Harry_Austryn_Wolfson_The_Double_Faith_T

Prévia do material em texto

THE DOUBLE FAITH THEORY IN CLEMENT,
 SAADIA, AVERROES AND ST. THOMAS, AND ITS
 ORIGIN IN ARISTOTLE AND THE STOICS
 By HARRY AUSTRYN WOLFSON, Harvard University
 THERE are many phases to the problem of the relation of
 faith to reason and on almost every phase of this problem
 there is a considerable literature. It is not our purpose in
 this study to go over this familiar ground. Our purpose is
 to answer a single question which we have put to ourself.
 In the discussion of any problem of religion in its relation
 to philosophy in the Middle Ages, we find that the problem
 had its origin in the identification of some Scriptural term
 or concept with some corresponding philosophic term or
 concept. In the problem of creation, for instance, it was
 only when the Scriptural Creator became the philosophic
 Immovable Mover and Scriptural creation became philo-
 sophic substantial motion that the problem as we know it
 made its appearance. We therefore asked ourself whether
 in the problem of the relation of faith to reason there is in
 philosophy any term, analogous to that of motion with
 reference to creation, with which the Scriptural term faith
 was identified and out of which indentification there arose
 the problem as well as the various mediaeval theories of the
 relation of faith to reason.
 In answer to this question we shall show how already in
 Greek philosophy there was a conception of faith, according
 to which faith is the acceptance of the truth of two kinds
 of knowledge, one a knowledge whose truth is self-evident
 and not based on demonstration and the other a knowledge
 [105] 213
 THE JEWISH QUARTERLY REVIEW
 whose truth is based on demonstration. Then we shall try
 to show that in the discussion of the relation of faith to
 reason in the religious literature of Christianity, Judaism
 and Islam, the problem, as it formulated itself, was whether
 faith in its religious sense as used in the various Scriptures
 meant, as the term faith meant in philosophy, (1) the
 acceptance of the teachings of Scripture both as a self-
 evident truth and as a demonstrated truth, or whether,
 unlike the term faith in philosophy, it meant the acceptance
 of the teachings of Scripture (2) only as a self-evident
 truth or (3) only as a demonstrated truth. We shall refer
 to these three views respectively as (1) the double faith
 theory,' (2) the single faith theory of the authoritarian
 type, and (3) the single faith theory of the rationalist type.
 We may also observe that what we call here the double
 faith theory is to be distinguished from what is generally
 known as the double truth theory. The latter, in the
 opinion of the present writer, has a history and significance
 quite different from that which it is generally supposed to
 have. Following this analysis of the problem, we shall
 examine representative discussions of the relation of faith
 to reason in the various philosophic literatures.
 In the present paper we shall deal only with the double
 faith theory. As exponents of this view, we have selected
 Clement of Alexandria, representing the Greek Church
 Fathers, Saadia and Averroes, representing Arabic philos-
 ophy, Jewish and Moslem respectively, and St. Thomas,
 representing mediaeval Latin philosophy. We shall analyze
 the views of these four philosophers and present them
 according to a common pattern. In the case of each one of
 1 The description of faith as "twofold" occurs in the Stromata, II,
 11, of Clement of Alexandria, discussed below ? 2, at n. 56. But this
 "twofold" faith refers to what he calls "faith of scientific knowledge"
 and "faith of opinion," which is not the same as that which we call
 here "double faith."
 214 [106]
 DOUBLE FAITH THEORY-WOLFSON
 them we shall first try to show that he was conscious of the
 double faith theory as a purely philosophical concept, and
 then, we shall try to show, how each one of them similarly
 developed a double faith theory as a religious concept.
 ?1. THE ARISTOTELIAN 'FAITH' AND THE
 STOIC 'ASSENT'
 The particular use of the term faith in Greek philosophic
 literature, which is necessary for our purpose as the historical
 background for the use of that term later in the various
 religious literatures, occurs first in Aristotle. In Plato, the
 term faith is used only as a designation of a special kind of
 knowledge in his general scheme of dividing knowledge
 into sensation (alOrfaots), opinion (66}a) and scientific
 knowledge (ia7rtor-/1r).Ia Faith (7rLTLrs) to him, together
 with conjecture (CiKaoa'a), is a subdivision of opinion.2
 Now in Aristotle, the term faith is still sometimes used in
 the sense of a special kind of knowledge, in such a passage
 as that, for instance, in which he says that faith is a vehe-
 ment assumption (vbr6X\7t'Ls o'poSp&),3 assumption itself
 being used by him as somewhat the equivalent of opinion
 or as a sort of opinion.4 But from many other passages in
 which Aristotle speaks of faith, we may gather that he
 also uses that term not in the sense of a particular kind of
 knowledge by the side of the other kinds of knowledge, but
 rather in the sense of a judgment of the truth of all the
 other kinds of knowledge, that is to say, as a judgment
 that the knowledge we have of a thing, be it sensation or
 opinion or scientific knowledge, corresponds to the reality
 Ia Theat., 187A ff.; cf. Zeller, Plato and the Older Academy, p. 171;
 Grote, Plato, II, p. 374, n. 1.
 2 Zeller, op. cit., p. 174, n. 14.
 3 Top., IV, 5, 126b, 18.
 4 Cf. J. A. Stewart, Notes on the Nicomachean Ethics, IV, 3, 1139b,
 17 (II, 34); J. Geyser, Die Erkenntnistheorie des Aristoteles, pp.147-49.
 [107] 215
 THE JEWISH QUARTERLY REVIEW
 of a thing- correspondence to reality being that which,
 according to him, constitutes truth.5
 In this new sense of the term, faith would seem to be
 regarded by Aristotle, in its relation to the various types of
 knowledge, in the same way as he regards consciousness in
 its relation to the knowledge of sense-perception. With
 regard to consciousness, he says that "there is also a kind
 of common faculty that is associated with all the particular
 senses, by the virtue of which a person is conscious (alaO&-
 verca) that he sees and hears."6 Similarly with regard to
 faith, he could have said, though he does not explicitly say
 so, that there is a certain faculty of judgment in the mind,
 by virtue of which a person comes to believe and to be
 certain that the knowledge he possesses is true. Thus,
 speaking of what he calls indemonstrable and immediately
 known "primary premises" (apXal), he says, "things are
 true and primary which obtain faith not on the strength of
 anything else but of themselves."7 Again, speaking of
 opinion, of which he says that "it may be true or false,"8
 he also says that "opinion is attended by faith, for it is
 impossible to hold opinions without faith in them,"9 and
 "further, every opinion implies faith (7TrlTrt), faith implies
 persuasion (r6 7rE7TrerOaC), and persuasion implies reason
 (X6oyos)."1? Finally, referring to sensation and demonstra-
 tive reasoning, he says that "faith is not only on the
 strength of sensation but also on the strength of reason
 (X6'oVo)"i or that "faith concerning this is from induction
 and syllogistic proof."I2
 s Metaph., IV, 7, 101b, 27.
 6 De Somno, 2, 455a, 15-16; cf. De An., III, 2, 425b, 12.
 7 Top., I, 1, lOOa, 30-100b, 18.
 8 De An., III, 3, 428a, 19.
 9 Ibid., 19-21. Io Ibid., 22-23.
 -" Phys., VIII, 8, 262a, 18-19.
 12 De Soph. Eleuch., 4, 165b, 27-28; Anal. Post., III, 3, 90b, 14;
 Top., I, 8, 103b, 7.
 216 [108]
 DOUBLE FAITH THEORY-WOLFSON
 From all these passages we may gather, then, that by
 faith in this new sense of the term Aristotle means the
 consciousness of the certainty of the truth of our knowledge,
 and that this consciousness of certainty, called faith, may,
 according to him, be associated with two kinds of knowl-
 edge: (1) animmediate kind of knowledge, such as sensa-
 tion and primary premises, and (2) a derivative kind of
 knowledge, such as that which is based on reason. Though
 of this latter kind of knowledge, to which Aristotle says
 that the term faith is applicable, he mentions explicitly
 only 'opinion,' yet from his statement that faith follows
 syllogistic reasoning one would be justified in inferring that
 the term 'faith' would apply also to scientific knowledge
 (f'rt~ar'/jr), for scientific knowledge, to him, is the con-
 clusion (arv/utrpaoarua) of a syllogistic proof.I3 Indeed,
 conclusions in scientific proofs, unlike opinions, are said by
 Aristotle to follow from their premises by necessity,I4 and
 hence their acceptance would seem to imply no persuasion.
 On the basis of this, one would therefore be tempted to say
 that, unlike opinion, scientific knowledge is not a matter of
 faith, for the latter, as Aristotle himself says, does imply
 persuasion.15 But the persuasion which Aristotle says is
 implied in faith is itself said by him to imply reason,'6 and
 nowhere does he say that this is to be only a faulty kind of
 reason and one which is to lead only to an inconclusive
 demonstration and hence only to opinion. On the contrary,
 it may be also that sound kind of reason which leads to a
 conclusive demonstration and hence to scientific knowledge.
 The very fact that he applies the term faith to primary
 premises whose truth is self-evident argues for the assump-
 '3 Eth. Nic., VI, 3, 1139b, 31-36.
 '4 Ibid., 18-24; Metaph., V, 5, 1015b, 6-9; cf. below ?2, n. 53;
 ? 3, n. 84; ? 5, n. 160.
 's Cf. above n. 10.
 I6 Ibid.
 [109] 217
 THE JEWISH QUARTERLY REVIEW
 tion that he would also apply it to conclusions which follow
 from primary premises.
 Still, despite all this, it is true that explicitly Aristotle
 does not use the term faith with reference to scientific
 knowledge; only the terms sensation, opinion and primary
 premises are explicitly used by him in this connection.
