpa.rrjv tw vo(T€iv '^uiKparrjv ivavTLov icTTLv. Cf. M., p. 278 d: Sanum namque esse Socratem ad languere Socratem contrarium est. Sometimes he used the collocation id quod dicitur. Cf. Com. in Caiegorias, p. 208 b: in eo quod ci . . . ce . . . ro dicitur; Editio prima irepl 'Epfj.r)V€La<;, p. 49, 11 : in eo nomine quod est homo ; and again T)e Sy/iogismis Categoricis, p. 795 b : in Ciceronis nomine, I.ater for this construction he generally used id ' This is all that is implied in Rand's discussion of (/uii/cm . . . vi-ro {^Jahrbucher Jur Klassische IViilolo'^ic, XXVI, Supplementhand, jip. 42S ff.), a point that Stangl (^IVochenschri/i fur Klassische J'/iilolo^ie, 1903, p. 179) seems not to understand. Stylisfii 7\sts and Chrono!os;y of tlic Works of lUn'tlnns 127 quod est. Cf. InirodNctio ad Sy//oi^ismos Categoricos, \). 763 it: in eo quod est Cicero. On this matter Boethius himself remarks {Editio prima ircpl 'Ep/i7;v£(.tts , p. 62, 17 ff.) : unde Graeci quoque his per se dictis verbis aliquotieiis adduiit articularia praepositiva, ut est to rpe- ^etv, Tov Tptx^iv, si qiiis enmi dicat : velocius est id quod est currere eo quod est ambulare, in illo nominativum iunxit articulum dicens id quod est currere, in illo vero ablativum dicens eo quod est ambulare. (See also below, p. 147). These illustrations make it evident that (Ireek idiom had great influ- ence on the style of Boethius. Consequently it is rather surprising that Friedlein should have been uncertain about the reading of the manu- scripts in the Dc Ariihinctica, p. S6, 1. 4 f. : quam secundum ad [?] (so Friedlein) aliquid speculamur. For we often find ad aliquid^ 171^6% tl not only in the translations, but also in the commentaries and the other writings. See Com. in Ca/i'i^-on'as, p. 213 b, and especially Sophistici Elenchi, 25, 4: 'O/Aotws Sc koX iirl tcuv Trpo's Tt, Cf. M., p. 1034 c: similiter autem et in ad aliquid. Here the preposition in accompanies the expression ad aliqind. Among other striking Graecisms in Boethius it will suffice to cite merely the following. The impersonal gerundive governs the accusative case. Sophistici Elenchi, 34, i : ttws Xvjiov tous Aoyous ^at cruAAo- yto-Mou's. Cf. M., p. 1039 I! : quomodo solvendum est orationes et syl- logismos. This use appears frequently in the translations. Again, dpa in questions is turned by piitasne {Sophistici Elenchi, 20, 6: apa olSas ^ M., p. 1030 A : putasne vidisti), or by uf pntas {ibid. 10, 10 = M., p. 1019 c). The participle with a.v Xo.vQ6.voi^ the participle with latebit. Ibid. 17, 19: fJi€Ta(f>ep(Dv av Tis Xavddi'OL to. ovofxara. Cf. M., p. 102SC: Transferens quispiam nomina latebit. OvM =1 neqiiidem (not ne ... qiiidcm). Cf. Aristotle, Topica, i, 4, 5 =z i\l., p. 913 A. oTL TocrauTa^ws = qiiod totidem modis. Cf. Sophistici Elenchi, 4, 2 =z M., p. loio a). Wherefore it is clear that literal translation is a feature of the style of Boethius. Still he was no slavish transcriber. His object, as has been shown above, ^ was accurately to convey the meaning of the original. ' See p. 124. 128 A rth ur Patch McKinlay Consequently he was not at all loth to depart from mere verbal trans- ference, if he could thereby better attain his aim of clearness. His use of examples attests this point. Some of the most striking are : Sophistici Elenchi, 4, 8 : koX to irtpX to ivv-nviov rov *Aya/u.€/[Avovo5, oTi ovK auTos 6 Zeus elrrev, ' St'So/Acv Se 01 eu;(os dpe'o-^ai.' Cf. M., p. loi I B : et id de Niso et Euryalo cum Rutulos vino somnoque sepultos intellexissent, Cetera per terras omnis animalia somno Laxabant curas et corda oblita laborum. Ilept 'Epixrjveux^, p. 4, 7 : KaAAtTTTros =: equiferus. Ibid., 1. 20 : $iAe Boethii Universo Colore, Pars Prior, 18S3, p. 32). This trait, though important, is more or less transient. Graecisms are most abundant in the translations, less frequent in the commentaries and comparatively scarce in the remaining works. This is doubtless the influence of Boethius's studies of Cicero. For example, see the table on page 139 under qiioniain, quia, quod. There is one further way in which the influence of translation made itself felt. Out of several modes of expressing a thought there would be a tendency for the translator to adopt one to the exclusion of the others, with the result that out of a number of synonyms one would prevail. The following table illustrates the point. 