Thinking with Rosa: assent in philosophy of the Islamic world

ABSTRACT In Thinking with Assent: Renewing a Traditional Account of Knowledge and Belief, Maria Rosa Antognazza offers a historical narrative of pre-modern epistemology. She argues that until very recently, philosophers generally held that “knowing and believing are distinct in kind in the strong sense that they are mutually exclusive mental states”. This paper tests, and ultimately confirms, that account by applying it two thinkers of the Islamic world, al-Fārābī (d.950 CE) and Ibn Sīnā (Avicenna, d.1037 CE). It is shown that both of them used the term ‘assent (taṣdīq)’ as an umbrella term covering two very different states, knowledge and belief. In the case of Ibn Sīnā, this contrast is ultimately tied to his sharp distinction between immaterial intellective thinking and embodied thinking that uses a physical organ.

'renewing' promised in the title.The second section of the book develops and defends the traditional account in philosophical terms; the third section applies the insights of the first two to problems in religious epistemology.
My aim here will be much more modest.I simply want to complement (and compliment) the first section of Rosa's book by looking at a non-European tradition, namely philosophy in the Islamic world.Actually, what I will offer is less ambitious even than that, since I will have space only to treat two of the most famous of the falāsifa, that is, thinkers directly inspired by the Greek tradition: al-Fārābī (d.950 CE) and Ibn Sīnā (Avicenna, d.1037 CE).They will provide us with a test case for Rosa's claim that the 'traditional account' was indeed pervasive in philosophy up until early modernity.My brief investigation will ultimately ratify that claim, especially in the case of Ibn Sīnāa significant bit of confirmation, since Ibn Sīnā's stature and influence in the Islamic world was unparalleled.It is especially worthwhile to take up this particular case study, because al-Fārābī introduced, and Ibn Sīnā carried on, the practice of making the notion of 'assent' central to Arabic epistemology, just as it is central to Rosa's book.She uses 'assent' as the covering term for both knowledge and belief, and as we'll see, this too corresponds to what we find in the Islamic tradition.
Before we get to that, though, let's dwell for a moment on Rosa's reconstruction of the traditional account.As already mentioned, this account makes knowledge not a type of beliefe.g.justified true belief, or justified true belief with the addition of something furtherbut rather a sui generis mental state.Knowledge is grasping the "presence" of an object of knowledge, a "natural, primitive, effective contact between cognizer and cognized" (Antognazza, Thinking, §I.5).It can be, and in the tradition often was, compared to seeing.Rosa thus quotes Thomas Aquinas' remark, "The reason why the same thing cannot simultaneously and in the same respect be known and believed, is that what is known is seen whereas what is believed is not seen".1By contrast, 'belief' is reserved for less direct modes of cognition.Thus, says Rosa, "knowing is not 'the best kind of believing'; nor is believing to be understood derivatively from knowing as 'knowledge minus something'" (Antognazza, Thinking, Introduction).One implication of this is that knowledge is binary -'on/off' so to speakwhereas belief comes in degrees, the phenomenon modern-day epistemologists refer to as levels of 'credence'.
The upshot is that, whereas modern-day epistemology has most often (though not without exception) embraced a picture like this: the traditional account adopts the following analysis: Notice that in this second diagram, assent without knowledge is not designated as 'mere belief'. 2 Far from denigrating the epistemic status of belief, Rosa adopts the slogan 'knowledge first, but give belief its due'.She is convinced, and convinced that many pre-modern and early modern philosophers were convinced, that belief plays a crucially important role in our lives.It need not be epistemically deficient.Often, belief is in fact the best mental state we can reach.Rosa remarks that "a crucial part of our successful cognition is constituted by justified belief, which tracks truth if and when knowing is out of one's cognitive reach" (Antognazza, Thinking, §12.1.1).