 Let us keep this observation in mind for future reference; it
 will become of importance in our discussion of St. Thomas.'7
 The fact that Aristotle always speaks of faith as a judg-
 ment of the truth of our knowledge, and never as a judg-
 ment of the goodness of our desires, would seem to indicate
 that, according to him, faith was only a function of the
 speculative intellect (vovs OewpfrnK6s), for it is the
 speculative intellect to which he ascribes the power to
 distinguish between truth and falsehood in the abstract.'8
 It would not be a function, according to him, of the practical
 intellect (vovs TrpaKTLKOS), to which he ascribes the
 power to distinguish between what is good and desirable
 and what is not good and undesirable.19 And if we are right
 in this conclusion that Aristotle considered faith as a
 function of the speculative intellect only, and not of the
 practical intellect, then we would be right also in concluding
 that Aristotle did not make faith dependent upon free
 choice (7rpoalpec?s) or will (/3ovXfraLs), for free choice
 and will, according to him, are the result of the cooperation
 of practical intellect with desire.20 In fact, we find that at
 least in one place Aristotle makes use of another term as a
 designation of the judgment of the practical intellect as to
 the desirability or the good of a thing. In that one passage,
 speaking of the method of determining "which is the more
 17 Cf. below ? 5.
 I8 Eth. Nic., VI, 2, 1139a, 27-28; cf. De An., III, 9, 432b, 27 ff.
 '9 Eth. Nic., VI, 2, 1139a, 29-31; cf. De An., III, 10, 433a, 15 ff.
 o2 Eth. Nic., VI, 2, 1139a, 21-26; cf. De An., III, 10, 433a, 22-25.
 [110] 218
 DOUBLE FAITH THEORY-WOLFSON
 desirable, or better, of two or more things,"2I he says that
 "our reasoning faculty (&L&voLa) will agree or assent
 (v'yKaTraOTerraL), that whichever side happens to excel
 is the more desirable."22 According to this passage, then,
 the judgment of the mind that a thing, on account of its
 superior excellency, is more desirable and better, is not a
 matter of faith but rather one of assent (o,vyKaraOeats)
 and that this assent is in our reasoning faculty (&L&voCa),
 which term is undoubtedly used in this passage in the sense
 of the practical intellect.
 This analysis of Aristotle's conception of faith, based
 upon carefully selected and controlled texts, could furnish
 us with all the necessary historical background for the
 problem which we have undertaken to investigate. But it
 happens that, by the time the problem of faith in its reli-
 gious sense makes its appearance, the Aristotelian view has
 already become combined with similar views propounded
 by the Stoics. Let us then see what was the Stoic view,
 and how it became combined with that of Aristotle.
 In the literature which records the teachings of the
 Stoics, only one definition of faith is to be found. It is
 said of them that they defined faith as "a strong assumption
 (V7r6Xr/LS iohxvpa), confirming that which is assumed."23
 Now tbis is nothing but a reproduction of Aristotle's
 statement that faith is a "vehement assumption,"24 which
 we have quoted above and which, as we have pointed out,
 refers to faith as a special kind of knowledge by the side of
 sensation, opinion and scientific knowledge. Of the use of
 faith in the sense of a judgment of the truth of all the kinds
 21 Top., III, 1, 116a, 3.
 22 Ibid., 11-12.
 23 Stobaeus, Eclogae II, 112.12 (ed. Wachsmuth); Arnim, Stoicorum
 Veterum Fragmenta, III, 147.11.
 24 Cf. above n. 3. Wachsmuth's change of Vro6X0\Lts to KaTarLX?p)I
 in the Eclogae (loc. cit.), which is adopted by Arnim (loc. cit.), is
 therefore unnecessary.
 [11] 219
 THE JEWISH QUARTERLY REVIEW
 of knowledge, such as we have found in Aristotle, the Stoics
 do not speak. They do speak, however, of assent (arv'yKaTa-
 Oeats) as being such a judgment. Now in Aristotle, as we
 have seen, the term assent, as a verb, is used with reference
 to the judgment of the practical intellect as to whether a
 thing is desirable and good. What the Stoic seems to have
 done here, therefore, is to extend the meaning of that term
 so as to include also the meaning of Aristotle's term faith
 and thus to use it as a judgment of the truth of a thing as
 well as a judgment of the goodness of a thing.
 This extension on the part of the Stoics of the meaning
 of the Aristotelian term 'assent' so as to include both the
 judgments of practical intellect and the judgments of
 speculative intellect has led them to speak of assent as an
 act of free will (ktp' 7/u,tv),25 even when it is used in the
 sense of Aristotle's use of the term faith, namely, as a
 judgment of truth. This would seem to be a departure from
 Aristotle. To Aristotle, 'assent,' as a judgment of the
 good of a thing, is indeed an act of free will, for virtue is
 defined by him as an act of deliberate choice (TrpoalpEoLs),26
 but, as for 'faith,' he would probably say that it is an act
 of deliberate choice only when applied to opinion, of
 which he says that it implies 'persuasion."27 With reference,
 however, to scientific knowledge, to which the term faith
 is undoubtedly applied by him,28 he would probably say
 that it is not an act of deliberate choice, for the determina-
 tion of truth is according to him an act of the speculative
 intellect, in contradistinction to deliberate choice which is
 an act of the practical intellect.29 Furthermore, he explicitly
 25 Cf. Index to Arnim, op. cit., p. 136, col. 1, 11. 7-12; cf. Zeller,
 Stoics. Epicureans, and Sceptics, p. 87, n. 1.
 26 Eth. Nic., II, 6, 1106b, 36-1107a, 2.
 27 Cf. above n. 10.
 28 Cf. above text at n. 16 ff.
 9 Cf. above n. 20.
 220 [112]
 DOUBLE FAITH THEORY-WOLFSON
 says that demonstrations by which the conclusions of
 scientific knowledge are established belong to those things
 which are necessary and that similarly the conclusions in
 such demonstrations, by reason of the fact thatthey cannot
 be otherwise, must likewise be necessary.30 Still this
 difference between the Stoics and Aristotle is not of
 real significance. The free will which the Stoics attribute
 to the act of assenting to the truth of our knowledge
 does not mean a freedom of choice in an absolutely
 undetermined sense. Their freedom of choice in the assent
 to the true, like Aristotle's freedom of choice in the assent
 to the good, merely means a choice free from any per-
 ceptible external compulsion and one which is made by the
 agent with a full knowledge of the rightness or wrongness
 of his act.31 In this case, as in many other cases, one may
 quote Cicero's statement to the effect that the difference be-
 tween the Stoics and Aristotle is only one of terminology.32
 With the extension of the meaning of the term assent to
 include under it what Aristotle calls faith, the Stoics use
 this term as a judgment of truth with reference to all those
 immediate and derivative kinds of knowledge in relation
 to which Aristotle uses the term faith. With reference to
 sensation, they say that "all sensation is an assent,"33 or
 that the perception by the senses "cannot take place
 without assent."34 Similarly with reference to indemon-
 strable primary promises, they say that "the notions of
 things" (notitiae rerum), in the same way as memory and
 30 Metaph., V, 5, 1015b, 7-8; 1015a, 33-35.
 3I Cf. Zeller, op. cit. (n. 25), p. 87, n. 1; p. 205; J. S. Reid, M. Tulli
 Ciceronis Academica, p. 223, n. to 1. 20.
 32 De Fznibus, IV, 9, 21.
 33Aetius, Placita, IV, 8, 12 (Diels, Doxographi Graeci, p. 397);
 Arnim, op. cit., II, 26.39-40.
 34 Cicero, Academica Priora II, 12, 37; Academica Posteriora I, 11,
 40; Arnim, op. cit., II, 35, 12; I, 19, 1-3.
 [113] 221
 THE JEWISH QUARTERLY REVIEW
 art, can have no existence without assent.35 And so also
 with reference to acquired knowledge, they say that "all
 opinion (66aa) and judgment (KpnLts) and assumption
 (v7r6Xr\/tLs) and learning (lu07criLs), by which we live and
 have perpetual intercourse with the human race, is an
 assent."36 While, as in Aristotle, there is no definite state-
 ment by the Stoics that assent attends scientific knowledge
 which results from demonstration proofs, the term ,u&arOaLs
 used in this last-quoted passage is undoubtedly the equiv-
 alent of CirtOrrn/ur and means scientific knowledge.37
 These then are the essential elements in the discussion
 of faith in Greek philosophy as a purely epistemological
 term. The salient point in this conception of faith is that
 it is of a double nature. It may be faith with reference to
 something immediately known as true without any demon-
 stration, such, for instance, as sense-perception and primary
 premises, or it may be faith with reference to something
 known by demonstration, such as opinion and undoubtedly
 also scientific knowledge. Now it was quite natural for
 Christian and Jewish and Moslem philosophers, to whom
 revelation was a new immediate source of knowledge in
 addition to the sources enumerated by the Greek philos-
 ophers, and to whom the teachings of Scripture, while
 having its origin in revelation, could also be demonstrated
 by reason, to raise the question whether faith in these
 teachings, which is enjoined by Scripture itself, means the
 same as the term faith used in philosophy or whether it
 means something else. In the former case, it would have a
 double meaning, namely, assent to the teachings of Scrip-
 35 Cicero, Academica Priora II, 12, 38; Arnim, op. cit., II, 35, 15-17.
 36 Clement of Alexandria, Strom., II, 12 (Migne, Patrologia Graeca,
 8, 992C; ed. Stahlin, II, 142.26-28).
 37 The term /iaOr7aLs Ltavor7TK77 in Anal. Post., I, 1, 71a, 1, is
 undoubtedly used by Aristotle in the sense of iTrLo'rflyV, which he
 discusses subsequently in 2, 716, 9 ff.
 [114] 222
 DOUBLE FAITH THEORY-WOLFSON
 ture both as immediately known by revelation and as
 derivatively known by demonstration. In the latter case,
 it would have only a single meaning, namely, assent to the
 teachings of Scripture either (1) only as immediately known
 by revelation or (2) only as derivatively known by demon-
 stration, the two views which we have designated respec-
 tively as (1) the authoritarian single faith theory and
 (2) the rationalist single faith theory.