1 Dial.' in For. Interpr. Isag. Com.' in Tor. Interpr. Catcg. Com. in Categ. N.-ini £nim Namque 211 229 71 12 70 I 98 412 6 22 212 17 212 636 125 ' For convenience I ailhcre to the okl icnuinology. See note on p. 155. Stylistic Tests and Cliyo}iology of the Works of Boetliius 129 We see here that luimque is fre(iuent in the Dia!o.;i : that it almost entirely lapses in the Interpretatio Isin^o^i^iie, a fact which accounts for its rarity in the Conimentarii in Pi>r/>hyrii/m. It ajjpears more fre- ([uently in the Inferprctatio Catei^^oriarion, a fact which accounts for its frequency in the corresponding commentary. The predominance of efiiiii over nam in the translations explains the similar relation in the commentaries. It will be observed that this tendency to use one synonym to the exclusion of its competitors makes for the unification of vocabulary. There is another feature of Boethius's style that makes for the opposite, that is, his evident striving after variety. This, next to the effects of translation, is the most marked characteristic of his style. A few of the innumerable examples I have collected will be enough to enforce the point; a {al>) with the ablative, the ablative alone, and quam are used after comparisons. Cf. Com. in Ciccronis Topica, p. iioia: minus est animal rationale a simpliciter animali. Ibid, c : animal maius est homine. Ibid. : minus est animal rationale quam proprie animal, Compare also the constructions after diipliis. De Arithinctica, p. 162, 15 : duplus a (ab). Il>id., p. 141, 11 : duplus ad. Ibid., p. 165, iS: duplus with the ablative. Com. in Categorias, p. 218 b: duplus(um) with the genitive. Item, rursiis, ampliiis. Com. in Ciceronis Topica, p. 1166 a: Item, causarum aliae sunt non spontaneae. Ibid. : Rursus, causarum aliae sunt constantes. Ibid. : Amplius, causarum aliae sunt voluntariae. Tamqiiam, quasi. Com. in Porphyrin m, p. 91c: Fieri autem potest ut res, . . . non quasi genus, sed tamquani species sub alio collocatur. Quoniam with finite verb and accusative with infinitive in indirect discourse. Editio Secunda Trepl 'Ep/xT;vetas, p. 362, iSf. : Siquis dicat Socratem animal esse. Siquis praedicet quoniam Socrates bipes est. Ac, atquc, et, que. Com. in PorpJiyriuin, p. 134c: Itemque species ac differentia et proprium atque accidens.^ Therefore, to sum up the foregoing points, any stylistic study of Boethius must take into account two marked influences on his style — his methods of translation and his desire for variety. The former influ- ' Further illustrations of this tendency may be noted in Engelbrecht's treatise on the style of the Coiisolatio, in SitzungsbericlUe der IVicncr Akadeinie Jer Wiiseii' schaften, 1901, pp. 15-36. 130 Arthur Patch McKinlay ence helps to account for the appearance and disappearance of certain usages ; it tends to unification of vocabulary. Though important, it has in many cases only a transient effect. In contrast to this, the second influence — the evident aiming at variety — tends to diversity of diction. Bearing in mind the foregoing facts, we are now ready to Lake up our chronological study of the writings of Boethius. Any such research must be based on the painstaking and masterly investigation^ of Samuel Brandt. Utilizing all the references made by Boethius to his own writ- ings, he has fixed beyond all question the chronology of most of the works. He has made out an almost complete framework, leaving now and then a gap of more or less uncertainty which, I hope, may be at least partly supplied by my investigations. Brandt divides the extant writings of Boethius into fi\-e classes and arranges them chronologically as follows : I. Works on the quadrivium; De Arithmetica, De Musica, De Geome- tria. 2. Works on the principles of logic; Dialogi in Porphyrium, Commentarii in Porphyrium, Coninientarii in Cafegorias (510 a.d.). 