To give an example that is prominent in her book, for many religious thinkers our grasp of God consists mostly or entirely of beliefs, not knowledge.In this life we do not enjoy God's 'presence' as we will in the beatific vision eagerly anticipated by such Christian thinkers as Aquinas.Another example might be 'knowledge by testimony', which on the traditional account is not in fact knowledge, unless testimony somehow puts us in contact with the objects at stake (Antognazza, Thinking, §5.3.2).Still, through testimony, we can get into a very favourable relation to the truth of a given propositionwe might be overwhelmingly 'justified' and even rightly take ourselves to be 'certain' that the proposition is true.It seems wrong, or at least misleading, to describe such cases as instances of 'mere' belief.In fact, as Rosa points out, the contrast between beliefs based on extremely reliable testimony and knowledge typically makes no practical difference (Antognazza, Thinking, §5.3.2).But they are still beliefs, and not knowledge, because they are assents that do not involve direct grasping.These considerations should forestall any suspicion that the above diagrams are in fact equivalent, with belief in general being re-named 'assent', and then subdivided into knowledgeable beliefs and mere beliefs.Instead, one may capture the contrast between highly certain, wellfounded beliefs and mere beliefs by integrating that contrast into the second diagram: Depending on one's epistemological predilictions, one could invoke 'warrant', or whatever is represented by the '+' in 'JTB+' models of knowledge, to spell out what a 'well-founded' belief is.But no addition will change belief into knowledge, because beliefs are a fundamentally different cognitive phenomenon from knowledge.It would be like trying to turn a cat into a bird by adding enough feathers.Now let us turn to the Islamic tradition, beginning with the author who, as already mentioned, was the first to make 'assent' a key term in Arabic epistemology.The word I am translating here as 'assent' is taṣ dīq, which literally means 'deeming true'.This is the standard translation in the secondary literature, though other renderings have been offered, including 'belief'; we'll come back to this.Taṣ dīq forms a pair along with taṣ awwur, which is usually translated as 'conception' (for the contrast see Butterworth, "À propos"; Lameer, Conception and Belief; Maróth, "Taṣ awwur and Taṣ dīq"; Wolfson, "Taṣ awwur and Taṣ dīq").This word comes from the verbal root used for 'forming', as of a picture or representation; thus the Greek word eidos, 'form', was translated into Arabic as ṣ ūra, which has the same root.
Conception is grasping the meaning of a term, phrase, or proposition, but without commitment to truth.Most often, the object of conception is a single term, in which case the question of truth would not arise anyway.By contrast assent is committing to the truth of a proposition.To put it in other, equally sketchy, terms, conception is to understand what something is or what something means, while assent is to think that something is the case.Here is a more technical explanation from al-Fārābī: Knowledge (maʿrifa) is of two types: conception (taṣ awwur) and assent (taṣ dīq).Each of these may be either complete or deficient … Complete assent is the same as certainty (huwa al-yaqīn), while complete conception is conceptualizing something in a way that encapsulates the thing's essence.Namely that the thing be conceptualized in terms of what is signified by its definition.Now, we can build on these two [ideas] to explain what precisely we mean by complete assent.We say that assent in general is when a human judges (yaʿtaqidu) that something is the case and accordingly judges that the existence of that thing outside the mind is in accord with what is held in the mind.Truth, then, is when the thing outside the mind in fact is in accord with what is held in the mind … and certainty is when we judge about that to which the assent applies, that it cannot at all be otherwise than we judge it to be. 3  (Burhān, 19-20)   There's a lot to unpack here, principally the relationship of conception and assent to knowledge, and the concept of complete and incomplete conception and assent.
Starting with the former, it is immediately striking that al-Fārābī makes conception and assent types of knowledge.This already is a hint that Rosa's 'traditional account' may be playing a role.Obviously if beliefs fall under this scheme at all, they should appear on the assent side, since believing something is taking it to be true.The fact that some cases of knowledge are 'conceptions' thus shows that whatever knowledge is, in general, it is not just a species of belief.On the other hand, when al-Fārābī introduces conception and assent as types of knowledge, he does not mean to suggest that every case of assent is knowledge.This will be clear as we go on, but it is in any case quite intuitive: people frequently assent without knowing, as when they assent to false propositions.Rather, al-Fārābī's point is that all conceptions, and some assents, count as knowledge: The reader is advised not to put any weight on the phrase 'epistemic states', which is not taken from al-Fārābī; I have introduced it as a purposefully vague umbrella term that can cover the very different phenomena of conception and assent.The reader is, by contrast, advised to put plenty of weight on the lack of any contrast between knowledge and non-knowledge within the category of conception.This is because to conceptualize something is to know it, at least partially.Ibn Sīnā will later say that something "may be unknown by way of conception" (Ibn Sīnā, Ishārāt, 41), but by this he does not mean having a false conception.Rather, it means simply lacking any conception of that thing.Hence he adds that in such a case, "its meaning is not conceived until one learns [other] concepts" (Ibn Sīnā,Ishārāt,41).