 ?2. CLEMENT OF ALEXANDRIA
 The first to apply this Aristotelian double faith theory
 to the problem of the relation of philosophy to Scripture is
 Clement of Alexandria. He is also the first to combine the
 Aristotelian term faith with the Stoic term assent. He thus
 speaks of faith as an assent of the soul (/vx7s rv?yKara-
 o0eLs)38 and, reflecting more closely Aristotle's statement
 that "our reasoning faculty (tlavota) will assent ovy'Ka-
 raOrfoerat)," he uses also the expression "on assent of
 the reasoning faculty" (oavyKaTaOeLats &Lavolas).39 Once,
 after quoting in the name of the Stoics their statement
 that "all opinion and judgment and assumption and
 learning... is an assent,"40 he adds "which is nothing else
 than faith."41 Again, following the Stoics he also maintains
 that assent is an act of choice (TpoaCpecfts),42 or that it is
 voluntary (gKocLos) ,43 or that it is exercising free will
 (avTrEov(Lotos),44 or that it is in our power (ehp' ,/Uv).45
 38 Stromata, V, 13 (Migne, Patrologia Graeca, 9, 128A; ed. Stahlin,
 II, 383. 1). English translation in The Ante-Nicene Fathers, II.
 39 Ibid., VIII, 5 (9, 581A; III, 90.3). 40 Cf. above, ? 1 n. 36.
 41 This seems to be Clement's own addition. It is so taken by
 Arnim, who does not include it as a quotation from the Stoics (cf.
 Arnim, Stoicorum Veterum Fragmenta, II, 291.1-3).
 42 Strom., II, 2 (8, 940B; II, 117.15).
 43 Ibid., V, 13 (9, 128A; II, 383.1).
 44 Ibid., V, 1 (9, 12C; II, 327.24).
 4s Ibid., II, 12 (8, 992C; II, 142.26).
 [115] 223
 THE JEWISH QUARTERLY REVIEW
 In this instance, however, though the language used by
 Clement is the same as that of the Stoics, its meaning is
 different. To the Stoics, free will merely means act free of
 external compulsion and based upon knowledge. To
 Clement it means, as it does to Philo, real freedom which
 man enjoys as a gift from God.46 His assertion, therefore,
 that assent is a voluntary act really means that we are
 absolutely free either to believe that a thing is true or not
 to believe that it is true.
 Again, as faith in Aristotle and assent in the Stoics, so
 also faith which is assent in Clement is a judgment of the
 truth of our knowledge, whether it be of the immediate
 kind of knowledge or of the demonstrative kind. With
 regard to the application of the term faith to immediate
 knowledge, such as what Aristotle calls "primary premises"
 and the Stoics as well as the Epicureans call "preconcep-
 tions," he makes the following statement: "Epicurus, too,
 who very greatly preferred pleasure to truth, assumes faith
 to be a preconception of the mind (7rp6Xr\7qts &tavolas);
 and defines preconception to be a grasping at something
 evident, and at the clear understanding of a thing; and
 asserts that without preconception, no one can either
 inquire, or doubt, or opine, or even argue."47 With regard
 to derivative knowledge, we have already quoted his
 restatement of the Stoic use of the term assent with refer-
 ence to opinion, to which he adds the statement that it is
 "nothing else than faith."48
 But it is quite certain that Clement understood Aristotle
 as having used the term faith also with reference to that
 kind of derivative knowledge which is technically known as
 scientific knowledge, for he states the following in the name
 46 Cf. "Philo on Free Will," Harvard Theological Review, XXXV
 (1942), 131-69.
 47 Strom., II, 4 (8, 948B; II, 121.9-13).
 48 Cf. above n. 41.
 224 [1161
 DOUBLE FAITH THEORY-WOLFSON
 of Aristotle: "Aristotle says that the judgment (KpTu)a)
 which follows scientific knowledge (Crtro'r7,n) is in truth
 faith. Accordingly,faith is more decisive than scientific
 knowledge, and is its criterion. Conjecture (edKaaoa), which
 is only a feeble assumption (aoOevt7s vir6Xr/ts), counter-
 feits faith."49 In this passage, his reference to "feeble
 assumption" as that which "counterfeits faith," evidently
 reflects Aristotle's statement, quoted above, that faith is a
 "vehement assumption."50 His statement, however, in the
 name of Aristotle that faith may also apply to scientific
 knowledge is not found in Aristotle. As we have shown,
 there is no direct statement in Aristotle to that effect
 though it can be indirectly inferred from some of his state-
 ments.51 This statement, however, is based on a passage
 in the work of Theodoret of Cyrus, in which this interpre-
 tation of Aristotle is to be found.52 That Clement himself
 accepted the use of the term faith with reference to scientific
 knowledge is quite evident. But how he could reconcile his
 own definition of faith as a voluntary assent with Aristotle's
 conception of scientific knowledge as knowledge which
 follows by necessity as conclusions from the primary
 premises,53 may only be conjectured. He would probably
 say that the human mind in its absolute freedom can refuse
 assent even to conclusions which follow by necessity from
 their premises, and consequently, when it does assent to
 them, the assent may be called voluntary. We shall see
 that St. Thomas has another view on this point.54
 Faith with reference to derivative knowledge is thus,
 according to Clement, of two kinds, namely, either of
 49 Strom., II, 4 (8, 945C-948A; II, 120.25-28).
 50 Cf. above ? 1, n. 3, and Migne, PG, 8, p. 947, n. 62.
 51 Cf. above ? 1, text at n. 17.
 52 Cf. Migne, PG, 8, 947, n. 61. The reference is to Theodoretz Graec-
 orum affectonum curatzo I, 90 (ed. Gaisford, Oxford, 1839).
 53 Cf. above ? 1, n. 14.
 54 Cf. below ? 5.
 [117] 225
 THE JEWISH QUARTERLY REVIEW
 opinion or of scientific knowledge. Now, according to
 Aristotle, opinion may be either true or false, whereas
 scientific knowledge, strictly speaking, refers only to
 knowledge which is true.55 It is evidently with reference
 to this distinction between opinion and scientific knowledge
 that faith is described by Clement as being twofold (&BTrr),
 the faith of scientific knowledge (TrLotrolovtLK) and the
 faith of opinion (5otaaTKfK).56 The former is a faith in
 knowledge which is true, whereas the latter is a faith in
 knowledge which may be either true or false.
 This, then, is Clement's restatement, in Stoic terms, of
 Aristotle's conception of faith as a purely epistemological
 term.
 Coming now to faith in its religious sense, Clement
 defines it primarily as "obedience to the commandments."57
 Such an obedience to the commandments is also described
 by him as the grasping of the teachings of Scripture "by
 faith," in the same way as one grasps an "indemonstrable
 primary premise" for, like a 'primary premise,' the teach-
 ings of Scripture are self-evidently true, inasmuch as they
 have been given to us "by the voice of the Lord."58 Faith,
 in its religious sense, thus has the meaning of assent to the
 teachings of Scripture as an immediate sort of knowledge,
 for which no demonstration is necessary. But, while the
 teachings of Scripture can be accepted as an immediate
 truth without any demonstration, it does not mean that
 they cannot be demonstrated by reason or that one ought
 not to attempt to demonstrate them by reason. Throughout
 his work he argues both for the possibility of their being
 demonstrated by reason and for the desirability of their
 55 Cf. Eth. Nic., VI, 3, 1139b, 16-18; Anal. Post., II, 19, 100b, 11-12.
 56 Strom., II, 11 (8, 984C; II, 138.20-21).
 57 Ibid. (8, 984C; II, 138.27).
 58 Ibid., VII, 16 (9, 532C; III, 67.25-28).
 226 [118]
 DOUBLE FAITH THEORY-WOLFSON
 being so demonstrated.59 But once these teachings of
 Scripture have been demonstrated by reason, they cease
 to be true in the sense that a 'primary premise' is true;
 they become true in the sense that a "conclusion" (arvu7repa-
 orua) in a valid demonstration is true.60 This, then, is the
 second meaning of the term faith in its religious sense.
 Now, strictly speaking, Clement does not use the term
 faith with reference to the assent to teachings of Scripture
 after they have become demonstrated by reason. He uses
 in that connection another term. He calls it knowledge
 ('yviaLs) - a term adopted by him from the Gnostics
 but it is often used by him interchangeably with the term
 scientific knowledge (Crtror-/r) .6I But sometimes this
 technical distinction is abandoned by him and "knowledge,"
 in the sense of demonstrated Scriptural teachings, and
 "faith," in the narrow sense of undemonstrated Scriptural
 teachings, are treated by him as two subdivisions of the
 term faith taken in a general sense. Thus in one place he
 says that scientific demonstration produces "scientific
 faith" (ETtrrntOVLKu) T7rTLrs) which becomes knowledge
 (yiw,l s).62 The implication is that 'knowledge' is not
 something different from faith, but rather a sort of faith.
 In another place he makes the following statement: "Faith
 is, then, so to speak, an epitomized knowledge (aov-
 rouos... . yv&o'ts) of the essentials; and knowledge is a
 strong and sure demonstration of what is received by faith
 (7rTirrew)."63 Here, again, faith and knowledge are not
 59 Ibid., I, 5-7, 9, 20; cf. J. Patrick, Clement of Alexandria, pp. 40-43.
 6o Ibid., II, 11 (8, 985A; II, 139.8-11).
 6i Ibid., II, 11 (8, 984C; II, 138.17); VI, 17 (9,388A; II, 511.27-30).
 Cf. Index to Clement of Alexandria's works, ed. Stahlin, IV, under
 Yvw&Lts, 10.
 62 Ibid., II, 11 (8, 985A; II, 139.6-8). According to another reading
 of the text (cf. MPG, 8, p. 985, n. 4, and ed. Stahlin), scientific dem-
 onstration is said to produce "faith" which becomes "knowledge."
 63 Ibid., VII, 10 (9, 481A; III, 42.3-5).
 [119] 227
 THE JEWISH QUARTERLY REVIEW
 different from each other; each implies the other; "knowl-
 edge," which is a "demonstration of what is received by
 faith," may, therefore, be called "demonstrated faith,"
 just as it is called by him "scientific faith." In still another
 place he says that "the discourse which consists of demon-
 stration implants in the soul of him who follows it the exact
 faith (aKpLt3l 7irtiLr)."64
 But, more than that, Clement throughout his writings
 tries to establish the equality of these two forms of faith.
