131 - Deborah Black on al-Fārābī's Epistemology
Deborah Black joins Peter to talk about al-Farabi's innovations concerning knowledge and certainty.
Themes:
• D.L. Black, Logic and Aristotle’s “Rhetoric” and “Poetics” in Medieval Arabic Philosophy (Leiden: 1990).
• D.L. Black, “Al-Fārābī,” A Companion to Philosophy in the Middle Ages, ed. J. Gracia and T. Noone (Malden, MA: 2003), 109–117.
• D.L. Black, “Knowledge (ʿilm) and Certitude (yaqīn) in al-Fārābī’s Epistemology,” Arabic Sciences and Philosophy 16 (2006), 11-46.
• D.L. Black, “Al-Fārābī on Meno’s Paradox,” in P. Adamson (ed.), In the Age of al-Fārābī: Arabic Philosophy in the Fourth/Tenth Century (London: 2008), 15-34.
Note: this transcription was produced by automatic voice recognition software. It has been corrected by hand, but may still contain errors. We are very grateful to Tim Wittenborg for his production of the automated transcripts and for the efforts of a team of volunteer listeners who corrected the texts.
PA: So we're going to be talking about al-Fārābī's views about knowledge. Maybe we'll get on to talking about Avicenna a little bit later because they have a lot in common. But before we start talking about what Fārābī actually says about knowledge and what it is, how we can get it, I thought maybe we should begin by talking about the sources that al-Fārābī draws on in developing his own ideas in epistemology.
DB: Okay. I also thought maybe I should say something a bit about the fact that we don't have an explicit treatise on, well, maybe that's a bit of an overstatement, but we don't really have an explicit notion of epistemology in Fārābī or in any of the Islamic Aristotelians. So most of what he's saying that we would classify as "epistemological" comes in the context of elaborating on Aristotle's theory of demonstration in the Posterior Analytics. So that kind of segues as an answer to your question about sources because that really is the main source for Fārābī - Aristotle's Posterior Analytics in the Arabic translation. And so epistemology is primarily a part of logic for Fārābī and his successors and to a lesser extent a part of psychology because some of the discussion comes up in the context of dealing with the cognitive psychology in Aristotle's De Anima. So those are the two main sources.
The other sources are related to those. So one of the interesting features of Fārābī's epistemology is that in addition to a theory of demonstrative science, so really knowledge in the highest form, he also addresses the status of lesser forms of knowledge and these are associated with other logical arts like dialectic, rhetoric, and poetics. And so Aristotle's treatises on those topics are also major sources.
Another important source, and I kind of discovered this myself by accident, closely related to the Posterior Analytics, is the paraphrase of one of the Greek commentators, Themistius from the 4th century. And the reason I discovered this is when both Fārābī and Avicenna discuss Meno's paradox, they keep talking about a runaway slave. So instead of Socrates questioning a slave boy about the principles of geometry, they have a puzzle about how you recognize the runaway slave if you don't know who he is. I stumbled upon that in Themistius, and so once I discovered that, I realized they are also reading a lot of Aristotle through Themistius.
PA: That's interesting because usually people think of al-Fārābī, maybe already al-Kindī, but al-Fārābī, Avicenna, as reading Aristotle through the lens of neoplatonist commentators on Aristotle. And so they say, oh, that's why we get this fusion of Neoplatonism and Aristotelianism. But you're saying that at least when it comes to the Posterior Analytics, one of the big influences on them from late antiquity would actually be Themistius, who is maybe some kind of Platonist, but certainly not a full-blown neoplatonist like, say, Philoponus or Simplicius.
DB: Exactly. And I think in relation to that, too, unless you're talking about the cognitive psychology side of this, there's a sense in which there are more properly Platonic elements, particularly in Fārābī, because this whole attempt to classify different forms of knowledge relative to the ideal of demonstration is in part cashed out in terms of what Fārābī knows about Plato's epistemology from the versions of the Republic that they have and the sort of hierarchical view of different forms of knowledge that you find in Book 6 of the Republic and other texts of that sort. And so even though Fārābī doesn't know a lot of Plato's dialogues directly or doesn't seem to - so the Meno example sort of gives you a flavor of that - if you read his philosophy of Plato, it's clear he knows a lot about the dialogues, but he doesn't really fully know their content. So I think there's a lot of Platonism in the sense of a very strong hierarchical ordering of the lesser modes of cognition to demonstrative science. Now, that's in Aristotle to some extent, but it's really emphasized by Plato. And in one of his rhetorical commentaries, there's this wonderful metaphor of rhetoric, poetics and dialectic kind of being the means by which you climb out of Plato's cave in the Republic. So I think that's a very strong flavor in his approach.