3. Further works on the principles of logic; Editio Prior nepl 'Epfirj- vetas, Priora Analytica, De Syllogismis Caiegoricis, Editio Sccunda Trepl 'F,pfir]veLa 5_ 1 1 1 C 1 u & ^ ^ a u 'i 1 I g 2 3 E i 3 3 « 3 S V 71 7 211 229 '44 6 n II I 12 °i 24 24 3 15 ; 78 43 14 27 8 4 98 412 22 I 3 14 7 5 12 12 9 10 ISO 9 20 40 62 6 85 292 25 9 42 7 18 3 3 6 57 37 108 8 125 79 212 636 19 5 22 12 122 "3 2 I 42 34 ISO 31 72 4 55 49 "3 473 5 I 18 3 27 93 109 3 6 91 39 125 I 12 91 159 3 3 5 3 I I 12 3 14 30 40 5 34 126 107 277 1249 II 2 63 33 97 192 2 2 287 4 31 384 93 235 2 15 H 17 75 I 7 4 6 I I 17 13 25 I 17 13 47 lOI 4 6 4 5 5 I 26 3 9 29 26 18 "5 181 9 I 3 26 4 20 18 6 19 4 153 19 26 26 17 7 76 223 3 4 36 21 2 6 6 10 I 120 10 99 43 56 47 201 401 IS i' 37 12 4 25 25 2 I I 2 192 24 42 54 9 8 60 134 3 19 6 4 4 4 2 67 9 20 18 2 20 41 2 13 2 4 I 19 I 6 3 3 3 3 3. I I I 4 15 I I 19 63 7 4 23 5 4 16 16 81 116 2 19 26 13 13 2 I (I) 79 22 3 I I 3 2 I 8 3 140 Arthur Patch McKinlay The classes presented in the preceding table are as follows : First Class. — Works in which the influence of translation is little felt, as the Dial, in Porphyriiiin, or in which such influence is manifest- ing itself, as the Com. in Poi'phyrium. I call this the transitional period. Second Class. — The influence of translation is paramount; Com. in Categorias, Ile/oi 'Ep/xr/vetas , Editio Prior and Editio Sccunda. I call this the Greek period. Third Class. — Works in which the influence of Cicero is felt; Com. in Cic. Topica, Dc Diffcretitiis Topicis. I call this the Ciceronian period. Fourth Class. — Consolatio Philosophiae. Since these classes are clearly defined, it remains only to fit in the somewhat less certain works, most of which Brandt has placed to a greater or less degree of certainty. Of these the De Syllogisviis Cate- goricis seems to go between the two Editiones irepl 'Epfxrjveia's and after the Priora Anahtica. The Introductio ad Syllogismos Catcgoricos was written after the Prior Editio. Whether it antecedes the Dc Syllo- gismis Categoricis Brandt cannot determine. The De Syllogismis Hypotheticis follows the Editio Secunda and precedes the Commentarii in Ciceronis Topica. The De Divisione was written before the De Diffe?-entiis Topicis and probably after the works on interpretation. Practically every test in thq preceding table shows that the De Divi- sione is transitional between the works of the second and third periods. The rarity of quidcni . . . an tern, ergo, qiioniam, quia, quod, sic, and the frequency of que, ac, atque, quo, quo fit ally the Introductio ad Syllogismos Categoricos with the third period. Hence I place it later than the second edition of Ilcpt 'E/3yu.r;vetas . The same may be said of the De Syllogismis Hypotheticis. The De Syllogismis Categoricis is a peculiar work. The most cursory perusal will show that the book is si/i generis among the writings of Boethius. 1 1 is characterized by a marked paucity of stylistic pheno- mena and by a brevity strained to the utmost. In fact, at first sight it might appear that this work is the breviarium referred to on j). 251, II. 9-15, of the second edition of lltpl 'Ep/xT/vetus . Branilt {Ent- stehungszeit, p. 257) has exploded this theory. His conclusion is strengthened by the character of the introduction of the De Syllogismis Categoricis. This ])roaemium is elaborate and by no means leads us to anticipate the brevity of the body of the work. After the elaborate Stylistic Tists and Chronology of the Works of Bocthiiis 141 introduction come a few words tiiat furnish a clue to the style to be expected (M., p. 794 c-d) : inchoandum nobis est illo jjrius dei^ulso periculo ne a quoquam stenlis culpetur oratio. Non enim eloquentiae compositiones sed planitiem consectamur : qua in re si hoc efficimus quamlibet incompte Icx-iuentes intentio quoque nostra nobis perfecta est. Sterilis and incompte are terms particularly applicable to the style of the work in question. A good example of the former quality is seen in the way Boethius employs illustrations. To illustrate a declarative sentence, we find, De Syllogismis