 Arguing against those who already by that time, out of their
 opposition to the teachings of the Gnostics, opposed the
 application of philosophy to Scripture, he tries to show how
 philosophy can contribute to the comprehension of divine
 truth.65 On the other hand, against the Gnostics, who
 considered the simple faith of believers as inferior to the
 demonstrated faith of the Gnostics, he argues that inasmuch
 as the teachings of Scripture are true, whether they are
 demonstrated by reason or not, the simple believer is not
 inferior to the rationalizing theologian. "He is not then
 imperfect," he says, "who knows what is perfect"66 and
 "nothing is wanting to faith, as it is perfect and complete in
 itself."67 The simple believer, therefore, who possesses only
 'faith' and the philosopher who possesses also 'knowledge'
 are equal before God, for "there are not in the same Word
 some 'gnostics' and some 'psychical men'; but all who have
 abandoned the desires of the flesh are equal and spiritual
 before the Lord."68 Indeed, even though undemonstrated
 faith is not less perfect in the eyes of God than demonstrated
 64 Ibid., I, 6 (8,728B; II, 22.2-3).
 6s Ibid., I, 20; cf. references above in n. 59.
 66 Paedagogus, I, 6 (8, 280A; I, 105.2-3).
 67 Ibid. (8, 285A; I, 107.17-18).
 68 Ibid. (8, 288A-B; I, 108.23-6); cf. E. De Faye, Clement d'Alex-
 andrie, pp. 204-5.
 228 [120]
 DOUBLE FAITH THEORY-WOLFSON
 faith, still, philosophically, he admits, "Knowledgeis, so
 to speak, a kind of perfection of man as man."69 Con-
 sequently Clement recommends that believers who started
 with mere assent to Scriptural teachings to acquire knowl-
 edge as far as it is possible for them.70
 The equality of the two kinds of faith is also implied in
 his view that both of them are roads to salvation, the one
 being the long road and the other the short road. This
 view is expressed by him in the statement that if any Greek,
 who has been brought up on philosophy, dispenses with
 reason and assents to the teachings of Scripture, he has
 chosen "the short-cut to perfection, viz., that of salvation
 through faith."7' Revelation, in its relation to philosophy,
 is thus only a short-cut to true knowledge.
 Why there should be need for such a short-cut and why
 mankind should not be allowed to work out their salvation
 by the efforts of their own reason, is explained by his
 younger contemporary Origen in terms which will prove
 of importance for us in our subsequent discussion of the
 problem. Answering Celsus, who spurned faith without
 reason, he says that "if it were possible for all to leave the
 business of life, and devote themselves to philosophy, no
 other method ought to be adopted by anyone, but this
 alone... But since the course alluded to is impossible,
 partly on account of the necessities of life, partly on account
 of the weakness of men, as only a very few individuals
 devote themselves earnestly to study," it is better for the
 great multitude of men "to believe without a reason" and
 "on the strength of mere faith" rather than to wait "until
 they could give themselves to a thorough examination of the
 69 Strom., VII, 10 (9, 477C; III, 40.21-22).
 7" Ibid. (9, 477C; III, 40.27-29).
 7' Ibid., VII, 2 (9, 416A; III, 9.20-21; Hort and Mayor, Clement of
 Alexandria's Miscellanies, Book VII, 18.19-21).
 [121] 229
 THE JEWISH QUARTERLY REVIEW
 necessary reasons.72 Here, then, we have a statement as
 well as a justification of the double faith theory.
 The definition of faith as a voluntary assent, which was
 introduced by Clement of Alexandria in the particular
 religious discussion of the relation of faith to reason, be-
 comes an established term in the discussion of the problem
 throughout its history in Christian literature, whether
 Greek or Latin, both among those who follow Clement of
 Alexandria's particular double faith theory and among
 those who in opposition to him contended for a single
 faith theory, either of the authoritarian or the rationalist
 type. The term appears again in modern philosophy under
 the guise of practical reason, under which Kant includes
 all speculations about faith. When William James flashed
 upon the world his phrase "the will to believe," he was
 merely giving expression to his own understanding of the
 old definition of belief as a voluntary assent. But it is not
 our purpose in this paper to trace the development of this
 conception of faith. Our purpose is to discuss three other
 examples, taken from Jewish, Moslem and Christian philos-
 ophy, of this double faith theory.
 ?3. SAADIA
 From the Church Fathers, through personal contact
 between Moslems and Christians, the problem of the rela-
 tion of faith to reason was introduced into Islam during the
 Umayyad period. Later, with the rise of the Mu'tazilite
 sect, under the second of the Abbasid rulers, speculations
 about faith and reason developed into a system. But in the
 writings which record the theological discussions among the
 Moslems before the time of Saadia (882-1042), with all
 their wealth of material on particular problems, there is
 72 Contra Celsum, I, 9 (MPG, 11, 761B ff.).
 [122] 230
 DOUBLE FAITH THEORY-WOLFSON
 to be found no extensive discussion of the general problem
 of the relation of faith to reason. Of the Mu'tazilites we
 know in a general way that they considered it as a "duty"
 to arrive at the "knowledge of God" by "speculation and
 proofs"73 or "by reason,"74 and from what is known of their
 general attitude it may be inferred that they held a single
 faith theory of the rationalist type. Similarly from what
 is known about the Ash'arites we may gather in a general
 way that they held a single faith theory of the authoritarian
 type. In Saadia's Emunot ve-De'ot, therefore, we find the
 first extensive discussion of the relation of faith to reason in
 Arabic philosophy, whether its Moslem or its Jewish
 branch.
 The view represented by Saadia, we shall try to show, is
 that of the double faith theory, like that we have met with
 in Clement. As a purely epistemological term, we shall try
 to show, the term faith is used by him in the sense of a
 judgment of the truth of both immediate and derivative
 knowledge. As a religious term, we shall also try to show,
 it is used by him in the sense of the judgment of the truth
 of Scriptural teachings both with demonstration and
 without demonstration, and, moreover, both these two
 kinds of faith in its religious sense are considered by him
 as of equal degrees of perfection, each of them being the
 perfect religion for those to whose needs it is adapted. In
 presenting now Saadia's view on faith and reason, we shall
 therefore pursue the method which we have chosen to follow
 throughout this study, namely, to state first his conception
 of faith as a purely epistemological term and then his
 conception of it as a religious term, though in his own
 discussion the two aspects of the term are merged together.
 Logically, though not actually, the starting point of his
 73 Shahrastani, p. 62, 11. 19-20.
 74 Ibid., p. 87, 1. 17.
 [123] 231
 THE JEWISH QUARTERLY REVIEW
 discussion is his formal classification of the types of knowl-
 edge.75 These types of knowledge are divided by him into
 three, corresponding to what Aristotle would call (1) sensa-
 tion, (2) primary premises and (3) scientific knowledge.
 Though this threefold classification of the types of knowl-
 edge was already a commonplace by the time of Saadia,76
 still the terminology used by him needs some comment.
 Sense perception is described by him as the knowledge
 of that which is visible.77 The term visible here reflects
 the Greek 6par6v, which is used as a description of cor-
 poreal objects which are perceptible by the senses.78 Then,
 what Aristotle would call primary premises Saadia calls
 "the knowledge of the intellect."79 This evidently reflects
 Aristotle's statement that it is the intellect (vovs) which
 apprehends the primary premises,8? on account of which,
 in his various classification of types of knowledge, Aristotle
 himself describes intuitive knowledge by the term vovs.81
 75 Emunot ve-De'ot, Hakdamah (ed. Josefov, 1885), ? 5, pp. 43 f.
 (Arabic, ed. Landauer, p. 12, 11. 17 f.).
 76 Cf. references to the Ikhwan al-$afa (Arabic in Dieterici, Die
 Abhandlungen der Ichwdn es-Safc in Auswahl, pp. 196, 212, 521) in
 Guttmann, Die Religsonsphilosophie des Saadia, p. 22, n. 5.
 77 Emunot ve-De'ot, loc. cit., p. 43 (p. 12, 1. 20- p. 13, 1. 1): [c
 a.'J.l n1 ,, ny,T,.
 78 See, e. g., Timaeus 46D. The same term is also used by the
 Ikhwan al-$afa in their description of this kind of knowledge; cf.
 op. cit., p. 196, 11. 4-5; p. 212, 1. 21; p. 521, 1. 4. In Arabic philosophy
 this term as a description of sensible things has gained currency evi-
 dently through the additional fact that it is used in that sense in the
 Koran 9.95.
 79 Emunot ve-De'ot, loc. cit. (p. 13, 1. 4): JiAJ I [': I ,n y.' .
 80 Anal. Post., II, 19, 100b, 12. But in Emunot ve-De'ot, Hakdamah,
 ? 2, p. 37 (p. 2, 1. 12) the term ".A.JI j \ J31: 'DWtn Y7D, as
 may be judged from the context, is used in the sense of ratiocinative
 thinking, the term J'ic: tv in that passage, unlike the same term in
 this passage, thus reflecting the Greek X6yos rather than vovs; cf.
 a similar double use of the Latin term intellectus, below ? 5, n. 154.
 8I Anal. Post., II, 19, 100b, 7-8; De An., III,3, 428a, 4-5; Eth. Nic.,
 VI, 3, 1139b, 16-17.
 232 [124]
 DOUBLE FAITH THEORY-WOLFSON
 Finally, what Aristotle calls scientific knowledge (CerL-
 a'Trlir), which is identified by him with conclusions of
 demonstrations, Saadia calls "a certain knowledge which
 necessity leads up to"82 or "the knowledge of necessary
 things."83 This expression, too, reflects a passage in Aris-
 totle which states that "a demonstration is a necessary
 thing, because the conclusion cannot be otherwise, if there
 has been demonstration in the full sense; and the causes of
 this necessity are the first premises, i. e., the fact that the
 propositions from which the syllogism proceeds cannot be
 otherwise. "84
 Now these three types of knowledge, as will be noticed,
 fall into two groups, of which one, that which includes
 sensation and primary premises, is immediate knowledge,
 and the other, that which includes logical conclusions, is
 derivative knowledge. Consequently, when Saadia, pre-
 vious to his classification of the types of knowledge, comes
 to define faith as the judgment of the truth of knowledge,
 he has two definitions of it, one with reference to imme-
 diate knowledge and the other with reference to derivative
 knowledge.