PA: So using Aristotelian logic to escape from Plato's cave, basically.
DB: Exactly. And if you're a prophet or a politician, you have to go back down again, just as in Plato. 5.33
PA: Actually, that mention of prophecy raises another question, though, which is that, I mean, you've been focusing so far just on the Greek sources that Fārābī is using. And you did call him an Aristotelian before, and I guess everyone would agree with that characterization of him. Do you think that Islamic sources fundamentally change his way of thinking about knowledge, or is he really just trying to develop a theory of knowledge by engaging with Aristotle and other Greek works in Arabic translation?
DB: I think Fārābī, unlike Avicenna later, who I do think in particular makes creative use of some of the traditions of the Mu'tazilite theologians, so the earliest rationalist school of theology, Fārābī is indebted to that background in his tradition, but in a far more negative or adversarial sense. So I think that, I mean, I'm not an expert in kalām, but it is clear from what I know of the treatises of kalām or theology, they always start with epistemological issues, so they are always concerned with the sources of knowledge, they give classifications and so on. They sort of treat metaphysical issues through the lens of epistemology. So I think the very fact that Fārābī focuses on epistemology as an important topic in its own right, and even, you know, does have one short treatise that is just on epistemological issues, I think that is a development that comes up from his rivalry with the Mu'tazilite theologians. So he thinks that they are wrong on almost everything and that their epistemology is wrong, so he has got to develop a more clearly perapetetic sort of epistemology to address that.
PA: They are so wrong about everything that they are even wrong about what it would mean to be right about something.
DB: Exactly, and whenever Fārābī is classifying the different grades of assent and the different types of knowledge, he always gets in a dig at the Mu'tazilites that they think they are getting knowledge and they think they are using the intellect, but it is just rhetoric and poetics. And in particular, he always rails against the description that the Mu'tazilites give for the goal of thinking, which they describe as an acquiescence of the soul, sukūn an-nafs, which is basically the idea that you come to assent to something, so that is their term for assent. And Fārābī always says they mistake psychological certitude - it feels like we're certain - for true epistemic understanding. And so he says their acquiescence of the soul is really rhetorical assent. But just having to deal with those kinds of issues and think about what exactly assent is and what certitude is and so on, I think he is forced to do that because of addressing those rivals. 8.30
PA: So the objection here, I guess, then would be that just because you feel like you are certain about something doesn't mean that you really are certain. It's as simple as that.
DB: Exactly. And of course later on, Avicenna tries to grapple with that, and there are places in which he sometimes sounds as if he's falling into that. But he seems to be a little bit more positively influenced by the Mutakallimūn, although he would never admit that.
PA: Well, speaking of being certain, let's turn to this little treatise, which I guess you were just alluding to, which is called "On the Conditions of Certainty", which is a work you've actually discussed in some of your publications. So this is a very short treatise, and he is trying to lay out, I guess, the different grades of certainty that someone can have. So in order that this doesn't take too long, maybe we should just go straight to the punchline and ask, what are the conditions under which I could take myself to be fully certain or as absolutely certain as he says?
DB: I don't know if that is even going directly to the punchline because that's all the treatise is about. And in fact, it's I mean, so that is the one exception to the rule. I mean, this is a treatise just on epistemology, right? I mean, who is the next person to write a treatise with that title? It's probably Wittgenstein. So he was way ahead of his time. So Fārābī actually has six conditions for what he calls "absolute certitude", and that's very important. There is certitude and there's absolute certitude. So we're already changing the focus a bit from what it is in the Aristotelian tradition, and even the discussion of certitude itself is quite un-Aristotelian. And I won't go into the details here, but as I have discussed, it is really a bit of an accident of translation, but it shifts the focus and allows Fārābī and his successors to do kind of interesting things.