Categoricis, p. 797 b: Socrates ambulat. With this it is interesting to compare the Introductio ad Syllogismos Categoricos, p. 767 B, where to illustrate the same kind of sentence we find Virgil's words : Est mihi disparibus septem compacta cicutis Fistula. How applicable incompte is to the style of our work is shown from the following, — not that the usages are found exclusively in this work, but that they are so frequent as to be particularly noticeable. The preposition ad appears very often in the sense of secundum ; p. 799 b: ad quantitatem (cf. p. 800 d : secundum quantitatem) ; p. 799 c: ad eundem ordinem ; p. 797 a: ad placitum (cf. p. 795 c: secundum placitum). The constructions with. participo are also noteworthy, p. 799 a : parti- cipat ad utrosque terminos ; with in and the ablative, p. 798 c; with the ablative alone, p. 798 c; with the dative and the ablative after in, p. 798 c: in nullo siSi participantes. In the Dialogi in Porphyrium we find the genitive, e.g., p. 62 a: sui participari / the accusative alone in the Inteipretatio Aristotelis Topicorum, p. 945 b: c; also the preposition a {ad) with ablative, Ilcpt 'EpfMrjveians (^iiod 16 31 Stylistu Jisis iDui Chronology of the Works of Bac/hiiis 143 The rarity of the quiikm collocations still calls for comment (the one instance of (/i/if Boetliius 145 Hence he argued that these works were later than the Diahi^ite. Com- paring the use of the conjunctions it.ique, i^itiir, eri^o, he was also disposed to deny the authorship of the De Fide to Boethius. Professor Brandt {Rntstehuni^szcit, pp. 146 f.) thought that the stylistic method, to have value, must be based upon the many references of our author to his own works. As Brandt himself has furnished us this basis for our investigation, I propose now to supplement Professor Rand's work by including many more tests. Referring to the table on pp. 13S f., I will take each test separately and show in detail its place in the general scheme. The (jiiidem collocations are rare in the Dia/oi^i. Hence works that show the frequent use of them will be expected to belong to a period later than that work. liaque is rare during the second period. Ergo is frequent in the first period, in the second vies with igitiir, and later almost disappears. Que and ac are far more frequent in the third than in earlier periods. The same may be said of atque. Namque appears less and less frequently. It is rarely postpositive at first, later usually so. Nam is as frequent as enim only in the Dialogi. Whx the latter particle outstripped the former is shown above, p. 128. At vera is frequent only in the early works. The same is true of at vera si. Qiiemadmodiim is very frec^uent in the first two periods, but later is hardly found at all. Quasi is frequent from the end of the second period. Quocirca and quare are frequent in the second period, which fact is due to translation. (See above, p. 132). Quo yf/ and quo in the sense of therefore, except for an eccentric appearance in the commentary on Porphyry, date from the transition between the second period and the third. Porro and porro autetn are frequent only in the Dialogi. The qiioniam, quia, quod clauses are frequent only duriug the Greek period. Sic as compared with ita is rare in the third period. Item, except for the abnormal De Syllogismis Categorieis, is rare in the second period. Now we are ready to apply these tests to the De Arithmetica and the De Musica. Professor Brandt considers that they belong together 146 Arthur Patch McKi)ilay and must be accounted the first of Boethius's extant writings. He does this, relying upon a seemingly plain statement of our author to that effect in De Arithinctica, p. 5, 11. 