The possibility of merely partial conception takes us to the second point about completeness and its lack.By 'complete conception' al-Fārābī means being in possession of a real definition, as when one conceptualizes human as rational mortal animal.An 'incomplete conception' would be either a partial definition, as when one conceptualizes human only as animal, or a merely nominal definition (one that mentions accidental features), as when one conceptualizes human as animal capable of laughing.The fact that conceptualization, whether complete or partial, counts as knowledge is another sign that al-Fārābī is working within Rosa's 'traditional account'.It looks very much like 'knowing by way of conception' is just grasping something, not unlike seeing.If there is no guarantee that the grasp will be complete, then that too is like seeing, since you can see something only in part.Again, all these points can apply to something more complex than a single term.If one is merely entertaining a proposition, that would involve conceptualizing it ('knowing what it means'), without assenting to it or to its negation.Though I do not know of a text on the issue, I assume that one could also 'incompletely' conceptualize a whole proposition, for instance if one were entertaining a proposition about humans and had only a nominal definition of human, or understood only one of the terms in the proposition.
What about complete and incomplete assent?In the passage quoted above, al-Fārābī tells us that assent in general is judging things outside the mind to be as they are in the mind.Here he is endorsing the sort of 'correspondence theory' of truth often ascribed to Aristotle; interestingly, he does so in terms of the matching of thoughts to the world, not the matching of statements to the world, as suggested by Aristotle at Metaphysics 1011b ("saying of what is that it is, and of what is not that it is not, is true"; see further Crivelli, Aristotle).This could be explained by saying that al-Fārābī is adjusting the definition of truth in light of Aristotle's doctrine in On Interpretation that words represent thoughts, and thoughts represent things (I take inspiration for this suggestion from David, "Correspondence Theory").In any case, believing that one's mental representation corresponds to the way things are is not enough for 'complete' assent.Also required is that the correspondence in fact obtains.In other words, that to which one assents must be true.Finally, it must be true in such a way that things could not be otherwise.
For further light on this last constraint, we should turn to another work by al-Fārābī, which has been well studied by Deborah Black, called On the Conditions of Certainty (Sharāʾit al-yaqīn, edited in the same volume as Burhān; see Black, "Knowledge and Certitude").He names no fewer than six conditions that need to be satisfied to reach complete or "absolute (ʿalā l-iṭlāq)" certainty, which partially overlap with what we have just seen (Sharāʾit 98): (1) It must be "judged of something that it is such-and-such".
(2) It must "suitably occur (yuwāfiqu) that it [sc.the judgement] corresponds to, rather than opposing, the existence of the thing outside [the mind]".(3) It must be "known that it corresponds".(4) It must be impossible that it not correspond.
(5) There is no time at which it does not correspond.( 6) All this happens essentially, not accidentally.
Is al-Fārābī effectively saying here that certain knowledge is a kind of belief, one that has a set of additional features?If so, he would be departing from the 'traditional account'.To answer this question we need to take a closer look at conditions 1-3, which Black calls respectively the belief, truth, and knowledge conditions. 4 On her reading, the first condition may be paraphrased as 'S believes that P' ("Knowledge and Certitude", 16), which seems fair enough.But as she points out, al-Fārābī is not using the Arabic term that would normally correspond to the Greek doxa, meaning 'mere belief'; that would be z ann, on which more shortly.Instead, he uses the rather generic term iʿtiqād, which I have rendered 'judgement'.He immediately goes on to offer two alternatives, raʾy ('opinion') and ijmāʿ ('consensus'), as alternative vocabulary to make the idea clear to his reader.But it is not only the verb for 'believing' that is important here, it is what al-Fārābī says about the content of the belief.Namely that one is judging 'of something that it is such-and-such', in other words, judging a predicate to hold of a subject.The fact that al-Fārābī is only interested in predicative judgements is hardly a surprise, since this is the type of proposition at stake in Aristotle's logic.We might then prefer to paraphrase the first condition as requiring that 'S judges that A holds true of B'.In other words, condition 1 requires that there is taṣ dīq, 'assent'.