So basically Fārābī presents a definition of certitude and he gives us the genus and then six differentiae and each one kind of builds on the one before it until you get to this very stringent notion of absolute certitude. So basically the idea is that under what conditions can we say that we are certain about some proposition P? And Fārābī is sometimes not entirely clear that it is always propositional knowledge, but that seems to be the main focus.
So first you have to believe that P is true. P has to be true, and Fārābī cashes that out in terms of correspondence with the way things are. Third, you have to know that P is true. And so if you take those three together, it sounds a bit like the contemporary definition of knowledge as justified true belief.
PA: Yeah, it actually sounds like kind of enough.
DB: Yeah, oh no, not enough. It is enough for certain kinds of certitude, but the other three conditions then narrow it down even more. So after believing a true proposition and knowing that it is true, you have to know that it's necessary for P to be true. And I'll come back to that in a second. And then you have to know that at no time can P not be true or become false. And then finally you have to know that conditions one through five are essentially and not accidentally true of P.
PA: Wow!
DB: So let me just say a couple of things about the third, fourth, and fifth, and sixth are somewhat related conditions. The first point is that you might take three, four, five, and six as being justification conditions in some sense. But in fact, Fārābī doesn't speak of them in those terms. And it is clear when he tries to explain what they are that they have some justificatory role, but that is not primarily how Fārābī is viewing it. The third condition that you must know that P is true is very important, and controversial from a philosophical point of view, because that is essentially what we would now call a KK condition, a condition of knowing that one knows. And so the idea is that if you know anything or in this case are certain of it, then you know that you know it. And Fārābī says you can actually generate an infinite number of propositions. So you also know that you know that you know, and know that you know that you know that you know, and so on, ad infinitum.
PA: Not very useful for most purposes! 12.52
DB: Not very useful, but it's a very strong condition. And it is one that captured the attention of Fārābī 's successors. So Avicenna always includes it in his accounts of what certitude is, and then he gives corresponding default conditions for lesser degrees of knowledge. So that is controversial because it seems too hard to meet. But of course, that's part of the traditional ideal that, you know, certitude is something special. It is not something we have all the time. 13.20
So the next condition that it is necessary for P to be true - that might sound again like it is something similar, or it is like the traditional Aristotelian notion that we can only have knowledge of necessary, universal, unchanging things. But when Fārābī explains what it is, he means that we recognize it is necessary in the sense that we have direct evidence for it. And he uses this to rule out knowing something is true by authority or report or something. So he treats it, and this is why it's interesting relative to the KK condition, he treats it as acquiring the knowledge through a reliable method. So he refers explicitly to some reliable way for getting principles or to a reliable inferential method. And usually those two things don't go together, that you both have a knowing that you know condition and a reliablist kind of condition. But that is partly because Fārābī is, again, really worried about people confusing the kind of assent that the theologians talk about and that you get through religious discourse, with real knowledge. 14.30
PA: Is the reason that I know that I know that I use the right method, and I know which method I used?
DB: Yes in part, but at this point, it still has nothing to do with the kind of thing you know. And it is only the last two conditions that throw that in. And that is one of the things that's fascinating about this treatise, because then Fārābī says, okay, in order to have absolute certitude, then you have to stipulate that there can't be any time at which P is not true and that P holds essentially and not accidentally. So that is the absolute certitude. That's what we are ultimately getting to. And so that's going to be only necessary universal truths. But by adding that as a further condition, Fārābī says, so up to this point, you can have a kind of attenuated or relative, maybe we might want to say non-absolute, certitude about temporal contingent propositions. And that is a really major concession. So I can know for certain that I am sitting in my office talking to Peter Adamson on Skype and looking at a blue wall, when I am experiencing it under the proper conditions. And you can call that certitude. And I don't think you would have standards for doing that on the traditional Aristotelian paradigm. 14.25
PA: Just that it wouldn't count as absolute certitude in this terminology, right?
DB: Yeah.
PA: Maybe turning away from this treatise for a second - can I ask you how this all relates to what I guess is the more famous thing about other al-Fārābī's epistemology, which is the role of this separate Agent Intellect, which helps us to come to know? Because what you just said implies that even if I can't get absolute certainty unless I am dealing with necessary universal essential things, which it seems like would be the kind of thing I would get from the Agent Intellect, I could still, for example, be certain that I am looking at a blue wall. So would the Agent Intellect just not be necessary in order for me to achieve that lower kind of certainty?