 With reference to knowledge that is immediate, he defines
 faith as "a concept which arises in the soul of anything
 known according to the condition in which it really is."85
 In this definition, the expression "a concept which arises
 in the soul," I take it, refers to both sense-perception and
 82 Emunot ve-De'ot, loc. cit. (p. 13, 1. 1): o,)jsJ1 c?l I. A s
 4JO: 1\b mnr n7m;lW Dr nyn'.
 83 Ibid. (p. 13, 11. 5-6): ot j.Jl31 L: n'nnDnn y-D.
 84 Metaph., V, 5, 1015b, 6-9; cf. above ?1, n. 14; ?2, n. 53, and
 below ? 5, n. 160.
 85 Emunot ve-De'ot, loc. cit., ?4, p. 42 (p. 11, 11. 4-5): '-ni' i3mrn
 1'7 m naiy (,5- ) l':Y n; l KDfr<1 .(rs^y l) =mc' n ^. = V
 ;p^y N1in itDw ;nzDn yl'n- ni- i (W9]3==-^_;)-
 [125] 233
 THE JEWISH QUARTERLY REVIEW
 primary premises, reflecting Aristotle's statement about
 sensation that out of particular sense-perceptions a univer-
 sal concept (Ka6bXov) is formed in the soul (ev rpj /'vx),86
 and also his statement about primary premises that it is
 the mind (vovs) which apprehends them.87 The expression
 "of anything known according to the condition in which it
 really is" is the conventional restatement of the corre-
 spondence theory of truth as given by Aristotle.88
 With reference to derivative knowledge, he defines faith
 in the following figurative language: "When the cream of
 speculation is extracted, the intellect takes it under its
 wings, folds itself around it, causes it to penetrate into the
 soul and to become intermingled so that man becomes
 thereby a believer in the concept thus arrived at and he
 keeps it in reserve for some later occasion or occasions."89
 In this passage, the opening statement "when the cream
 of speculation is extracted" means when by an act of
 speculation a conclusion is derived from primary premises.
 The remaining part of the passage means to say that the
 conclusion, which technically speaking is a "second notion,"
 derived by demonstration from a primary premise or a
 "first notion," may itself, after it has been fully approbated
 and accepted by the mind, become a primary premise out
 of which other conclusions may be derived.90 In Aristotle
 there is an allusion to the use of conclusions of one demon-
 86 Anal. Post., II, 19, 10Ca, 16.
 87 Ibid., 100b, 12.
 88 Cf. Metaph., IV, 7, 1011b, 27.
 89 Emunot ve-De'ot, loc. cit. (p. 11, 11. 5-8): ll'y;l ntin myn T
 ann 'I,n' nm tarm (nl3D]=@= ) nl:ni 'IwD'1 ,pD'p' iWn. nI)Zp'
 0PnO lN mnnN nyi Inyzzl IrH yani 1KN 3'ZYM I'DD -rtNI.
 90 On this meaning of "second notions," see Millot ha-Higgayon,
 Ch. 8, and Rua4 Hen, Ch. 3; Milzamot Adonai, Introduction (p. 4).
 It differs from the scholastic term "secundae notiones; cf. my The
 Philosophy of Spinoza, II, pp. 121-22.
 234 [126]
 DOUBLE FAITH THEORY-WOLFSON
 stration as a premise in another demonstration in his refer-
 ence to the geometrical proposition that "the angles of
 every triangle are equal as two right angles" as something
 which a student "knew beforehand" (rpo'j8E),9' that
 is to say, as something to be used as a premise in a demon-
 stration. Now the equality of the angles of a triangle to
 two right angles is not a self-evident primary premise; it
 is deduced, as Aristotle himself says elsewhere, from the
 proposition that a straight line set up on a straight line
 will make angles equal to two right angles92 or, in other
 words, it is the conclusion of a demonstration.
 Saadia then proceeds to show, in true Aristotelian fashion,
 how each of these three types of knowledge may be either
 true or false. In the case of the sense of sight, for instance,
 he says, the knowledge arrived at by it is true when there
 is a real external object which we see, for, in that case, our
 knowledge corresponds exactly to reality, inasmuch as our
 organ of sight is assimilated to the perceptible object and
 becomes identical with it, and this on account of the likeness
 between the combining ratios of the same four elements
 which constitute the forms of both the perceptible objects
 and the organ of sight.93 It is untrue, when we see the
 image of an object reflected in a mirror or in water and
 91 Anal. Post., I, 1, 71a, 19 ff.
 92 Cf. Metaph., IX, 9, 1051a, 24-26, and Euclid, I, 32 and I, 13;
 cf. also Phys., II, 9, 200a, 16-18, with Ross's Commentary ad loc.
 93 Emunot ve-De'ot, loc. cit., p. 45 (p. 15, 11. 15-16): ySn ': nltD3
 nnMK 1In btr yN-l 'IbD I ra3n1 1r3a - pl:nI 'nDon m13ln 91I' iv i: elnnn
 imnvtrn nvo; cf. I, 3(8), p. 74 (p. 20, 11. 7-8): o,,on m,s' 1i'3' ylu'i
 12'y y t3: y -D'llnnD Dn^'nn ' rr' nl-li y mK ? K4 H1 nM t DK 'D; II,
 12, p. 102 (p. 106, 11. 9-11): ',wm13n a,'r: ot a'D'n: m'K D':Tnm
 Y1 nl1Kn -W K nnlr Dy D'-nnO131 oD'Yty3 nYiKNI D1'DfnlDnT DorrnoPa
 lKl'l i' N11m nlyXzp)MrM t"P
 All this is based upon Aristotle's statement in De An., II, 5, 418a,
 5-6, with regard to the sensitive faculty, that "when once it has been
 acted upon, it is assimilated and has the same character as the sensible
 object."
 [127] 235
 THE JEWISH QUARTERLY REVIEW
 think that that reflection is a real object.94 Similarly in the
 case of primary notions, he maintains, "all those which are
 formed in our mind when in a wholesome state are true
 and not subject to any doubt, provided only that, after
 we know how to form by speculation such notions, we do
 form them in a perfect manner."95 As an example of per-
 fectly formed and hence true notions he mentions such a
 generally accepted ethical notion as that "truth is good
 and falsehood is horrid."96 Now these generally accepted
 ethical notions belong to what Aristotle calls maxims
 ('iJLgat) in rhetoric,97 describing them as being generally
 known or agreed upon or self-evident,98 corresponding to
 primary premises in demonstrative syllogisms, except that
 they deal "with objects of human actions, and with what
 should be chosen or avoided with reference to them."99
 In fact, Arabic logicians include such maxims under what
 they call by the general term "primary premises,"I??
 describing them specifically as "generally known primary
 premises."IoI But, on the other hand, he continues, there
 are also imperfectly formed and hence untrue notions in
 94 Ibid. (p. 15, 11. 18-21): n-i mI1n , ' vn nwv ny: r,n y HN il
 -N -IDlipn 1It -rWK . 3 NID ... nDV34 DvW UT1- T 1Y N '1 r ^' IN1D' nKin
 K'N, nym t-Dn n DK NIm nrosm D0,; DNzn.
 95 Ibid. (p. 16, 11. 4-6): i: 1nzn vzi -a ,,'ar -vN i:l mSmDn -1
 'r;y y'ri 1' nK (J-j) '1"Y3 1'N YIDV n ,P3D 1 1'S nmD pTI HiUm yp3
 (>J1).
 I take it that the terms Jl; : rl"Y and j1.Ji: l'yn, in this passage
 do not mean the art of speculation or of reasoning but rather the
 formation of concepts in the mind. In the passage quoted below in n.
 113, the expression_,JI 4;L. : pwyn n:DnO refers to the art of rea-
 soning, i. e., logic.
 96 Ibid., p. 43 (p. 13, 1. 5): ,m13 i rtnl ml prtlnw lO.
 97 Rhet., II, 20, 1393a, 23-24.
 98 Ibid., II, 21, 1394b, 10-15.
 99 Ibid., II, 21, 1394a, 24-26.
 100 Cf. Algazali, Maqaszd al-Faldsifah, I, p. 52, 1. 4: ,UjWl.
 o10 Ibid., p. 55, 11. 9 ff.: .lj)lJl; cf. Millot ha-Higgayon, 8: mnDDmDD.
 [128] 236
 DOUBLE FAITH THEORY-WOLFSON
 our mind, to which he refers as "phantasms and dreams."I02
 These he describes as being due sometimes to "the thoughts
 which happened to cross through the mind during the past
 day" and sometimes to "the foods, with reference to their
 having been hot or cold, much or little," and sometimes to
 "the humor which happens to gain dominance over the
 mixture; the warm and moist become the dream-images of
 merriment and pleasure; the cold and dry become the
 dream-images of grief and mourning."I03 Finally, in the
 o02 Emunot ve-De'ot, loc. cit., p. 45 (p. 16, 11. 6-7): oi,'nn 1n nDmai
 (^A.Yl,J) nlDlmnm [i..J1]. These terms evidently reflect Aris-
 totle's statement that "every image ((pavraao,a) in sleep is a mere
 dream (Yb'irviLov)" (De Somnits, 1, 458b, 24).
 103 Ibid. (p. 16, 11. 9-15 ff.): nwi 9inn Dm',n ,'a'yD n,nrr onp ,D yir, K5i
 oncpi ... DolDoi D1311 D1nipi Dml Ollron1t 'DsD Dnpil ... vDnnn y Il1zy
 nln- nnm oDinnm (:ljJl) aro1 ^ iman (XSvOs= ^ O lJ iDDU '"D
 (t; 1 ,-) nlmnl (;Lr) r nolrT wn,'ni npn pln ','i, mnDi (r^,,).
 From the fact that in this passage, the s.,: DD, which happens
 to become predominant in the T\lA: uT, is said to be either hot or
 moist, cold or dry, it is evident that it refers to one of the four humors,
 for the four humors correspond to the four primary qualities of the
 elements (cf. Hippocratis de humoribus liber et Galeni in eum com-
 mentarius, I, in Galeni opera, ed. Kuhn, XVI, 23-24). The literal
 Hebrew translation of xvI,os, >^S", juice, humor, is nrn,5. But,
 inasmuch as in Arabic the common word for humor is >;., literally,
 mixture, KparLs, /uLiLS, the translator renders cj,) here by 1DD,
 mixture. Conversely, too, we find that the Hebrew ninl,, literally,
 juice, is used as a translation of the Arabic l?;-, literally, mixture.