As for condition 2, it simply affirms Aristotle's correspondence theory of truth.As Black suggests, we can take this to be an 'externalist' constraint on absolute certainty; to judge that animal holds true of human with absolute certainty requires that animal does in fact hold true of human in the external world.By contrast, condition 3 identifies an 'internalist' constraint on the belief.What al-Fārābī says about this condition is interesting: it is intended to rule out cases in which the person making the judgement "is not aware that what is judged [to be the case] corresponds; rather, as far as he is concerned (ʿindahu) it might be (ʿasā) that it does not correspond".As al-Fārābī goes on to say, the person in possession of absolute certainty knows that their judgement is 4 My thanks to Fedor Benevich and Abdurrahman Mihirig for helpful discussion of the passage.
not a "mere belief (z ann, again, the corresponding term to Greek doxa)", that is, a belief that might be either true or false.So what is being excluded here is belief of a very specific kind: beliefs in which the person with the belief is explicitly aware that the belief may not correspond to the way things are.
The kind of assent that interests al-Fārābī here is very different, as we can see from conditions 4, 5, and 6, which effectively spell out the meaning of "could not be otherwise" in the passage quoted earlier.These conditions are inspired by Aristotle's constraints on knowledge in the strict and proper sense (that is, demonstrative 'understanding', the translation urged by Burnyeat, "Aristotle"), which also require that such knowledge concern itself with eternal, necessary, and essential truths.So there is no sign here that al-Fārābī imagines us to be in possession of a mere belief (z ann) that could be turned into certain knowledge (ʿilm yaqīnī ) by adding something further, like justification.Rather īʿtiqād ('judgement'), which I take to be tantamount to ṭ aṣ dīq ('assent'), is an umbrella term covering cases of certain knowledge and avowedly uncertain belief.Black too denies that al-Fārābī's account can be assimilated to the JTB model, and as if she had already read Rosa's book almost twenty years ago, comments that for al-Fārābī "knowledge, like vision, requires direct epistemic contact with the object known at the time when it is occurring.And it is that direct relation to the object of one's belief that must be present to guarantee certitude" ("Knowledge and Certitude", 16, 22). 5  In keeping with all this, conditions 4-6 suggest that for al-Fārābī, as for Aristotle, knowledge in the strict and proper sense deals with a highly restricted range of propositions.For, according to these conditions, knowledge in the proper sense has to be about eternal, necessary, essential truths.Other cases of assent simply cannot be instances of "absolutely certain" knowledge, regardless how ideal our epistemic state may be.These would include assents to propositions about contingent and transient matters, like 'Socrates is in the marketplace'.Al-Fārābī does allow that one could have "incomplete certainty" about such propositions, and thus know them to be true in a less demanding sense.Just as one can have incomplete knowledge on the side of conception, for instance through a nominal definition, one can have incomplete knowledge on the side of assent.This 5 I have a slight disagreement with Black on the question of whether what is known is propositional.She notes, as I did, that condition 1 seems to be envisioning propositional judgements, indeed predications.She is worried though that his talk of 'awareness' suggests something more like 'acquaintance' with an object, which would be in tension with the propositional account.But I would take this to be simply a sign of al-Fārābī's underlying assumption that Aristotelian syllogistic is entirely compatible with a 'knowledge-as-seeing' account: what one 'sees' when grasping human is, for instance, that animal is predicated of human.This is not to deny that there are difficulties in bringing together the epistemology of the Posterior Analytics with that of the De Anima, only that al-Fārābī was surely convinced that it is possible.
would be non-scientific assent, in other words, assent that does not satisfy the demands of the Posterior Analytics.
The upshot is that, though some scholars do translate taṣ dīq as 'belief' (as in the very title of Lameer, Conception and Belief), it would be truer to al-Fārābī's presentation to fill out our earlier chart as follows: As Rosa liked to say, though, "call them as you like".We should not insist dogmatically on the terminology, as long as the philosophical point is clear.Namely that for al-Fārābī, assent is (as Rosa's historical account would predict) an umbrella term for two fundamentally different and mutually exclusive epistemic states.