19-24: Recte ergo, quasi aureos Cereri culmos et maturos Baccho palmites sic ad te noi'i operis rudi- menta transmissi. Tu tantum paterna gratia nostrum i)rovehas munus. Ita ct laboris mei primitias doctissimo iudicio consecrabis et non maiore censebitur auctor merito quam probator. Primitias seems to imply that this was the maiden effort of Boethius, Jiovi operis referring to the recent accomplishment of the work. Still, as Professor Rand has pointed out, these words need imply only that our author has begun a new task. As to Brandt's contention that these works come first and go together, a glance at the tabular view, pp. 138 f., will show that they do not belong to the same period. Instead of resembling each other in most stylistic criteria, as the other works of a given period do, they markedly disagree. Professor Brandt in a personal letter has suggested that this disagreement is due to the fact that the sources of the two works are different, for as has been shown, the sources have a marked influence on the style of a work. I had already noticed that with a new subject new words and constructions would come in, e. g. Dico quia, De Miisica, p. 303, 4 ; Po/ie, De Arithmetiea, p. 78, 30 : 79, 7 : 14. These criteria, however, are not the kind on which I rely, criteria such that when they have started, appear on almost every page of a work, e. g. et, aiitem, etc. Furthermore, if Professor Brandt's suggestion holds, we should find the diction in the two works very uneven, for their sources are manifold. The following tables will show that the use of a given particle in either work is fairly consistent. In each work the first sum under a given word, e. g. aiitem, is the total number of times it appears in the whole work. The figure just under is the number of times that word is found in the first half of the book. Atque Autciii ]iiiiiii 133 203 292 76 102 146 220 170 223 105 66 91 I have chosen these criteria al])habctically. The figures are fairly constant. Hence I conclude that stylistic divergencies in the two works are nut due to diverse sources. Ac At De Arith. . . 17 36 12 17 De Musica • 243 155 II I i;t Ktiain I-itur lt;i Na.n 924 54 105 57 ^^5 441 24 51 26 39 621 7J 412 120 76 277 43 171 61 3S Stylistic Tests and Chronology of t lie Works of Boethiiis 147 I come now to the other of Professor Brandt's contentions, namely, that the works on the quadrivium are the earliest of Boethius' extant writings. I had accepted this as the true view all the while that my material was collecting. When I began to study my results it was borne home to me that the De Miisica was about as different as could well be from the Dialoi^, which must have followed it within a few months if Brandt is correct. A glance at the tabular view (pp. 138 f.) will enforce this divergence. Cf. also these tables. At Nam si Quemad- pp. $ % Etiainsi moduin * Dial, in For 57 46 i. 79 i- 10 i De Musica 89 11 .15 19 .14 o 12 Deinceps Invicein Ut puta \'erc Atsi Ideo \'cTum Dial, in For o 12 5 11 o 39 o De Musica 10 i o o 4 11 8 Besides there are 53 other tests consisting of particles that appear from one to seven times in one of the two works and not at all in the other. Hence it hardly, seems likely that the De Musica was written only a short time previous to the Dialogi. If the De Arithmetiea and De Musica do not belong to the first period nor together, where do they come in the chronolog)' of Boethius' works? A reference to the tabular view on pp. 138 f. will answer that question. To begin with the De Aritiimetica, the quidem collocations place it later than the Dialogi. Itaque places it at the beginning of the second class. Ergo places it before the third class. Que, rp/iyrii Co/ninenfa, 1906). Aside from the text there is a valuable introduction. Brandt's comparison of the two editions is especially good. Brandt holds to his former position that the works on the quadrivium must precede those on Porphyry. Noting that Boethius, M., p. 70 D, proposes a com- prehensive study of Aristotle's works on logic, he thinks that we should have had a reference to the-Z'if Arithmetica if Boethius had intended also to treat of the quadri- vium. But observe, once more, that Boethius nowhere gave notice of his intention to take up Cicero's Topica, to which he later turned. Brandt also believes that a man, so propositi tcitacissimus as our author, could not have broken into his interpretation of Aristotle by interposing works on the quadrivium. And yet Brandt himself has pointed out a similar