 For examples see the quotations in Klatzkin, Ozar ha-Munatim
 ha-Pilosofiyyim, s. v. All this interchange of terms is due to the fact
 that the terms humor and mixture of humors became interchangeable,
 owing to the fact that usually in every mixture of humors one particular
 humor predominates and lends its name to the mixture as a whole.
 But see another interpretation of the term lDD here in Philipp
 Bloch, Vom Glauben und Wzssen: Saadiah's Emunoth we-Deoth, Mun-
 chen, 1879, p. 27, n. 1, which is adopted by Klatzkin, op. cit., sub 1DD.
 The second and third causes of dreams mentioned here by Saadia
 are also given by Abraham Ibn Ezra in his commentary on Eccl. 5.2:
 ln nnM ilmnml D'wNDn nanynn Nlnm ylI n o''ly o'ayo a-lyo Doinn rn"mmr
 .tin ^in 'D yll MID by nt41 IlnnD In -1 n'n Ni oVwnw nmy:nim.
 [129] 237
 THE JEWISH QUARTERLY REVIEW
 case of derivative knowledge, he concludes, it is true when,
 as in the case of what Aristotle calls scientific knowledge,
 it follows certain rigid logical rules of demonstration. It
 may be false, however, when, as in the case of what Aris-
 totle calls opinion, there is some laxity in those rules of
 demonstration, and Saadia thereupon enumerates five such
 rules that are to be observed in demonstrative reassuring.104
 Since both immediate and derivative knowledge may be
 either true or false, faith with reference to both these kinds
 of knowledge may be either true or false. In Clement, as
 will be recalled, there is a similar distinction between faith
 of scientific knowledge, which is true, and faith of opinion,
 which may be false.?05 Drawing upon the terms used in
 the definition of the correspondence theory of truth, Saadia
 defines these two kinds of faith as follows: "True faith
 means that a thing is known to be as it really is, the many
 to be many, the few to be few, the white to be white, the
 black to be black, the existent to be existent and the non-
 existent to be nonexistent. False faith means that a thing
 is known to be the opposite of what it really is, the many
 to be few, the few to be many, the white to be black, the
 black to be white, the existent to be nonexistent and the
 nonexistent to be existent.1?6
 This, then, is Saadia's restatement of Aristotle's double
 faith theory as a purely epistemological concept. Let us
 now see how he treats of the same theory as a religious
 concept.
 Besides these three types of knowledge which may be
 "04 Ibid., p. 46 (p. 18, 11. 16 ff.): ,nnDn,I yrnn rri'nrr ' int 11'
 IDDn, ]D ml n"i o -i l'3t3., '-Ix.
 o"s Cf. above ? 2, n. 56.
 O6 Emunot ve-De'ot, loc. cit., ?4, p. 42 (p. 11, 11. 10-14): rmDnm
 ,-lnrw nrlnnl j1 l mn ,Ly D l -om yn - : in ,rn iN 'WH:- "y'tvw , N n,'n,'on
 -inU ,Kn -In D I.n 1 2In y't 7'V NK1 n n nlt .-n? C tr-y3 : 3 N1x3 m
 H2YP3 xn) 3l N m KD3II ,129 Tinwnl '*nrw imv , Y1 Dy.II DYD.
 238 [130]
 DOUBLE FAITH THEORY-WOLFSON
 either true or false, Saadia mentions a fourth type of
 knowledge, which he considers as being unqualifiedly true.
 He calls this fourth kind of knowledge "true tradition" and
 by this he means the teachings of Scripture as well of those
 of the oral Jewish tradition which are supplementary to
 Scripture.107 Now these teachings of Scripture and of the
 oral Jewish tradition may be accepted by the believer as
 true without any demonstration, in the same way as one
 accepts the truth of indemonstrable primary premises.
 But they may also be demonstrated by reason and thereby
 become accepted as true in the same way as one accepts
 conclusions of syllogistic proofs as true. To Saadia both
 these kinds of acceptance of the truth of these teachings
 constitute faith in the religious sense of the term. The
 former is described by him as the faith of him "who believes
 by tradition" and the latter as the faith of him "who
 believes by speculation and understanding."Io8 Thus, as in
 Clement, we have here a double faith theory in the religious
 sense of the term.
 Moreover, again as in the case of Clement, both these
 kinds of faith are regarded by Saadia as of equal perfection.
 Against those who might think that the religion of the
 simple believer is inferior to that of the philosopher, Saadia
 briefly declares that the religion of those-who are incapable
 of speculation is "perfect."109 But being more concerned
 with those who in his own time objected to the use of
 philosophy in connection with religion, he undertakes to
 I07 Ibid., p. 44 (p. 14, 11. 2 ff.): u,n D''DoD o'n,nrn Inp i3nai Jzb
 n3l nl'i ?m iniKa. la^3n '3'1 1 WD n'1K rmn DPn -rD -im aDwn -iwtn
 (jva23 .:;-1) fnDNKn anru n mm ' vWl; ); cf. the use of the terms
 nmn ,mannr and nnDK nmna ( r. . .-) in III, 6, pp. 113, 114
 (p. 126, 1. 13; p. 127, 1. 4).
 o08 Ibid., ? 2, p. 39 (p. 6, 1. 7): ,atn; l y l' -n1 p n 1'p n l'in mi,l.
 Io9 Ibid., ? 6, p. 51 (p. 25, 1. 18): ,?v 'Di 'c-iyy N , 1W 'D I 'I, ',D
 n'tw inn n,nn yl"VY Y'.
 [131] 239
 THE JEWISH QUARTERLY REVIEW
 refute all the popular prejudices against the application of
 reason to Scripture and, moreover, to show on many
 grounds that the application of reason to Scripture is
 desirable.II0 But while philosophic speculation about
 matters of faith are permissible and while also such specula-
 tions, if they are sound and founded on true principles,
 cannot conceivably be in disagreement with the teachings
 of Scripture,'IT one is not to make his belief in these teach-
 ings dependent upon the outcome of one's own rational
 investigation nor is one to relax in one's belief pending the
 outcome of such personal investigation. "We are not
 allowed," he says, "to put the books of the prophets aside
 and to rely upon what may occur to the mind of each one
 of us by his own reasoning when taking up for considera-tion such questions, for instance, as the beginnings of
 place and time. For he who employs speculation in matters
 religious in this manner may by himself either hit the right
 belief or miss it. Consequently, until he hits the right
 belief, he will be without belief. And if he happens to hit a
 right belief and gets hold of it, he will not be sure of himself
 that he will not part from it at the rise of some doubt in
 his mind, and thus his belief will be ruined for him."II2
 Belief, Saadia would say, is a habit as well as a conviction,
 and such a habit can be formed by a decision of the will
 and cultivated by practice.
 With his conception of the equality of undemonstrated
 faith and demonstrated faith, Saadia raises the question,
 which we have already met with in Origen, as to what
 lo Ibid., p. 48 (p. 20, 1. 18- p. 21, 1. 22).
 I Ibid., p. 49 (p. 22, 11. 6-10): 1i3 ts, nipmn 1Y": nWvt ' l3,'Trm
 rp112 w ln bid 13,nn; 1N'13 12 0 y 3-yn'l;l 1m3 yw i:: DhWIn -1in
 z3n3lnnv m'PD'nni c'K i1 lnn AminaYD 3 'uwy 0'791z?.
 112 Ibid. (p. 21, 11. 15-20): by Dn1m Yix a'm,',:nr '-=D ltyiW ly:D 15:3
 ' ,aip;l-rn Inrnlimnnn iny-t by inSynn oxy nyn nrlmi nnm I Kx,Y'W no
 t4D' DN1 ,n Him ;rnrr' H -'Ww nl ,nvyW zWB91 tSXDW iN nrn -tXn Inp IrnyDn
 innolN lhy -rDD'1 ,1vxsz -Tlny, pwSn mon) D InpnynD nr t m na p'lrm n-r;n.
 240 [132]
 DOUBLE FAITH THEORY-WOLFSON
 need was there for revelation, when men can attain to the
 same truth by their own reason. His answer is like that
 given by Origen. Speculative faith, he says, may indeed be
 suitable for those persons who by natural ability, training
 and leisure are able to carry through a logical argument to
 a successful conclusion. But there are those who, for
 various reasons, are lacking either the leisure or the ability
 to reason out for themselves the truths of religion. "God
 in his wisdom," he says, "knew that problems which are
 to be solved by the art of speculation can be brought to a
 conclusion only after the passage of some tract of time.
 Consequently, if God were to refer us to the art of specula-
 tion for a knowledge of His Law, we would each have to
 remain without the Law throughout the time that we
 were engaged in gaining proficiency in that art and per-
 fection in the use of it. And perchance many among us
 would never become proficient in the art of speculation on
 account of some deficiency in them; or they would never
 become perfect in the use of that art on account of some
 aversion to it which might overtake them or of some doubts
 which might overpower and perplex and bewilder them.
 From all these troubles God has quickly saved us by sending
 to us His prophets.""3
 Thus, as in Clement, revelation is to Saadia a short-cut
 to salvation, and the short-cut and the long road both lead
 to faiths which are of equal perfection. For the perfection
 of faith, argues Saadia in effect, consists in the certainty
 113 Ibid., p. 51 (p. 24, 1. 19- p. 25, 1. 5): ' minDmn yrp'v , DnMlI
 nnz', a~ , ,Drn Dl nun ^D 'o ,D= ' H, lVyn n5=4an 'a1,Inn a,'plDon
 pDyn anl ,nr6n tUrni y -R , bnn im lir mlD~y ,n, y in-nn ny-,M= anr
 1i IvH'1, bv ,1K .13C 11iDmn inay nzKiDn in o6wn ti 1ioDD mi t1nv .mn
 j~y ,iln mlpsDmw , mD 1K ,(4iJ., 3.J) a: Y'p'm reDy: n: pDynn9
 rmnlt 1ui nrti nnnmn , D': nnn ;1rHm nD :tin lnnn ,irnlvm ::I .