While one could further explore al-Fārābī's views on the difference between these two states, at this point it will be more fruitful to turn to the more elaborate epistemology and philosophy of mind offered by Ibn Sīnā.He takes over from al-Fārābī the contrast between taṣ dīq and taṣ awwur, agreeing that these are the two varieties of knowledge (e.g. at Ibn Sīnā, Madkhal 30; Burhān 51).So far I have been translating two Arabic words indifferently as 'knowledge': ʿilm and maʿrifa.They have no systematically different connotations, though ʿilm was the usual rendering of the Greek epistêmê and thus had more tendency to connote the concept of properly scientific knowledge.Sometimes maʿrifa means something more like understanding or awareness.But in his masterwork the Healing, within the section corresponding to the Posterior Analytics, Ibn Sīnā stipulates an unusual meaning for the two words, using ʿilm for the universal knowledge that qualifies as fully demonstrative and maʿrifa for the subsuming of a particular under such universal knowledge (Ibn Sīnā, Burhān 283; Adamson, "On Knowledge", 283).To use an example already given by Aristotle, it would count as ʿilm to know that all triangles have internal angles equal to two right angles, and maʿrifa to know that this particular triangle has internal angles equal to two right angles.While this terminological stipulation is artificial and unusual, even in Ibn Sīnā's own writings never mind Arabic philosophical works more generally, it does show him making the same sort of contrast we have seen in al-Fārābī.ʿIlm corresponds to al-Fārābī's 'absolutely' or completely certain knowledge, and maʿrifa to a kind of incompletely certain knowledge.
For Ibn Sīnā universal, necessary, and essential knowledge is the best form of assent, and is a kind of certain assent (Ibn Sīnā, Burhān 51; Black, "Certitude, Justification", 122; Strobino, Avicenna's Theory, 41). 6He pursues the topic of certainty in a way that will look familiar from what we saw in al-Fārābī: (1) Certain assent comes with a second-order judgement (iʿtiqād) that it must be true.(2) A first kind of uncertain assent comes with no second-order judgement one way or the other as to whether it must be true.(3) A second kind of uncertain assent comes with a second-order judgement that the assent could be false.
The first kind of assent is used in demonstrative science, the second in dialectic and sophistical arguments, and the third in rhetoric.So this classification assent types gives us a rationale for the range of argument forms considered in Aristotle's Organon. 7et's consider the three types in turn, starting with the assent that must be true (cf.al-Fārābī's condition 4).In the spirit of Aristotle's constraint that demonstrative knowledge is of eternal truths (cf.al-Fārābī's condition 5), Ibn Sīnā elsewhere specifies that in the first kind of assent, the proposition cannot cease to be true (lā yumkinu zawāluhu, at Burhān 256; see Strobino,Avicenna's Theory,44).Notice that this is stronger than saying that the proposition does not cease to be true.Ibn Sīnā thinks that some propositions are always true, but only contingently so, for example the proposition 'the universe exists'.The universe is eternally made to exist by an extrinsic cause, namely God, and it is only this causal relation that renders the proposition true, when in itself it might have been false.Assent to this proposition is therefore of the third type: if one forms a secondorder judgement about it, this should be that it need not be true.What underwrites the modal features of these sorts of assent is Ibn Sīnā's essentialism, according to which each essence has certain features that are conceptually "inseparable" from it (Ibn Sīnā, Ishārāt 46; see further Benevich, Essentialität; Benevich, "Avicennan Essentialism"; Strobino, "Per Se, Inseparability").Animal and rational are inseparable from human, whereas existent is not inseparable from the universe.Thus Ibn Sīnā would also agree with al-Fārābī's condition 6 (essentiality).Indeed this is the really important condition, of which conditions 4 (necessity) and 5 (eternity) turn out to be mere corollaries.
As for Ibn Sīnā's second kind of assent, the kind used in dialectic and sophistry, its status relies on the merely psychological fact that the person making the assent has not considered its modal status.Ibn Sīnā says explicitly that if the person were to do so, they would judge either that the assent must be true, turning it into an assent of the first type; or that, while true, it could be false, turning it into an assent of the third type.To flesh out what is going on here, we might imagine a dialectical disputation where a premise is introduced on the grounds that it is 'widely accepted' (endoxon in Aristotle's Greek, mashhūr in Ibn Sīnā's Arabic), e.g.'pleasure is good'.If the premise is granted (musallam), then the argument can proceed without either party stopping to consider whether pleasure must be good (which is tantamount to the question of whether goodness is essential to pleasure), or might after all fail to be good.This is entirely consistent with the character of Aristotelian dialectic, where arguments are pursued without being grounded in scientific first principles and thereby established as essential and necessary truths.Instead, premises are simply conceded according to certain rules.Something similar happens in sophistical argumentation, except that the victim of the sophistry is being caught unawares, tricked into granting something they should not, instead of deliberately conceding a premise for argument's sake as in dialectic.