circumstance. In the passage referred to above, Boethius proposes to take up Aristotle's logic. Nothing is said about a second commentary on Porphyry. Brandt also recurs to the scarcity of quidem collocations in the first commentary on Porphyry. He thinks that although Boethius may have used them in earlier works, yet he may have laid them aside, for the time being, not meeting with them in Victorinus's translation of Porphyry. As a parallel, Brandt adduces the use of porro auteni. This occurs 24 times (Brandt's figures) in the first commentary on Porphyry, rarely elsewhere (see table, pp. 138 f.). He supposes that Boethius, noticing the solitary instance in Victorinus's translation, with a few more that may have dropped out of our text, took a notion to porro aulein and used it freely, later abandoning it. Now this is exactly the sort of evidence to which I have been appealing in this discus- sion. We are concerned, first, with noting genuine peculiarities, and then, if we can, with explaining them. Brandt's explanations might perhaps suffice here, if other criteria did not clearly place the Dial, in Porph. and the De Arithmetica in the first period, but the De Musica in the third. It is therefore more natural to account for the rarity of qtiidetn in the Dial, in Porph. on the ground that this work precedes Boethius's translations. A different cause, as explained above, operates in the De Syllog. Cat., of which Book I may be spurious. Instances of sporadic preferences, like porro autem, may be noted in all the works of Boethius. These are interesting to obser\-e, but I have cited only such peculiarities as illustrate a constant use or some marked development. Brandt's discussion of ijuidciii and porro autem shows that he believes such evidence worthy of consideration. In the light of many more phenomena of the same nature, considered in the same way, I venture to draw a different conclusion from his. I cannot agree, therefore, that the peculiar character of the De Syll. Cat. should oblige us to abandon the stylistic method in our efforts to 156 Arthur Patch McKinlay which I take from a personal letter to me. " Nondum persuadere mihi possum Boethium cum iam diu secundum propositum suum ad libros organi Aristotelici Latine tractandos et ad artem logicam et dialecticam exponendam operam suam contulisset, ad artem musicam explicandam redisse quae pars esset quadrivii." In other words, we are asked to hold of Boethius what Schleiermacher held of Plato, namely, that a man's life work is in embryo in the youth ; that we must expect no deviation from the plan outlined by our author in his second edition of the Ilept 'Ep/xr;vctas, p. 79, 10-80, I : "haec fixa sententia est, ut . . . ego omne Aristotelis opus . . . transferam atque etiam . . . omnes Platonis dialogos vertendo vel etiam commentando in Latinam redigam formam." Though these words seem to substantiate Professor Brandt's conclusion, yet it were rash to deny that some outside interest might intrude for a time — in fact we know that this was the case with Boethius. For all must concede that before he had carried out his plan of translating and perhaps of commenting on all the works of Aristotle and Plato, he had begun to work on Cicero. In the same way, he may have undertaken the De Miisica as a parergon. determine the chronology of the works of Boethius. On the contrary, as I have indicated, this aberrant work may confirm, perhaps decisively, the validity of the method. UNIVERSITY OF CALIFORNIA LIBRARY, BERKELEY THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW Books not returned on time are subject to a fine of 50c per volume after the third day overdue, increasing to $1.00 per volume after the sixth day. Books not in demand may be renewed if application is made before expiration of loan period. Ki--^ ^' "" ''^^2^, M\ 1'65-3PW SJl 29 SENT ON ru. U Apr'52AI 0CT2'i1996 U. C. BERKELEY 5^pr'eo?w IN STACKS MAR 22 1960 REG'D LD . fm^^Wi \U STACKS FEB 111965 75m-7,'30 U.C.BERKELEY LIBRARIES 'fIVlTllfti' I III mil III ill I n| ill II iilill mil n il I <:DM4Dm3'^M UNIVERSITY OF CALIFORNIA LIBRARY •■NJff