DB: Okay. I am a little unsure about how to answer that. So on the one hand, I think I have a pretty minimalist reading of how the Agent Intellect functions in Fārābī. So I don't think that it functions as a guarantor. I think that Fārābī believes that we need sense experience to acquire even the primary intelligibles or first principles. And I know other people read him as arguing that those are kind of directly deposited in our minds by the Agent Intellect. But there are a few places where Fārābī seems to think that, in fact, he actually says that most ordinary people, the masses, think that we don't need sense experience to get primary intelligibles, but that's just wrong.
So I think that he, on my reading, he simply sees the Agent Intellect as a necessary mechanism for explaining how we get abstract universals, right? And that it mostly is a metaphor to explain how natural physical things in the sensible world can be illuminated so that we can abstract them. And I think many places he seems to suggest that it is our intellect, the so-called material intellect or potential intellect, that actually does the abstracting. So I think in many ways I am a minimalist generally about the function of the Agent Intellect in Islamic philosophy. I think it is important to see it as kind of a theoretical posit that the philosophers think needs to be there in order to account for certain features of our cognition, but it is not like some external entity that is sort of doing the work for us, even in the stronger cases. 18.20
PA: So just to make sure that I got you right there. So what you are saying is that any time that I come to know anything for al-Fārābī, basically what I am doing is I'm engaging with the world around me using my senses, and the only reason why the Agent Intellect needs to be invoked is to explain how I am able to abstract some kind of universal notions or truths from this sense experience. So it is not like the sense experience sort of prepares the way for me to be zapped with an intelligible form from the agent intellect.
DB: Yeah, it seems pretty clear to me that Fārābī doesn't see the picture as being that way. What I am not so clear about, because Fārābī doesn't have a lot of developed accounts of sense perception and so on, is exactly how he would explain propositions like "I'm looking at a blue wall" or "I know that I'm sitting here right now". I presume that in some sense, well, if you are assenting to them and they're propositional, I think they are going to have to have some kind of an intellectual element to them.
PA: Do you think that it's just that he's enough of an Aristotelian to think that issues like that aren't really germane to the study of epistemology proper? Because what we are really interested in when we do epistemology proper is this higher grade kind of knowledge that Aristotle talks about as the goal in the Posterior Analytics? 19.40
DB: Yeah, I think that ultimately he is very traditional that way and that the major insight is simply well, yes, it is obviously the case that we are certain about all kinds of temporal contingent truths and we should not deny that because that is not a good theory if it is not true, but it is not particularly interesting explaining how that happens.
PA: Right, so it is more like he is adding some detail around the edges of the Aristotelian theory rather than taking a huge jump away from it.
DB: Exactly, although I think in some ways it is a big breakthrough though simply to acknowledge, oh, you can say that because once you've acknowledged that then someone else is going to come along and have to deal with it and explain it.
PA: The floodgates are open.
DB: Yeah, and also, I mean, the other thing is we just don't have psychological treatises by Fārābī and we don't know, I can't recall any reports about those. So part of it is just not something he wrote a lot about.
PA: So speaking of a big jump away from the Aristotelian theory, let's go back to something we mentioned briefly before, which is the role of the Islamic tradition in his epistemology and in particular what he has to say about prophecy. So I guess now I'm a little bit puzzled actually what you might say about prophecy because I would have thought that the explanation of prophecy, maybe this is reading Avicenna back into Fārābī too much, but I would have thought that the explanation of prophecy in Fārābī is that the prophet does get some kind of knowledge from the Agent Intellect. So would you just say that the Agent Intellect is doing something of a completely different kind there than happens in the case of normal human cognition?