Finally, the third kind of assent occurs when one is fully aware that one's assent could be false.This might look strange at first glance: 'I assent to P, while cheerfully admitting that P might not be true'.But in fact most of our assenting is of this kind, since it covers all propositions about the accidental features of things.After all, an accidental feature is precisely a feature that could be separated from its bearer.Significantly, and in contrast to what we found in al-Fārābī, Ibn Sīnā does use the word z ann for this kind of assent, that is, the word that corresponds to the Greek doxa.So we can update our chart to reflect Ibn Sīnā's vocabulary, as follows: Where knowledge grasps propositions that are necessary, because they are essential, and belief grasps propositions that are at best contingently true (and at worse, just false).
One might wonder why scientific knowledge as applied to particulars (maʿrifa) falls under knowledge and not belief, since particulars are contingent.For instance, why would 'Socrates is rational' not qualify as an assent of the third type, just as much as 'Socrates is in the marketplace'?After all, it is not eternally the case that Socrates is rational, given that Socrates is not eternal.The full answer to this question is a bit complicated, but it boils down to the following.When one has scientific knowledge that human is (essentially) rational, one thereby knows that every particular human is (essentially) rational.So a necessary and eternally true proposition is still in play here, and it is what is known in the strict and proper sense; this knowledge is then deployed in the case of Socrates simply by recognizing that Socrates is a member of the kind human, that is, a particular falling under the relevant universal, and applying to him an attribute that is inseparable from human.
This connects to a notorious, and for our purposes highly relevant, discussion in Ibn Sīnā, concerning the question whether God knows particulars.As I have argued elsewhere (Adamson, "On Knowledge"), Ibn Sīnā's provocative response to this questionnamely that God only knows them 'in a universal way'is intimately related with what we've just seen.As a purely intellective being, God only grasps universals as such, but He can also know about particulars insofar as they are subsumed under those universals.It is only through His remote causal influence over them that He knows them to fall under just these universals.For example, because He is the ultimate cause of Socrates' existence, by knowing Himself God can know indirectly that Socrates is an existing human.Furthermore, He has the universal knowledge that human is rational, so He can know that Socrates is rational.Humans are in a different situation.We are possessed of sense-perception, which gives us direct access to particulars as such.So we can grasp Socrates as a particular human, and then straightforwardly apply to him any attributes that belong to the universals under which he falls.(Recent studies of this much-discussed topic include Nusseibeh, "Avicenna"; Lim, "God's Knowledge"; Kaukua, "Future Contingency"; Zadyousefi, "Adamson, Avicenna".) The reason I am going into this in such detail is that it is precisely in view of this faultline that Ibn Sīnā embraces the traditional account as laid out by Rosa.For Ibn Sīnā, knowledge in the strict and proper sense is assent of the first type, either in the form of demonstrative propositions (this sort of knowledge he calls ʿilm) or in cases where we apply such propositions to particular cases (this sort of knowledge he calls maʿrifa).Other kinds of assent concern themselves with non-scientific propositions, with contingent matters.Like Rosa, Ibn Sīnā wants to 'give belief its due', precisely because the vast majority of assents we make in everyday life fall into the latter category: Socrates is in the marketplace, that giraffe over there is tall, this cake would be delicious to eat.Ibn Sīnā even allows that we can have certainty about such assents and that we make them 'by necessity', not in the sense that the propositions at stake are in themselves necessary, but in the sense that we cannot help assenting.This would apply, for instance, to assenting on the grounds of sufficiently strong testimony, as well as assents that are based on sensation.