 The Hebrew yip in Gen. 27.46, Num. 21.5, and Prov. 3.11, is trans-
 lated by Saadia himself by ,. For pnrt In inD, see Eth. Nic. II, 1,
 1103a, 15-17: Acquisition of intellectual virtue requires time.
 [133] 241
 THE JEWISH QUARTERLY REVIEW
 and strength of one's conviction in the truth of what is
 believed, irrespective of the manner by which that truth
 was arrived at. The ideal believer is described by him as
 "he who has arrived at the truth and knows it and rejoices
 in it."II4 Imperfection of belief, similarly, is not to be
 identified either with a person's lack of rational arguments
 in support of his faith or with his use of such arguments.
 Nor is it even to be identified exclusively with a person's
 belief in what is not true, or with his lack of decision as to
 what beliefs are true. A person's faith may be called im-
 perfect also if he lacks firm conviction in what he has
 chosen to believe, even though he happens to have chosen
 a belief which is true. He thus distinguishes from the type
 of the perfect believer quoted above three types of imperfect
 believers. These are first, "he who has arrived at the
 truth but he is still doubtful about it, being neither assured
 of it nor holding fast to it."II5 Second, "he who assures
 himself of that which is false on fancying that it is true, so
 that he clings and holds to falsity and abandons that which
 is right.""6 Third, "he who preoccupies himself with a
 certain view for some time, but gives it up on account of a
 certain defect he notices in it. He then transfers himself
 to another view for some time, but parts from it on account
 of something in it of which he disapproves. He then trans-
 fers himself to still another view for some time, but after
 "4 Ibid., ? 2, p. 38 (p. 3, I. 22- p. 4, 1. 1): y-n' mmt ni n ri y,nw , '
 ir7 nDBn ,nim.
 IIs Ibid. (p. 4, 1. 3): nn4riD r3rm nn pmnoDD tmni n=n t?4 y3w ?D
 ,: p'rno Ki1 1SeH.
 II6 Ibid. (p. 4, 11. 5-6): Di1t nmn(i nm p,nLm i xN nrrnnu t)w ,D r on
 r,'pl t,(;j.) n.1..n ) nD, (.)jJl: ) Nm'I p'rnmD mr n DonD 1W _nm lpWn,
 irmon urn r,nn KW y nym Iva 13N' 5 (w'5 ,'ll r1't) nDn. The mirr in the
 printed text is to be read mslw. In his translation of, and commentary
 on Job, Saadia takes the wi in the verse of Job quoted by him here
 as being the equivalent of nsw; cf. Bacher's in Version Arabe Du
 Livre de Job de R. Saadia ben Josef al-Fayyium, Paris, 1899, p. 51;
 cf. also D. Kaufmann, Gesammelte Schriften, III, p. 439-40.
 242 [134]
 DOUBLE FAITH THEORY-WOLFSON
 a while he rejects it on account of something which has
 rendered it faulty in his eye. And thus he vacillates from
 one view to another throughout his lifetime.""7
 ?4. AVERROES
 The development of the problem of the relation of faith
 to reason in the Moslem branch of Arabic philosophy was
 the reverse of that of its Jewish branch. Arabic Jewish
 philosophy started with Saadia's double faith theory, then
 developed, in Hallevi, a single faith theory of the authori-
 tarian type, and ended up with Maimonides' single faith
 theory of the rationalist type. Arabic Moslem philosophy,
 on the other hand, started with the Mu'tazilite view of a
 single faith theory of the rationalist type, then developed,
 in Algazali, a single faith theory of the authoritarian type,
 and ended up with Averroes, a contemporary of Maimon-
 ides, with a double faith theory like that of Saadia and
 Clement of Alexandria.
 Unlike Saadia who takes as the subject of his discussion
 the term i'tiqad, faith, Averroes takes as the subject of
 his discussion the term tasdzq, which from its original
 meaning "the consideration of a thing as true" it came to
 mean also "affirmation" and "belief" and "judgment."
 We shall try to explain Averroes' use of this term before we
 undertake to explain his double faith theory.
 Early in the history of Islam there appeared the question
 as to what constitutes religious faith (iman), by virtue
 of which a believer could be properly called a Moslem.
 Some said that faith is only a confession (iqrar) of the
 tongue of certain articles of the creed, even it were not
 II7 Ibid. (p. 4, 11. 7-10): lz nn-i mlyrn 1n nygm iry nK rr,nw n o DDni
 I3i 113 l3y: lDKDI o li nmnNt nyn im lo= p'nynl ,O1 wityi 1312 1l3y3 lDKD
 1"'Y "113Y3 itryi r in nnn nyn n i Doo p'nyn l: nnl ,13a r13y nImi
 1D' i: nl"Ennn3 n1nl ,l"py3 lnl '1n TDW.
 [135] 243
 THE JEWISH QUARTERLY REVIEW
 accompanied by a sincere conviction of the truth of the
 creed. Others said, quite on the contrary, that it is a belief
 (tasdiq) in the heart. Still others said that it is a combina-
 tion of the two. There were finally others who said thatin
 addition to the confession of the tongue and the belief of
 the heart, it is also the performance of good works."8
 Now all this was a problem of practical religion, which
 concerned itself with the question as to what were the
 duties required of a true Moslem. Historically, it would
 seem to be a repercussion within Islam of certain discussions
 among the Christians with whom Moslems had come in
 contact, which had their origin in such New Testament
 statements as "That if thou shalt confess with thy mouth ...
 and shalt believe in thine heart. . . thou shalt be saved""9
 and "faith without works is dead."120 It had nothing to do
 with the problem of the relation of faith to reason. The
 expression "belief of the heart," which may also mean
 "the conviction of a truth in the heart" had no implication
 that that conviction must be based upon reason and
 demonstration. We may gather this from the fact that
 such exponents of opposite views with reference to the
 relation of faith and reason as the Mu'tazilite al-Najjar
 and Al-Ash'ari both declare that faith is a belief (tasdiq) of
 the heart.121
 Now, Averroes happens to believe that faith is not only
 confession or good works but also a belief of the heart.
 Accordingly he uses the term tasdiq instead of imdn or
 i'tiqdd in his formal discussion of the relation of faith to
 reason.
 With this preliminary comment, we shall now try to
 118 Cf. L. Krehl, Beitrage zur Charakteristik der Lehre von Glauben
 im Islam, pp. 7 ff.; F. A. Klein, The Religion of Islam, pp. 39 ff.
 I9 Romans, 10.9.
 120 James, 2.20.
 I2I Shahrastani, p. 62, 1. 20; p. 73, 11. 2-3.
 244 [136]
 DOUBLE FAITH THEORY-WOLFSON
 recast Averroes' entire discussion of the relation of faith to
 reason, in his special treatise devoted to this problem,122 in
 the general pattern of the problem as we have presented it
 in this paper. We shall discuss first Averroes' conception
 of faith in its general epistemological meaning and then
 we shall discuss his conception of the term in its special
 theological meaning.
 Epistemologically Averroes starts his discussion of the
 meaning of the term faith with the statement with which
 all Arabic works on logic usually begin, which in his version
 reads: "Instruction is of two kinds, namely, simple appre-
 hension (tasawwur) and belief (tasdiq), as the masters of
 the science of the Kalam have explained." He then
 proceeds to say that the methods of tasdiq are "demonstra-
 tive, dialectical and rhetorical."I23 Now we are going to
 show that in this threefold division of the methods of
 tasdiq Averroes means to say that tasdZq may refer either
 to belief in immediately known truths or to belief in demon-
 strated truths. In Arabic philosophic literature prior to
 Averroes, the term tasdzq is said to be of two kinds, the
 primary (awwaliyy) and the acquired (muktasib),I24 the
 former being a judgment which is "not preceded by another
 judgment upon which it is dependent"125 or which is "with-
 122 Kitdl Fasl al-Maqdl wal-TaqrIr md bain al-Shari'ah wal-Hikmah
 min al-Ittisal, in J. M. Miiller's Philosophie und Theologie von Averroes,
 Arabic: 1859; German: 1875; French by L. Gauthier, Accord de la
 Religion et de la Philosophie: Traite d'Ibn Rochd (Averroes), in Recuezl
 de Memozres et de textes publie en I'honneur du XIVe Congres des Orien-
 talistes par les professeurs de l'Ecole Supe'rieure des Lettres et des Meder-
 sas, Alger, 19C5, pp. 269-318; English by M. J. Rehman, The Philosophy
 and Theology of Averroes, Bardoa, 1921.
 23 Fasl, p. 19, 11. 10-12. This passage of Averroes is discussed more
 fully by the present writer in a paper on the terms taSawwur and tasdiq
 to be published in The Moslem World.
 24 Shahrastani, p. 349, 1. 1.
 25 Alfarabi, 'Uyun al-Masd'il, in Alfarabi's philosophische Ab-
 handlungen by F. Dieterici, p. 56, 1. 16.
 [137] 245
 THE JEWISH QUARTERLY REVIEW
 out investigation and research,"'26 the latter being a judg-
 ment "which one cannot comprehend without having
 comprehended beforehand some other things"127 or "which
 is obtained only by investigation."128 As illustrations of the
 primary kind of judgment, in Alfarabi and Algazali, men-
 tion is made of certain common notions in geometry or, in
 general, of primary premises in logical demonstrations,129
 and as illustrations of the acquired kind of judgment men-
 tion is made of several propositions which are arrived at
 only by demonstration as conclusions from primary pre-
 mises.130 Since, therefore, tasdiq is said in Arabic philosophy
 to be either primary or acquired, when Averroes says here
 that the methods of tasdzq are "demonstrative, dialectical
 and rhetorical," he undoubtedly means to say that tasdiq
 may be either primary or acquired. As to how this distinc-
 tion of primary and acquired is used by Averroes here with
 regard to tasdsq may be gathered from a further study of
 what he says subsequently about the rhetorical method of
 tasdiq as distinguished from the demonstrative and the
 dialectical method.