DB: Well, it's another sort of yes and no perspective, I think. So Fārābī clearly invokes the Agent Intellect in his accounts of prophecy and he usually changes the terminology in different treatises, but sometimes he speaks about what the Agent Intellect does as revelation. However, when you read the account of what revelation is, it is essentially the same as Fārābī's account of the so-called acquired intellect, which is basically the perfected ordinary human intellect when it is acquired all intelligibles, and Fārābī thinks that that entails or maybe we would say that there supervenes on that some kind of a direct knowledge of the Agent Intellect. So that is the intellectual component of prophecy, but that is attainable in philosophy and that is attainable through totally natural means. So in some sense it sounds supernatural because he's invoking a separate intellect, but it's just the way that he invokes a separate intellect as a natural mechanism in the way we were just talking about a minute ago. 22.30
What makes the Agent Intellect distinctive in prophecy for Fārābī is that the prophet has a very strong imaginative faculty, which is a faculty in the body and because of that the influence of the Agent Intellect can spill over or overflow into the imagination and so the prophet receives imaginatively the knowledge that his intellect acquires naturally, but the imagination is a symbolic imitative faculty. So in order to express what the imagination has received, it has to, if you will, translate intelligibles into appropriate images and symbols - and that is what the prophet does and that is what religious texts show us. They show us a symbolic imaginative expression of the truths that the prophet knows through his intellect, but the truths are exactly the same as the truths of philosophy and in a virtuous religion they should map onto the rational deduced truths that you would find in say, Aristotelian philosophy.
PA: So actually you get a kind of parallel structure there where the relationship between demonstrative philosophy and religious discourse is the analogue to the relationship between the human intellect and the human imagination, so the two things are running alongside each other.
DB: Exactly, exactly. And so that view of there being something special about the prophet's intellect is something that you find in Avicenna to the extent that Avicenna thinks that perhaps Fārābī's account of the uniqueness of the prophet is too much focused on the imaginative faculty and so he suggests the prophet has an especially strong intellect, but he still agrees that the content is the same as the content of philosophy, just the prophet acquires it all at once and without having to go through all the work. And then for the consummate prophet, the overflow into the imagination and so on would be exactly the same.
PA: Okay, well actually already from a lot of things we have said, it has been clear that there are several respects in which al-Fārābī is anticipating the epistemological views we find in Avicenna. So even though I haven't covered him yet, I thought maybe I could close by just asking you to what extent does al-Fārābī already kind of anticipate what we find in Avicenna, and what would you say are the main shifts away from Fārābī that we then find in Avicenna? 25.10
DB: Well, I think a couple of things I have already mentioned that I think are central that Fārābī provides a framework that Avicenna and all the other traditional Aristotelian philosophers take over. First of all, like Fārābī, Avicenna takes over the focus on certitude versus knowledge. He includes this "knowing that you know" condition in the definition of certitude. He also takes over the expansion of the organon to include the rhetoric and poetics, which goes back to the Greek commentators, and the general stratification of modes of cognition that that brings in. So that structure is in all of Avicenna's accounts of the principles of knowledge. And something I didn't mention but is the basic distinction between conceptualization or concept formation and assent that is used to frame all epistemological discussions to solve Meno's paradox and so on. That is something that you find clearly formulated in Fārābī and that Avicenna takes over.
Where I think Avicenna makes particular strides as far as epistemology goes is first of all in the discussion of the principles of knowledge themselves and their epistemic classification. So it is a sort of aspect of the discussion of certitude. And Avicenna has these elaborate classifications of all of the different possible sources of knowledge based on their psychological mechanisms, the sorts of use that is made by the intellect of other faculties and so on. And all of his treatises corresponding to the Posterior Analytics start with this. And so this is partly inspired by Fārābī and partly Avicenna taking over an approach of the Islamic theologians and making it his own.
Along with that, one thing you see in Avicenna is a slightly more positive development of some of the aspects of epistemology that we now call social epistemology. So discussion of the sources of knowledge that we derive from the testimony of others. And so in a couple instances Avicenna actually tries to explain how testimony is a source of certain knowledge which Fārābī would never admit. And that is something he takes directly from theology. In other cases, while he still treats testimonial knowledge as inferior and as a form of dialectical or rhetorical assent, he also develops an account of cases in which knowledge that we acquire on the basis of what we learn through social interaction and through conventions. Those kinds of knowledge can only be learned that way, and in particular this is the case with ethical dicta and so on. So there is a real, further broadening of the scope of types of knowledge that are of interest to the philosopher and have different degrees of assent. 28.15
Also, one thing that Avicenna takes over just from a small hint in Fārābī, occasionally Fārābī will talk about empirical knowledge and he will mention it as a source of certitude but he won't really explain it. And Avicenna develops an elaborate justification of why we have necessary knowledge of sort of causal empirical propositions like aspirin relieves pain. He usually likes to talk about scammony purging bile but that's not as attractive as an aspirin example. And so he tries to explain the cognitive mechanisms that go into that and why it produces certitude and produces reliable knowledge.