So if I see Socrates in the marketplace, or get powerful testimony that he is there at the moment, I may take myself to be certain that he is in the marketplace.I would make a second-order judgement that I cannot be wrong about this, even though I could never have scientific understanding of his being in the marketplace, since this is not necessary, essential, eternal, or universal, and in this sense lack 'knowledge' of it.A further situation arises when I simply grant something for the sake of argument, as in a dialectical encounter,8 or have a belief while lacking certainty.As Black notes, Ibn Sīnā will sometimes use the word z ann ('mere belief') for that latter category.Taking all this into account, we can now give our chart of epistemic states its final form: One reason to insist on a fundamental divide between knowledge and belief in Ibn Sīnā's theory of assent is that this divide is reflected in his philosophy of mind.According to him, we have only one psychological power that is not exercised through a bodily organ: intellect (ʿaql).Here we have a metaphysical basis for Ibn Sīnā's adherence to the traditional account, one he shares in common with many, if not most, other pre-modern thinkers who accept this account.Knowledge and belief are mutually exclusive mental states not just for the epistemic reasons we've been treating so far, but also because they are states of very different psychological powers.Beliefs about particulars always involve the brain, which is the seat of such powers as sensation, memory, and the imaginationall the powers that deal with particulars.A particularly important brain-centred power for our purposes is 'thinking (fikr)', which is the power that would be engaged when we are forming beliefs about particulars.While one function of 'thinking' is to prepare the way for the full-blown knowledge achieved by the intellect (Gutas, "Intuition and Thinking"), that kind of knowledge cannot be realized in the brain or in any other bodily organ.This is because it is universal in character, and Ibn Sīnā thinks that reception of a universal intelligible in a bodily organ would 'particularize' it, rendering it no longer suitable for scientific understanding.Indeed, this is the basis of Ibn Sīnā's proof for the immateriality of the rational soul.Because the human intellect grasps universal 'intelligibles (maʿqūlāt)', it must operate without using an organ, so it can survive and continue to enjoy knowledge after the death of the body (Alpina, Subject, Chapter 5; Adamson, "From Known").
Ibn Sīnā is aware that he is departing from Aristotle in some respects.He presents his own proof of the soul's immaterial activity as an improvement on Aristotle's, and criticizes Aristotle for suggesting that the soul is only the 'form of the body', since this would apply to the soul only insofar as it is the source of the body's perfection, not insofar as it has an intellective function in its own right.Still, Ibn Sīnā is basing himself throughout on originally Aristotelian premises, above all the rigorous demands placed on knowledge in the Posterior Analytics (see further Strobino, "Avicenna's Use"), which are also a driving consideration in al-Fārābī's epistemology.
In light of which, it is no surprise that what Rosa says about Aristotle in Thinking with Assent turns out to be directly relevant to these thinkers of the Islamic world.She writes: Judgement or taking-to-be-true (hupolēpsis) is, in turn, the generic cognitive mode under which, in De Anima, Aristotle groups more specific cognitive modes, including epistēmē and doxa … Epistēmē is not a matter of having justified true beliefs.If anything, as a grasp of demonstrations or as an explanatory capacity, epistēmē is closer to understanding than to justified true belief.
(Thinking with Assent, Section 1.2.ii) As Joep Lameer has shown (Lameer,Conception and Belief,(18)(19)(20)(21)(22)(23), in coining the term taṣ dīq al-Fārābī was thinking of a passage at Posterior Analytics 1.1, with the original Greek term being hupolambanein.9This might encourage us to modify Rosa's remark about hupolēpsis by inserting Arabic terminology in place of Greek, which would yield the following result: Judgement or taking-to-be-true (taṣ dīq) is a generic cognitive mode under which fall more specific cognitive modes, including ʿilm and z ann … ʿIlm is not a matter of having justified true beliefs.If anything, as a grasp of demonstrations it is closer to understanding than to justified true belief.
Which captures remarkably well what we have found in Ibn Sīnā.