 The rhetorical method of tasdiq is said by Averroes to be
 a method which endeavors to establish belief by means of
 mathal and shibh.'I3 These two terms, I take it, represent
 respectively the Greek y'c,na?, maxim, and 7rapa6ely7/a,
 example,I32 which, according to Aristotle, are two of the
 three rhetorical methods of establishing faith (rlo'TeLS),
 the third being the enthymeme (evObvulJlua).133 Now both
 126 Algazali, Maqasid al-Faldsifah I, p. 5, 1. 4.
 27 Alfarabi, op. cit., p. 56, 1. 13. I28 Algazali, op. cit., p. 5, 1. 5.
 29 Alfarabi, op. cit., p. 56, 11. 17-18; Algazali, op. cit., p. 5, 11. 7-11.
 30 Alfarabi, op. cit., p. 56, 1. 14; Algazali, op. cit., p. 5, 11. 12-13.
 31I Cf. Fasl, p. 15, 11. 8-13, quoted below in n. 144.
 I32 Muller's "Bilder und Gleichnisse," Gauthier's "des figures et des
 symboles" and Rehman's "examples and parables" have all missed
 the technical use of these two Arabic terms here.
 133 Rhetorzca, II, 20, 1393a, 23-24.
 246 [138]
 DOUBLE FAITH THEORY-WOLFSON
 the maxim and the example, in contradistinction to the
 enthymeme, are considered by Aristotle as immediate
 kinds of knowledge. Maxims are described by him as
 "conclusions or premises of enthymes without the syllo-
 gisms"134 and as being generally known or agreed upon or
 self-evident.I35 They are in fact like the immediately known
 primary premises of demonstrative syllogisms, except that
 they deal "with objects of human actions, and with what
 should be chosen or avoided with reference to them.I36
 Similarly the example is defined by Aristotle as a "rhetorical
 induction,"'37 in contradistinction to the enthymeme
 which is a "rhetorical syllogism," for induction is considered
 by Aristotle as being the opposite of syllogism and demon-
 strationI38 and as a primary but indirect sort of know-
 ledge based upon our senses.139 When therefore Averroes
 says that the methods of tasdiq in the sense of belief are
 "demonstrative, dialectical and rhetorical," he means there-
 by that belief may be either 'acquired,' such as are obtained
 syllogistically by arguments, or 'primary,' such as are
 learned not syllogistically but rather directly by 'maxims'
 and indirectly by 'examples.'
 Belief or faith, then, in its general epistemological sense
 of the term, is used by Averroes in its two Aristotelian
 senses. In the first place, it is the judgment of the truth of
 undemonstrated knowledge, such as the truth of ethical
 maxims or generally known primary premises. In the
 second place, it is the judgment of the truth of conclusions
 which have been arrived at by demonstration. Like
 Clement, therefore, Averroes applies the term faith to
 '34 Ibid., II, 21, 1394a, 27-28.
 '35 Ibid., II, 21, 1394b, 10-15.
 I36 Ibid., II, 21, 1394a, 24-26.
 I37 Ibid., I, 2, 1356b, 4-5.
 138 Anal. Pr., II, 23, 68b, 13.
 '39 Anal. Post., I, 18, 81a, 38-81b, 9.
 [139] 247
 THE JEWISH QUARTERLY REVIEW
 scientific knowledge,i. e., to conclusions which follow by
 necessity from their premises. But he differs from Clement
 in his description of the nature of this kind of faith. Ac-
 cording to Clement, as we have seen, faith of any kind,
 even faith with reference to scientific knowledge, is an
 act of free will.140 Averroes, in contradistinction to this,
 explicitly says that "the belief which arises in the soul of
 the truth of a thing by reason of a proof is something
 compulsory and not voluntary, that is to say, we have not
 the power to consider that thing as untrue or as true in the
 same way as we have the power to rise or not to rise."'W4
 Coming now to his conception of faith in its theological
 sense, we find that Averroes starts with the common
 assumption of all religious philosophers ever since Philo,
 in asserting about the Koran what both Jews and Christians
 asserted about the Hebrew Scripture and what Christians
 asserted also about the Greek Scripture, namely, "that
 this our divine Law is true."142 Faith, then, in the religious
 sense, means the acceptance of the teachings of the Koran
 as true. Now as to the question whether one is to accept
 the truth of the teachings of the Koran without investiga-
 tion or whether one should try to demonstrate them by
 reason, his answer is like that of Clement and Saadia.
 Each man is to have a faith according to his capacities.
 "The natures of men," he says, "differ with reference to
 their conviction of the truth of a thing. One will be con-
 vinced of the truth of a thing through demonstration.
 4? Cf. above ? 2, text at n. 53 f.
 141 Facl, p. 13, 11. 18-20. Horten (Texte zu dem Strezte zwischen
 Glauben und Wzssen im Islam, p. 16) is not right, I think, in inferring
 from this statement that "faith," according to Averroes, is a voluntary
 act and hence essentially different from "reason." What this state-
 ment means is that, though scientific truth is not voluntary, still the
 term faith can be applied to it. See discussion on this point in St.
 Thomas, below ? 5.
 142 Ibid., p. 6, 11. 14-15.
 [140] 248
 DOUBLE FAITH THEORY-WOLFSON
 Another will become convinced of the truth as a result of
 dialectical arguments, just as the former becomes so by
 demonstration, and this because he is by his nature in-
 capable of more than that. Still another will become
 convinced of its truth as a result of rhetorical arguments,
 just as the master of demonstration becomes so by demon-
 strative arguments."I43 The same view is expressed by him
 elsewhere. "With regard to things," he says, "which on
 account of their abstruseness cannot be known except by
 demonstration, God has shown his favor to those of his
 servants who have no access to demonstration, either
 because of their natural predispositions or because of their
 habits or because of their lack of means to instruct them-
 selves, by making use for their benefit of maxims (amthal)
 and examples (ashbah) of these things, and with these
 maxims he has invited them to faith, for with these maxims
 one can attain faith by the use of arguments which are
 commonly accessible to all, that is to say, by dialectical
 and rhetorical arguments.I44
 In still another place, after mentioning the three methods
 of convincing people of the truth of a thing, namely, the
 demonstrative, the dialectical and the rhetorical methods,
 he says: "Inasmuch as not all men are by their natural
 dispositions capable of mastering the art of reasoning by
 arguments, -not even reasoning by dialectical arguments
 and still less so reasoning by demonstrative arguments, in
 addition to the intrinsic difficulty of the task of learning
 these demonstrative arguments and the length of time it
 '43 Ibid., p. 6, 11. 17-21.
 I44 Ibid., p. 15, 11. 8-13. Strictly speaking, "dialectical" arguments
 are not based upon "maxims" and "examples." But Averroes' coup-
 ling here of Rhetoric and Dialectic reflects Aristotle's statement
 that "Rhetoric is a counterpart of Dialectic; for both have to do with
 matters that are in a manner within the cognizance of all men and
 not confined to any special science" (Rhet., I, 1, 1354a, 1-3).
 249 [141]
 THE JEWISH QUARTERLY REVIEW
 must take those who are capable of learning them, and
 inasmuch also as the Law has no other purpose than to
 instruct all the people, the Law of necessity has to include
 all the various methods of affirmation or belief and all the
 various methods of simple apprehension."'45
 So far, then, we have in Averroes a double faith theory
 like that in Clement, in Origen and in Saadia. There is
 one truth underlying both forms of faith, the truth of the
 revealed teachings of the Koran. There is only a difference
 in the method by which different believers arrive at that
 truth. Some accept it implicitly and hence take the teach-
 ings of the Koran literally; others are supported in their
 acceptance of it by demonstration and hence understand
 some of those teachings as interpreted philosophically. To
 the former, the truth of the teachings of the Koran is like
 an indemonstrable primary premise or a maxim or an
 example; to the latter, it is like the conclusion of a syllogism.
 But as Averroes goes on in the development of his double
 faith theory he advances a view which is unique with him.
 The two conceptions of faith are not only defended and
 justified as true types of faith for those who by their
 capacities and opportunities happen to hold them, but they
 are set apart from each other as the only types of faith for
 their respective possessors. The simple believer is not only
 assured that his simple belief is good enough for him, but
 he is prohibited from trying to profess a belief as it is under-
 stood by philosophers, and the latter are not only allowed
 to rationalize about certain religious beliefs, but they are
 also prohibited from trying to profess such beliefs in the
 manner in which they are understood by the common
 people. In the case of such statements, he says, as can be
 understood either literally or rationally, as, e. g., the
 statements that God "proceeded to the heaven"I46 or that
 45 Fasl, p. 19,11. 13-17.
 250 [142]
 146 Surah, 2.27.
 DOUBLE FAITH THEORY-WOLFSON
 he comes down every night to the earth,I47 "for philosophers
 to take them according to their external literal meaning it
 is disbelief, and similarly for those who are not masters of
 demonstration to interpret them and to take them out of
 their literal sense it is accounted as disbelief or heresy.I48
 Moreover, not only must each particular class of believers
 keep to its own kind of belief, but it is also not allowed for
 philosophers to propagate their rational beliefs among the
 common people, or to expound them in popular books which
 may be read by the common people,149 and this prohibition
 applies even to philosophic interpretations which are true,
 not to mention those which are untrue.'50 He justifies this
 position of his by the analogy of medicine where, he says,
 it is commonly considered improper to discuss in popular
 work medical problems which might lead the common
 people to disregard the practical prescriptions of physicians
 for the preservation of health.Is5
 Now this is a new element introduced by Averroes into
 the double faith theory. When one considers the political
 conditions that existed in Moslem Spain under the Almohad
 rule during the time of Averroes one cannot help feeling
 that this element of his view was dictated by political
 necessity. But his essential view as to the relation between
 faith and reason, apart from the rigid boundary he has
 set up between the two camps of believers, is nothing but
 the old double faith theory, which, as we have shown, has its
 basis in the Aristotelian epistemological conception of
 faith.152
 47 Cf. M. J. Muller, Philosophie und Theologie von Averroes, German
 translation, p. 8, n. 1, for the source of this statement in Moslem tradition.
 148 Fasl, p. 16, 11. 2-4.