I guess the main move in a completely different direction from Fārābī is Avicenna's including introspective knowledge and self-awareness as among the foundations for knowledge. And this of course is something that Avicenna is very interested in and thinks is sort of the core kind of knowledge we have is our knowledge of ourselves. But Avicenna also brings that into his epistemological discussions and it also seems to license for him the use of thought experiments which he loves to do. And that is something I don't think that Fārābī really considered to be an important or even perhaps he didn't even entertain that as a possible philosophical method and that's a very distinctive of Avicenna.
PA: Right, so that's looking ahead to things like the famous flying man argument and so on.
DB: Yeah, exactly. Exactly.
PA: Well, I will be getting on to Avicenna in a few more episodes. First I'm going to be looking at philosophical and scientific developments in the time between al- Fārābī and Avicenna or around the time of Fārābī's death and taking up to the time of Avicenna starting with the mathematical sciences and in particular the use of geometry to study optics which will be the topic of the next episode. But for now I'd like to thank Deborah Black very much for coming on the show.
DB: Thanks very much for having me. It's always fun to talk about al-Fārābī.
PA: Yeah, I agree.
Comments
Agent Intellect
That is a very fine discussion with Deborah, Peter. But I will have to write to her or Skype her for a few more minutes discussion of the Agent Intellect in al-Farabi. I know she inclines to seeing it as a final cause rather than efficient but she did not state it that way in the podcast. At any rate, I am not quite certain how she sees the AI to function in cognition and in the constitution of the world. In particular I am not clear on how prophecy works if the AI is primarily or solely a final cause. I could, of course, construct an account of how it draws the prophet to actuality through finality but I don't see that in al-Farabi.
But, as you know, I regard Averroes not only as trying to be a pure Aristotelian in the 12th c. but also as one who takes the principles and leanings of al-Farabi to the bitter end, and the bitter end is the collapse of Classical Rationalism with Averroes. What am I talking about? Well, in his (mature or perhaps nearly all) philosophical writings there is no afterlife for Averroes, there is no particular providence, there is no de novo creation, there is no theory of prophecy (and so there is no philosophical account of scripture as a divine donation [unless one gives a contorted account through final causality]), and — following al-Farabi — the purest truth is philosophical, even to the point of Averroes saying that the most perfect form of worship is to study God and God's 'creation' in the science of metaphysics (the shari‘a of the philosophers). The collapse I will explain another time but you see where I am going.
But I am off for another morning walk and I will listen to you and Deborah one more time attentive to the issue of the AI in al-Farabi.
Thanks for these fine discussions. I will be requiring them for my undergrads in Ancient Philosophy in Spring 2014 and for grad students in my Fall 2013 Aquinas course with MU and KUL. I wonder when you will reach Aquinas. Probably not this year, I suppose. But keep up the good work. For me it makes for entertaining and clarifying walks.
Richard
Other philosophers adopt the idea that prophet must align with r
Peter,
Sorry again to return to much earlier episodes. I am looking for any other philosophers who may have made similar claims to those discussed in this episode by Farabi. As I understood this part of the discussion, Prof. Black maintained that Farabi asserted that though the Prophet may get knowledge at once, and "without doing all the work", that it still had to align with what knowledge could be obtained through the methods and requirements discussed. Were other philosophers holding that or discussing it?
In reply to Other philosophers adopt the idea that prophet must align with r by Dave Lavery
Prophecy
Yes indeed, in fact there is a passage in al-Kindi (in "On the Quantity of Aristotle's Books") where he says this, almost verbatim. Other "naturalistic" or "rationalistic" accounts of prophecy were put forward by other philosophers, for instance Ibn Sina (Avicenna). It is something of a hallmark of Greek-influenced philosophy (falsafa) in the Islamic world.
Oh and don't hesitate to comment on old episodes, that's what the comment function is for!
Add new comment