Having said that, there is not a perfect fit between Rosa's own epistemology and that of Ibn Sīnā.As we saw, for Rosa sensation and perhaps testimony can give rise to knowledge in the strict and proper sense.So Rosa has a much more liberal understanding of 'knowledge' than the strict, intellective account of knowledge in Ibn Sīnā.She would apply the term to any primitive grasping of something that is 'present' to an epistemic subject.If we consult the final chart given above, this means that Ibn Sīnā's 'certain beliefs', or at least some of them, would count as knowledge in Rosa's terms.The difference between them is the result of significant commitments on Ibn Sīnā's part that Rosa does not share: in epistemology, his essentialism and its attendant modal theory, and his highly restrictive approach to 'knowledge' in the strict and proper sense, which is an inheritance from Aristotle; in philosophy of mind, the stark contrast he draws between an immaterially realized power for grasping universal intelligibles, and an embodied power that deals with particulars.None of this means, though, that Ibn Sīnā fails to exemplify Rosa's historical account.It is one thing to claim that knowledge and belief are mutually exclusive epistemic states, and another to say what belongs on which side of the ledger and why. 10s it happens, subsequent thinkers of the Islamic world challenged Ibn Sīnā's epistemology, and devised alternatives that come closer to the position adopted by Rosa.Two developments were especially relevant.First, a number of thinkersespecially Abū l-Barakāt al-Baghdādī (d. in the 1160s) and Fakhr al-Dīn al-Rāzīrejected wholesale the Aristotelian and Avicennan psychological theory according to which separate 'faculties' or 'powers' (Gk.dunameis, Ar. quwan) are the subjects of cognition.Rather, there is just the single 'self' that performs all cognition (Kaukua, "Self, Agent, Soul"; Tiryaki, "From Faculties"; Adamson and Benevich, "Fakhr al-Dīn").Against Ibn Sīnā's position, Fakhr al-Dīn al-Rāzī invoked one of the phenomena we have discussed above: the subsuming of a particular under a universal judgement.In such a case, Ibn Sīnā would say that brain-seated powers like sensation and memory, which grasp particulars, are working in tandem with the immaterial intellect, which grasps universals.Thus if we perceive that Socrates is a human, we can deploy the intellective knowledge (ʿilm) that human is rational to derive the knowledge (maʿrifa) that Socrates is rational.Fakhr al-Dīn argues that the real 'judger' or 'knower' for this latter proposition must just be a soul or self that can grasp both particulars and universals.How else would the two kinds of cognition be brought together?In one fell swoop, this plausible move undercuts Ibn Sīnā's bifurcated epistemology.
In a second development, some philosophersespecially those in the 'Illuminationist' (ishrāqī) tradition inaugurated by Suhrawardī (d.1191)stressed the analogy between directly knowing something and seeing something.In terms that are, again, strikingly similar to those used by Rosa, these Illuminationists spoke of a form of cognition they called 'knowledge by presence (h  ud  ūr)' (Kaukua, "Suhrawardī's Knowledge").They tend to define this sort of cognition negatively, as the absence of any impediment to grasping the cognized object.Just as we can see something so long as nothing is blocking our view, so the human mind will know its object as long as that object is 'present' to it and nothing prevents the cognition (e.g.interference from the body).This is also how we are aware of the contents of our own minds (including our beliefs) and our own bodies (Eichner, "Knowledge"), which are simply present to us.It is in the same way that God does after all know particulars as such, since nothing is hidden from Him, or ever could be (Benevich, "God's Knowledge").This second challenge to Ibn Sīnā even more obviously shifts the discussion in Rosa's direction.Now the phenomenology of unhindered sensation is being used as a paradigm of knowledge, doing away with the Avicennan stricture that knowledge in the proper sense has to do with universal, intelligible objects of cognition.
One could certainly extend this investigation by adding other thinkers of the Islamic world to Rosa's historical panorama.Figures like Fakhr al-Dīn al-Rāzī were critiquing Ibn Sīnā in light of epistemological ideas taken from the tradition of Islamic theology (kalām), which defined knowledge as a 'relation' between knower and known rather than the forming of a mental representation that corresponds to the way things are.This already suggests that the kalām theory maintained a mutually exclusive opposition between knowledge and belief, since knowledge is a sui generis relation on this account; but more discussion would be needed to establish that.On the Aristotelian side of the story, Ibn Rushd (Averroes, d. 1198) shared with Ibn Sīnā a wholehearted commitment to the epistemic strictures of the Posterior Analytics and its association of knowledge with universal, necessary truths.Indeed, this underlay Ibn Rushd's notorious claim that there is only one potential intellect for all of humankind.An intellect belonging to just one human would be embodied and thus particularized, therefore unable to perform the requisite kind of cognition.Furthermore, there can be only one act of grasping a given intelligible.As Stephen Ogden puts it, "the best way to explain how we can all think the same thing is that there is only one and the same thing that is thoughtin one intellect" (Ogden,Averroes on Intellect,109).One could hardly ask for a more dramatic instance of Rosa's historical thesis that knowledge and belief are mutually exclusive mental states: here knowledge as described in the Posterior Analytics is a cognitive state in an immaterial, universal intellect, while other kinds of thinking are realized in the brains of individual humans (Taylor, "Remarks").As these examples show, the widereaching and ambitious claim made in the first part of Rosa's Thinking About Assent is not just a historical thesis about the many philosophers she discusses, but also a hypothesis that can be tested on other philosophers whom she does not discuss.If the foregoing analysis is correct, then it is a hypothesis that is even more powerful and illuminating than she supposed.

Disclosure statement
No potential conflict of interest was reported by the author(s).